Some useful context: this is almost certainly being driven by Apple’s Private Cloud Compute architecture and not tariffs, as an investment of this magnitude is not planned overnight.
Why is PCC driving Apple to spend billions to build servers in the states? Because it is insane from a security standpoint (insanely awesome).
PCC is an order of magnitude more secure server platform than has ever been deployed for consumer use at planet scale. Secure and private enough to literally send your data and have it processed server side instead of on device without having to trust the host (Apple).[1] Until now the only way to do that was on device. If you sent your data for cloud processing, outside of something exotic like homomorphic encryption[2], you’d still have to trust that the host did a good job protecting your data, using it responsibly, and wasn’t compromised. Not the case with PCC.
To accomplish this Apple uses its own custom chips with Secure Enclaves that provide a trust foundation for the whole system, ultimately cryptographically guaranteeing that the binaries processing your data have been publicly audited by independent security auditors. This is the so called hardware root of trust.
It is essential then that the hardware deployed in data centers has not been physically tampered with. Without that the whole thing falls apart. So Apple has a whole section in their security white paper detailing an audited process for deploying data center hardware and ensuring supply chain integrity.[3]
You can imagine how that is the weak point in the system made more robust by managing it in the US. Tighter supply chain control.
Trusting Secure Enclaves custom chips over processing locally is going to be a hard to impossible sell for those who truly care about privacy.
Thankfully for Apple that's a very low number in a world where people demand tiktok remain legal when shown how their data is being used by foreign actors. People only care about privacy when it's local (don't want mother to find out, neighbours to talk, friend to think a certain way about you or classmate stalking) and that's why ai fakes are much more concern then a company knowing everything you do.
But this product is great for fortune 500 businesses.
I think this is a level of security Apple is providing at additional cost to themselves that only a tiny fraction of consumers would even pay an extra cent for.
From that perspective I really appreciate this effort by Apple.
No dog in this fight, and I agree with the premise, however there was never a time Apple made a ton of noise about MacBooks being manufactured in the US.
There was a ton of noise about Mac Pros being manufactured in the US, but sadly, I am not nearly as familiar with Apple after, say 2018*. Not even sure if they have a Mac Pro anymore. :X and if they do, I assume it's not the same model (the black trashcan), so it makes me wonder if they bothered retooling here, or quietly moved it somewhere else
* TL;Dr at some point it became clear to me Cook is Sculley 2.0. I date it to around walking around NYC and seeing an absurd amount of Apple News bus-stop ads. Services! (TM)
Much as it might be pandering to Trump's nationalist (America first / American only) policies, or simply an action to avoid some of the effects of tariffs that might be imposed, this time around, I see no such connection for the 2021 announcement. Unless they are connected as they are a more generic "pleasing the incoming administration" to try curry favour for when decisions that might affect the company are being made.
"Apple’s most recent announcement on US investment was a 2021 promise to spend $430 billion over the following five years, including a 3,000-employee campus in North Carolina, though development on that project has since paused."
- a larger investment number in a previously announced Austin campus
- new factory in Houston "which will create thousands of jobs"
- "doubling its $5 billion US Advanced Manufacturing Fund to $10 billion"
- "It will also open an Apple Manufacturing Academy in Detroit in which Apple engineers and other experts will offer consultations to local businesses on “implementing AI and smart manufacturing techniques,” along with free classes for workers."
Coincidentally, construction isn’t set to start until late November 2028—convenient timing. If this mess blows over, they can quietly backpedal and carry on like nothing happened.
Good timing because Trump should be significantly weaker, and it'll be clear where Trumpism is headed in the culture, but also it will be more clear where AI will end up.
Even if the move forward with investment, they will be a bit of a 'late' mover, but will have had a chance to see what is working and what isn't working for everyone else.
1. JD Vance independent of Trump will have the same policies.
2. JD Vance will have enough popularity for a serious 2028 run. He might fall out favour with Trump as Trump tries to mount a bid for a third term, Trumpism might just generally lose popularity if policies lead to bad outcomes.
3. Dems don't figure their shit out. They should be able to take back some control in mid-terms, and then start to push their own policies, or at least credibly show that the most extreme policies from the exec branch don't have teeth anymore.
And you're assuming that the 2028 election (and 2026 for that matter) will be business as usual elections, against all evidence to the contrary staring us directly in the face.
The federal government doesn't control elections in the US so they don't have much power here. Also, firing all the FBI agents is a bad first step to using them for interference. They have no clue how to be authoritarians; to do that, you need to be popular and have the security forces like you.
Going to be watching closely - but cynically, a promise of investment (for avoidance of tariffs) only needs to last one news cycle until tariffs are no longer top of mind. Then it can be walked back without tariffs being imposed.
Maybe instead of saying the t-word tariff, US gov can charge Apple a special fee on each iPhone. They can call it something catchy, like say, a Core Technology Fee.
This should be the top comment. Apple are doing business the way business is done, just like last time. Results don’t matter, it’s economic policy via press release. Form over substance.
Results matter, it's not hard to imagine that Apple considers the real risk of its promise and market position of being the privacy option being undermined by their supply chain risks, and leverage being used against them by privacy unfriendly actors.
What's the added risk here? It's fine to "risk" almost the entire iPhone itself to be manufactured in China but the servers for some random AI features need to be pure?
Sounds more like technical marketing and the company will treat any decisions around it as a marketing exercise.
Apple's commented previously on why they build in China, and it's beyond just the pricing - the supply chain for every single part they use is in China and mostly in the same geographic region, so there's a level of flexibility there they couldn't get in the US. It wouldn't surprise me if it was genuinely a goal for Apple to manufacture more in the US - they're a notoriously privacy-focused (corporate, not end-user) company, and China's known for IP wandering its way off campus. They're not going to sacrifice the iPhone economics until the US option is actually viable, but I'm not surprised they keep kicking the tires on US manufacturing.
> "the supply chain for every single part they use is in China"
Not entirely true. Some of the highest value components in an iPhone, including the CPU/SoC, baseband, and the majority of OLED displays, are sourced from countries that are not mainland China.
> They're not going to sacrifice the iPhone economics until the US option is actually viable, but I'm not surprised they keep kicking the tires on US manufacturing.
Apple could, with its immense cash hoard and cash flow, _make_ the US viable, but it chooses not to because it'd rather take the easy way out and have China or India or $COUNTRY fund it and return money to shareholders. They've returned money to shareholders rather than invest it in US operations, by design.
This is a classic feint to protect Tim Cook's entire raison detre. He built his career on super high efficiency operations by outsourcing to cheap labor countries. It relies on the low-to-no tariff access to US consumer money.
And I don't care that it's better for their stock price; that's Apple's problem not mine as a US citizen. And even as an Apple investor I would rather the money be spent on US on-shore operations.
It's mostly about cost and market access to China.
Most smartphone supply-chain for Samsung and Apple exist outside China -- primarily in Japan (camera, sensors), South Korea (DRAM/NAND, OLED), and the US (various ICs fabbed at TSMC in Taiwan). There are quite a few reliable estimates/teardowns showing that these three countries account for close to about 90% of iPhone BOM (bill of materials). That's one reason why Samsung's smartphone unit was able to pull out of China without much disruption back in 2019 -- ie, low dependence on China.
I feel that Apple has pushed this misleading narrative a bit too long to defend their massive China outsourcing.
They've actually been diversifying iPhone manufacturing away from China for a few years already. As of April 2024, 14% of all iPhones were already manufactured in India. That's around 30 million phones per year. And Apple plans to double their India manufacturing again by 2028.
Results don't matter as much as PR, this is time when this is unfortunately valid. Just look at US elections.
Measurable results affect rational aspects of our minds, PR attempts to attacks directly emotions bypassing the former, ie to induce impulsive shopping.
Also, what actual security? Apple is as vulnerable as cheap chinese phones against state actors using 0days. Apple devices are still being stolen for spare parts, Apple doesnt secure each component AFAIK and thieves know this (very recent case with friend of a friend, they even knew how to bypass that built in airtag tracking). I haven't seen anything but very well crafted PR statements on this topic. All money-accessing apps on absolutely any phone are a security risk.
> This should be the top comment. Apple are doing business the way business is done, just like last time. Results don’t matter, it’s economic policy via press release. Form over substance.
If the Trump administration has any competence, they will rub those old promises in Apple's face until Cook actually does something meaningful.
The whole Trump administration is all about form over substance, though. I would not expect Trump to do anything actually productive about it, as long as Tim Cook sings his praise (and pays his dues).
Charitably, tariffs exist so POTUS can either lower taxes or increase jobs in US, but both would take time to pan out assuming things go well. So if a company is willing to onshore money or jobs, its achieving its intended purpose in their eyes.
No I think Mexico/Canada were largely about stopping immigrants and fentanyl smuggling. But China was targeted for not doing enough to stop manufacturing of fentanyl precursors.
Sorry to say but that's already been walked back on after Canada committed $1 billion dollars for extra northern border security and it made no difference in the tariffs discussions.
- According to CIS, the number of Canadian crime groups producing synthetic drugs doubled between 2023 and 2024
- There's a lack of Canadian agents who are tasked at preventing this and current legislations make it very inefficient between federal and provincial law agents
- There's an upward trend in Fentanyl seizures in Canada the last 2 years
- Fentanyl is now being produced domestically in Canada
All of that is within the control of Canada with better policies.
Let’s put it into perspective, because those numbers don’t give a baseline for what the problem is. Also they don’t necessarily have anything to do with trafficking.
Last year there were 45 lbs of fentanyl intercepted crossing into the US from Canada. Thats a backpack. There’s 500x as much coming from Mexico.
It’s unrealistic to expect that zero fentanyl will come into the US from Canada, and until that happens we will tariff all trade with them.
Yes this is one reason tariffs are so valuable to a corrupt POTUS. They have essentially unilateral and very fine-grained control over them, down to exempting specific companies or products outright.
Congress needs to step up on this, honestly. The entire idea that the President can unilaterally implement trade policy is as plain a violation of separation of powers I can think of, and SCOTUS is a fan of non-delegation doctrine.
Legislators step up when enough of their voting constituencies make it clear that they value something as a non-negotiable (assuming votes still matter).
Which means those who care about this are back to not only contacting legislators but also persuading a lot fellow voters that separation of powers is crucial and worth prioritizing over familiar well-handled and loved heuristics.
Yes a million times. For all the rhetoric about authoritarians, the Democrats never seem to want regin in Executive power when they are they majority. It is like a game of chicken where America winds up with a populist dictator from either the left or right.
Yes because populism is a reaction to government being generally unresponsive to people’s needs.
Congress has become increasingly unproductive and unresponsive. There are many popular policies that Congress essentially ignores, and many problems that go unsolved. So trust in government dwindles and people crave strongman solutions.
I’m not sure there is a solution. There are so many interlocking problems gumming up the process that any “we just need to fix X” solutions (where X is gerrymandering, money in elections, lobbying, the two party system, first past the post, corruption, income inequality, the electoral college, the slow death of journalism, consolidation of industries, etc) are nearly impossible and also probably insufficient because they all feed back into one another: they are both causes and effects.
So when people are mad about a downstream effect like the price of eggs and digging any deeper touches one of the topics above (“to fix egg price gouging you need to reinvent the political system” sounds a lot like “to make an omelette first you need to create the universe”), it’s really easy to throw your hands up.
Clearly a non-sequitur, but I'll bite anyways. Paul Pelosi is a VC, and I'm sure that most of the Pelosi net worth is due to his income, not hers. Since the laws preventing members of Congress from trading on information they receive as part of their duties, you can't say that the Pelosi's have violated any securities laws.
And both sides of the aisle benefit from this. Whether it's legalized insider trading or jumping to corporate jobs when out of office, it's a corrupting influence. All members of Congress, SCOTUS and POTUS should have to place their assets into blind trusts. That won't stop this corrosive influence, but it is the bare minimum.
> Since the laws preventing members of Congress from trading on information they receive as part of their duties, you can't say that the Pelosi's have violated any securities laws.
> Since the laws preventing members of Congress from trading on information they receive as part of their duties, you can't say that the Pelosi's have violated any securities laws
The feds can and do go after people for using family members and friends to execute trades. So if Nancy Pelosi told her husband some material non-public info, and Paul Pelosi traded on it, that would still be insider trading.
There was a guy at Microsoft who was caught once using a friend to place trades. He said he talked himself past his ethical concerns by reasoning that members of Congress do it.
Edit: to be clear, the absence of a prosecution does not mean that the Pelosis did not insider trade. Nor that they did. We can't tell from this distance, only speculate.
"Paul Pelosi, 83, sold 30,000 shares of Google (GOOGL) stock in December 2022, just one month before the tech giant was sued over alleged antitrust violations."
iirc as soon as anything becomes beyond the border the President holds the keys for various reasons including the ever-vague “national security” but also due to being prescribed as the primary negotiator https://en.m.wikipedia.org/wiki/Treaty_Clause
Tariffs are not treaties, they're taxes/duties - only Congress has the power to raise them. Article I is extremely clear on that. Historically, tariffs were always raised by acts of Congress, not by Presidential fiat.
Trump is using extremely misguided legislation from the 1960s/70s where Congress allowed the President to enact tariffs for national security and emergencies. There is a very strong argument (in the sense it resonates with the conservative SCOTUS majority) that Congress cannot delegate its fundamental powers to the executive by legislation alone.
I think people are just too cowardly to bring a case in front of the courts to challenge the constitutionality of it all. Non-delegation doctrine is what the Federal Society want to use to kneecap all federal regulation. Trump operates on a spoils system so it's not in the interest of conservatives or businesses to challenge him, for fear of retribution.
Trump is using tariffs not to raise revenue, but rather use it as a stick to force companies to invest in USA.
Previously they were outsourcing and offshoring as much as they could get away with it. Which led to transfer of advanced technologies outside USA and America losing its manufacturing and technology edge
So how's that going? Outsourcing seems to be going strong, the tarriffs instead pissed off allies who are preparing counter-tarriffs, and the CHIPS Act is being dismantle as we speak (there goes our investment.
I have already explained my thinking up this comment chain. I'm mostly replying to GP who misunderstands that the intent of the tariffs is besides the point.
TL;DR read Article I section 8, read up about the Trade Expansion act of 1962 and Trade act of 1974, and "non-delegation doctrine", and you can trivially find legal debate about the constitutionality of IEEPA. Rather than listen to random nerds on HN you should seek out this information yourself.
Tariffs are a form of taxation. If I want to import say tea, and the government is placing a tariff on that imported tea, I am effectively taxed by the government. And only Congress can impose new taxes.
Not saying you're wrong, but... I have seen claims that tariffs are a source of government income that Congress doesn't control. You're claiming they do.
I haven't seen a citation from either side. Can you substantiate your position?
There is no jobs problem in the US though. Unemployment is at 4% which is mostly just job churn. Long term unemployment is only 1%.
US consumers, that’s all of you, are being hammered with taxes on imported goods most of which can’t realistically be produced in the US anyway, to solve a problem you don’t have.
A commitment like this takes years to plan. It can’t possibly be a response to tariffs announced weeks ago. This is all optics.
I fully believe that the real ("main street not wall street") economy is in worse shape than government numbers on unemployment suggest and both sides are to blame for different aspects of this problem.
But nothing Trump is doing is going to fix your situation.
In fact he (or rather Elon/doge) is very actively making things worse for you with the massive government layoffs, flooding the market with even more people to compete with you for jobs making finding work more difficult and also eventually dropping all of our wages.
>But nothing Trump is doing is going to fix your situation.
I'm aware. I'm sure he's responsible for at least 3 job freezes I ran into mid-interview this year. He's literally costing me job opportunities because no one can budget around this chaotic government.
4% too many and probably understated. The BLS repeatedly underestimated unemployment during the previous administration. Also the labor participation rate, which is harder to game, still hasn't reached pre-Covid levels yet: https://www.bls.gov/charts/employment-situation/civilian-lab...
In February 2020 it was 63.3% and in January 2025 it was 62.6%, for a difference of 0.7%. Also note the steady decline post-2008 and the multi-year plateau that jitters around 63%.
Having the plateau change from ~63% to ~62.5% isn't an unreasonable scenario.
False, USA has a big problem with manufacturing. All US jobs are service jobs to prop up consumer economy, that have no strategic benefit.
A lot of fake employment and low productivity jobs are in the government/NGO sector, paper pushers, DEI jobs, law/compliance type jobs - that should have been manufacturing jobs instead.
USA has no shipyards and infrastructure is crumbling precisely because of misallocation of resources and labor
The government determines the employment rate via surveys, i.e. they just go and ask people if they're employed. It's not a calculation from taxes or from employers or anything.
So it's up to the gig workers if they think they're employed or not. Presumably this depends on how often they do it.
Apparently, yes. I saw mention of discussion around the Trump administration potentially giving Apple a tariff waiver. And I believe in Trump’s last term, Apple did have some sort of waiver.
I’m on mobile but Googling for “Apple tariff waiver” and “Apple tariff exemption” will point you to several news items.
I am a free trader in principle. However you have a country (China) with an authoritarian government that makes favored industries subsidized.
Of course the standard economic argument is that China using its GDP to make goods cheaper for our own citizens to purchase is better for us - they are subsidizing our economy. However it ignores the strategic disadvantage by our country losing its manufacturing capabilities.
The graphs may show economic advantage. It’s hard to quantify the long term strategic and militaristic disadvantage to not being able to make anything yourself if a world war occurs.
> However you have a country (China) with an authoritarian government that makes favored industries subsidized.
This is overlooking the forest for one tree. The thing is, mean chinese manufacturing wages are $25k/year (purchasing parity adjusted! $15k unadjusted) for a 49h week.
That is the reason that so much manufacturing/industry has shifted there, not some nebulous "Chinese government subsidies" (not saying those are not a thing, just that they don't really matter all that much).
> It’s hard to quantify the long term strategic and militaristic disadvantage to not being able to make anything yourself if a world war occurs.
Certainly. But forcing low-skill industry to stay at a relevant size in a high-wage country is expensive business (compare agriculture, which is subsidized basically for exactly this reason) and not straightforward (see Jones act).
Presenting tariffs as a viable alternative to taxation is just beyond ridicule, but that has not stopped people so far either...
Salaries are just a small part of the reason industry works in China.
The bigger picture is that China invests in the development of an industrial chain. This has many aspects: infrastructure, education, training, housing, and of course tax incentives. The USA decided to stop investing in practically all of these. Even scientific research, the last area in which the US used to lead, is now in jeopardy from both sides: competition from China and internal cuts.
Absolutely. You do need a minimum baseline for infrastructure, government stability and workforce.
Most of Africa is just starting to slowly get there, Bangladesh is already very relevant for textile production.
I would expect the same basic trend to repeat that we saw with electronics manufacturing in 90s Japan:
First cheap products move (very wage sensitive), then the local sector expands, wages rise with the whole local industry moving up the value chain, then at some point local wages become high enough for the whole process to repeat with the next low-wage country...
I think trying to block this trend off with tariffs is a futile waste of taxpayer money which american consumers are gonna pay for.
Spending tax money to keep some degree of self-sufficiency in critical industries (like with agriculture) can be a solid idea if done sparingly and cleverly, but that is not how the current US admin has approached this...
"The company responded by intensively lobbying the U.S. government to intervene and mounting a misinformation campaign to portray the Guatemalan government as communist.[18] In 1954, the U.S. Central Intelligence Agency armed, funded, and trained a military force that deposed the democratically elected government of Guatemala and installed a pro-business military dictatorship.[19]"
Sour grapes. Most economists were just happy with this situation until recently. What I mean is, the current situation arises by the desire of Western businesses of getting hid of productive investments and concentrating only on capital investments. It has nothing to do with trading with an authoritarian government or not, which almost everyone believed was Ok until recently.
> The graphs may show economic advantage. It’s hard to quantify the long term strategic and militaristic disadvantage to not being able to make anything yourself if a world war occurs.
Is the United States at risk of not being able to make anything ourselves? We have the second largest manufacturing output in the world.
Maybe the next administration should keep up the tariffs (as Biden did to a degree). Cheap trade with China distorts the tech sector too. Jobs and Wozniak were the products of a system in which americans had to build products at home. Tim Cook is the product of a system where you can become a trillion company by hyper-optimizing foreign supply chains. Which is better?
You’re incorrect about history. Mercantilism not only restricted foreign trade, but restricted domestic industrial development by requiring the colonies to sell raw materials to Britain and buy finished goods from the Britain. Tariffs were a core pillar of the Lincoln Republican Party.
There’s been an isolationist wing in tech as long as I’ve been in it (early 2000s). I remember chatting with someone at Cisco/Juniper in the late aughts about Huawei ripping off their router designs down to the silk screening. Of course today Huawei makes their own state of the art routers with their own silicon, and some lower-end Cisco/Juniper gear is white boxed foreign equipment. And of course tech folks were complaining about immigration and outsourcing back in the early 2000s when Republicans were enthusiastically supporting both.
Having access to cheap oversea steel allows Americans to focus on building companies with significantly higher value-add. Onshoring low-value industries is a massive human capital waste and an easy way to depress wages.
It allows a very small portion of Americans to build companies with significantly higher value-add.
It destroyed the futures of a larger number of Americans.
Then again, why do we make the distinction "American"? If you have people who became unfathomably wealthy by shipping off strategic industries to the lowest bidder regardless of geopolitical implications, does nationality matter anymore?
No, the analysis (and it’s not exactly rocket science) says just the opposite: Way more downstream manufacturing jobs that rely on steel as input are lost, vs. domestic steel production jobs gained.
Do "normal Americans" pay taxes? From the numbers I've seen, ~1/3 - ~1/2 of tax filers receive more money from the government than they pay. To them, "refund season" is a cause for celebration rather than a stressful event.
The on-average crossover between negative and positive net total federal income (individuals will differ because of individual circumstance beyond just income level) tax when taking into account refundable credits (most notably, but not exclusively, EITC) is a bit below the median personal income but not that far below it, so certainly lots of individual "normal" (by most reasonable definitions) Americans do not pay net federal income tax .
But even if they don't pay net federal income taxes, they probably still pay a net positive amount in a variety of state taxes, federal payroll taxes, and federal consumption taxes (e.g., gas tax.)
The problem with taxes is that it's a prisoner's dilemma. You need global cooperation at some base level of taxes, otherwise companies move to more favorable tax jurisdictions in the long term and offshore from there, which would hurt the US even more. It doesn't have to be all-or-nothing, but any marginal dollar of increased taxes in one place will have some non-zero effect of encouraging the next investment dollar to be spent elsewhere.
To be clear, I do think capital gains taxes are criminally low in the US relative to income tax, so I'm not arguing in _favor_ of lower taxes. I'm just saying why raising taxes isn't a panacea.
Creating an underclass that relies on economic elites paying taxes rather than being economically independent because you want to optimize for "high value add industries" is a terrible long term strategy.
> tax those people appropriately and pump that money back into the economy
So make the US to be like a far less successful country? Kill your economy by increasing taxes? The US economy is singularly successful because it has incentives to build businesses - see YC.
Have you tried living in a country that doesn't encourage businesses? They are often great tourist destinations. I'm in New Zealand and too many ambitious young people leave here: we have an emigration problem because our economy sucks. The government fixes the economy with 30% immigrants (disclaimer: I love immigrants). I have many friends that are never coming back here except for holidays. I hate the New Zealand government incentives for businesses (taxation and regulation) and I can see no way to fix them. Even our "business" political party ACT is completely fucked (latest story - they will be selling everything profitable to overseas "investors" - destroying the economy).
Taxation incentives matter to businesses. Be careful what you ask for because the majority have little understanding and vote for the wrong incentives.
Even business owners don't seem to understand incentive systems that well. Perhaps game designers do?
However I believe that incentives need to be marginal. If you already have a lot perhaps you need a big carrot as your incentive? I don't know any billionaires that I can ask how they feel about taxation incentives: I reckon you are making assumptions about what you think they should feel.
What makes Tim Cook make the US more money?
Taxation cliffs are shit. In New Zealand our Green party decided that 1 million was enough. Why would you bother growing a business after you reached 1 million? Retirement? A business is defined as being about making money (albeit some people do run "businesses" for other outcomes - why is Warren Buffett still working?).
High marginal taxation is also shit IMHO.
The hard part is to design the incentives so that productive people build your economy for the benefit of everybody.
If a government discourages business then the economy is crap and everybody suffers. See other economies.
Few people understand the incentives of others, and few people understand how wealth is created for all: the hoi polloi dismiss the wealthy as vampiric money grubbers. Anyone who uses the word capitalist in a derogatory way has been brainwashed. Most everything that makes our economies work is invisible non-monetary rewards. https://news.ycombinator.com/item?id=43162596
I can speak for my own financial incentives. My perception is that I have an effective tax rate of well over 50% in New Zealand (any retirement savings are not safe because our demographics and governments will screw our economy).
I do not feel the incentive to work in a business - My attitude means I now produce marginally less than I could for the New Zealand economy (I still pay taxes so they are advantaged but they could get a lot lot more from me). I now mostly selfishly concentrate on those closest to me. Why should I work if it isn't marginally beneficial enough for me? I'm no more selfish than my retired friends that I know (a wide variety of people from many walks of life).
(Reëdited to expand and clarify).
We can't decide how much is fair. Compare yourself to a dead king - what is fair? We can design systemic incentives so that we each make the world better for everyone. Not that that it is easy... Trite thoughtless dismissals of the most productive members of society are not helpful.
Edit 2: I guess this discussion is as close to work as it gets for me. Too much adulting. Should I get into politics? Are morals an impediment to helping others? There are too few politicians I admire, and too many I wouldn't want to shake hands with or be associated with. Every idiot has political opinions - how much of an idiot am I? Every politician is smart enough to win an election - they are not stupid yet they make too many horrific mistakes. What about the cryptically smart ones? I see how systems affect people that join a system. What would I become if I join our political system? Understanding our different systems is hard because they grow so weirdly with vestigial complexities due to history, complex interactions, and reflexivity.
I get the value add argument, but lots of people just need income to pay for living expenses. Without an income, those people become disaffected and sometimes violent. Then they embrace right-wing protectionism because, while their gadgets are cheaper, they have no income to buy cheap gadgets.
Nor can they move to these offshore places (where the cost of living is lower) because immigration laws exist in part to control worker mobility.
> Having access to cheap oversea steel allows Americans to focus on building companies with significantly higher value-add. Onshoring low-value industries is a massive human capital waste and an easy way to depress wages.
That's the talking point, but it's bullshit. A lot of those "low-value industries" are fundamental capabilities, and China sure as hell isn't going to let the US own the "higher value-add" areas. They dominate those next, and the US free-trade business elites will be fine with it as long as they get to make some money for themselves.
Being a high value-add area is endogenous to how hard it is for others i.e. China to reproduce. In other words, if it were easy to make GPUs they wouldn't be so damn expensive.
China's economic power is certainly not rooted in their isolationist social policies. They're just as bullish about foreign investment as the US was at the height of the free trade era.
>> Isn't this why we declared independence in the first place? To get away from the British restricting free trade?
No. I'm not sure where you got that idea. If you look at something like the Boston Tea Party, it wasn't high taxes on tea that were being protested against, it was lowered taxes on tea that undercut the smuggling operations of people like Sam Adams and John Hancock. "No taxation without representation" makes better press than "No undercutting my smuggling operation" though.
In the early years of the US, between 80 and 90 percent of federal revenue came from tariffs. Not exactly free trade.
> In the early years of the US, between 80 and 90 percent of federal revenue came from tariffs...
To be fair, the Federal Budget back then was 2%-ish of GDP. And their political consensus gave the Federal Gov't very few things that it had the power to tax.
I want to say “that’s not what isolationism means”, but I realize it starts to feel vague just like the word “fascism”, used when convenient but varies wildly in rhetorical meaning… to be more specific is better, I like what George Washington had to say about it in his farewell address because it shows the nuance of the topic across the spectrum, it’s not as simple as isolation good vs bad:
The great rule of conduct for us in regard to foreign nations is, in extending our commercial relations to have with them as little political connection as possible. So far as we have already formed engagements let them be fulfilled with perfect good faith. Here let us stop.
Europe has a set of primary interests which to us have none or a very remote relation. Hence she must be engaged in frequent controversies, the causes of which are essentially foreign to our concerns. Hence, therefore, it must be unwise in us to implicate ourselves by artificial ties in the ordinary vicissitudes of her politics or the ordinary combinations and collisions of her friendships or enmities.
Our detached and distant situation invites and enables us to pursue a different course. If we remain one people, under an efficient government, the period is not far off when we may defy material injury from external annoyance; when we may take such an attitude as will cause the neutrality we may at any time resolve upon to be scrupulously respected; when belligerent nations, under the impossibility of making acquisitions upon us, will not lightly hazard the giving us provocation; when we may choose peace or war, as our interest, guided by justice, shall counsel.
Why forego the advantages of so peculiar a situation? Why quit our own to stand upon foreign ground? Why, by interweaving our destiny with that of any part of Europe, entangle our peace and prosperity in the toils of European ambition, rivalship, interest, humor, or caprice?
It is our true policy to steer clear of permanent alliances with any portion of the foreign world, so far, I mean, as we are now at liberty to do it; for let me not be understood as capable of patronizing infidelity to existing engagements. I hold the maxim no less applicable to public than to private affairs that honesty is always the best policy. I repeat, therefore, let those engagements be observed in their genuine sense. But in my opinion it is unnecessary and would be unwise to extend them.
>when belligerent nations, under the impossibility of making acquisitions upon us, will not lightly hazard the giving us provocation; when we may choose peace or war, as our interest, guided by justice, shall counsel.
Ironic that as a Canadian, the US is moving from the nation that would be guided by Justice into the belligerent nation in this situation.
It also serves as a lesson to us that we should have learned from you and George Washington, and stood on our own first and ensured our own security before cooperating with others. We have a long way to go to get back there now, unfortunately under the position of potentially our closest ally and economic partner being belligerent, untrustworthy and unreliable.
Apple went bankrupt under Jobs and Wozniak and was saved by hyper optimizing foreign supply chain company Microsoft only to rise 10 years later by focusing on hyper optimizing foreign supply.
There was a lot more than that going on and I think you've pretty generally mischaracterized the main problem with the mid-80's era Apple—which had nothing to do with domestic manufacturing and everything to do with not delivering new products that people wanted, at a reasonable price. You can claim overseas manufacturing solved the pricing component of that, but that's not at all clear: other companies were manufacturing in the US at the time and still out-competing Apple.
I don't know about Microsoft, but I'm very clear that the "miracle" operated by Apple was exactly to perfect foreign supply chain at a time when Intel/Dell/HP and others were still heavily focused on the US. The quality of Apple products was already there since the beginning, but they had no way to compete with the PC market until they figured out Asian supply chains.
Anyone keeping count of how many trillions in hypothetical investments and millions of jobs large American corporations have promised in the next 3-5 years?
> Does this mean moving to US based fabs for the M-series chips?
This is not really a practical option. A big part of the M-series success is TSMC's lead in cutting edge process nodes. And Taiwan does not allow export of technology for the latest nodes. It is available only there.
* What type of jobs? - "The 20,000 additional jobs, Apple said, will focus on research and development, silicon engineering and AI."
* Does the US have the required people, in terms of numbers and skills? - "The company is opening up what it calls a manufacturing academy in Detroit, where it will help smaller companies with manufacturing. It already operates an academy for app developers in the city. It’s also doubling its manufacturing fund in the US to $10 billion." - Sounds like they are upskilling, and will count the employees of companies joining the academy as "jobs created"
* Does this mean moving to US based fabs for the M-series chips? - "[M-Series] chips themselves, however, continue to be produced in Taiwan.
* Is this actually profitable, or is this just a political move? - Define profitable. It is cheaper than paying tariffs.
I might be reading it wrong, but that's the 20,000 ADDITIONAL jobs, which is going to be R&D, engineering and "AI".
Those 20,000 people won't be staffing the production lines. So how many manufacturing jobs, especially low skill, entry level with decent pay, will this create? The whole thing is framed in a way that makes it sound like Apple is creating thousands of manufacturing jobs.
> Does the US have the required people, in terms of numbers and skills?
For 30 years, IT managers at blue chip US corporations have exploited the H1-B visa program by saying, "No," and then hiring a never-ending stream of barely-capable Java coders from programmer mills in India, take 5 times longer to make an app than it should have taken, get promoted, and leave everyone holding the bag with shitty web app that we all hate because it's too slow, too bloated, and doesn't work like it needs to. And the companies who can't get enough of that bullshit in-house just hire it out to sub-sub-contractors that do the same thing. Can we not invest in our native population and education systems this time around? I'm so tired of the fact that 90% of the IT staff in my Fortune 250 is Indian, and I know people who would be better at their jobs living in my home town. It hurts our community and our country, in the long run, and by the VERY same logic as re-homing our chip production.
It sounds like you should be directing more of your anger to the C-suite than the people they’re hiring. If they couldn’t get even cheaper Indian immigrants you’d be complaining about code boot camp hires instead - what you need is a tech union which would give you the ability to push back against short-sighted decisions which make your life worse cleaning up messes.
Well, those Indians living in the US will have families of their own, and over time become part of the community you claim to be a part of. Very much like your ancestors did, except they likely didn't face the arbitrary constraints on immigration that Indians (and any other nationality) face today.
The same thing that happened in the UK will happen in America.
People in the UK who are against immigration are often talking about Poles who moved to the UK after the EU and not Indian families who have lived in those neighborhoods for generations.
The crazy thing is, it's not that long ago that Irish and Italian immigrants were not discriminated against. They didn't even consider Italian immigrants to be white.
It will actually happen because it’s nothing new. The 500b is almost all wages for existing US-based employees. They are looking for a carve out from the new China tariffs (same as last time). Note - they made a very similar announcement 4 years ago https://www.apple.com/newsroom/2021/04/apple-commits-430-bil...
Thanks to ireland all big US corporations saved hundreds of billions of dollars past few years so now they can get back to US with this massive cash for anything they want (ofc. nothing will get back to EU as long as they ignore tax heavens)
The US is a great place to have your headquarters and a terrible place to have your not-so-cheap labor.
Their actions will drive prices higher and, indirectly, wage higher. Businesses without a war chest will not be able to keep going and fold, the labor market will collapse.
The rest of the world will trade amongst each other and I suspect to save themselves, some big tech companies will relocate their headquarters.
That could also lead to a depression. I haven't heard a lot of politicians here (Andrew Yang in 2020? does he even count as "a politician"?) with good plans for what to do when automation hits jobs even harder.
“We grew from 10 customers to 100 customers in a year. At this rate we will have 20% of the world’s population in a decade!!!”
The first cohort of customers of any company is always the easiest to obtain with the lowest acquisition cost. You solve the easiest problems first.
This is Cohort Analysis 101. Not to mention Waymo still hasn’t shown to be able to operate in less than ideal weather conditions or proven that the unit economics will make sense or be economical especially taking into account maintenance, or utilization ratios.
It’s been operating safely in each market they’re in. The AI keeps getting better. They have no competition (please don’t bother mentioning Tesla vapor ware). Path to high growth seems pretty sure at this point.
Which is very very few markets, and all of them share weather patterns that are very similar.
When Waymo can demonstrate reliably going from Chicago to Ann Harbor in the middle of a snow storm thats when we can start talking about how its good enough.
And the markets they are in are low hanging fruit with good weather. I’m not saying Waymo is less safe than human drivers. I am saying that it will only take one fatal accident by any self driving car for people to lose confidence, investigations to start, rollouts to be paused etc. I’m also not saying that is a logical response.
> I am saying that it will only take one fatal accident by any self driving car for people to lose confidence, investigations to start, rollouts to be paused etc.
Uber and Cruise are both great examples of this, but it seems like the effect is mostly localized to the company itself that has the issue.
Uber hit and killed a jaywalking pedestrian, resulting in their self driving tech being sold to Aurora. [0]
Cruise hit a pedestrian that was flung into the cars path that a human driver hit previously. This resulted in GM completely abandoning Cruise and their future seems foggy at best. [1]
Google also had literally hundreds of failures and Android is not an amazing financial success by any means and Google still ends up paying Apple over $20 billion a year because people with money buy iPhones.
Google is not exactly known for its success rate getting products out of the door that aren’t ad related.
In the phone market. The Motorola acquisition was a major failure and Pixels aren’t taking the world by storm.
The entire “Other bets” haven’t led to any major successes.
There are only two tech companies that have shown any ability to do hardware at scale as mass consumer products in the last 25 years - Apple and Tesla.
This is all intended policy to benefit Trumps super donors. They can then scoop up marketshare and competition for pennies, then lobby to get the tariffs lowered or removed, but the higher prices - that we will be used to paying at the point this all comes together - will not go down.
All else being equal, companies are going to use the source of labor that results in the cheapest product they can produce. No one is forcing companies to move this kind of manufacturing to the United States. A 10% (let me reiterate that: TEN PERCENT) tariff on incoming goods is inflationary, but by very little, and quickly absorbed by companies and consumers. No one is moving their labor supply from China to the US to avoid a ten percent tariff; US labor is more expensive than that, and there are fifty other places around the planet you could find cheap low-skill labor that aren't on Trump's shitlist.
But you won't believe any of that, because you want all this to happen. You're a doomer; doomers and preppers secretly want the doom they predict to happen, even if they won't admit it to themselves.
Your "logic" (masking) conveniently avoids the point I raised: That these tariffs are being enforced to the tune of 10%. That isn't enough to alone justify this level of investment, or relocation of significant production capacity. Obviously, Apple agrees with this, because the investments they're making aren't as far as I can tell a relocation of capacity from China to the United States, but rather greenfield investment in high-skill research and development. Apple has also made significant investment into advanced silicon manufacturing in the United States; something they also did not rely on China for previously.
> The world is a bigger market than USA and just about every other country has cheaper labor and no tariffs.
Have you done zero research into this? The EU imposes a tariff on Chinese EVs. India imposes insane tariffs on all imported electronics. China tariffs Australian wine. Russia tariffs agricultural products from the EU. Brazil tariffs all imported automobiles. The list goes on. Tariffs are everywhere, everyone uses them for something.
Until things are actually built, I take press releases like these with a grain of salt. Similar to the stories about Mark Zuckerberg removing tampons from men's washrooms the week before the Presidential inauguration, I believe that a lot of these stories are intended for an audience of one.
Deepseek seems to prove there's no super secret sauce that makes these models irreproducible outside the US and that the companies here are suffering a bit from the glut of cash/credits leading them to burn tons of extra processing power that could have been optimized away.
A nice model, does not a billion dollar company make. The hard part of AI is not the model; Apple needs people to do the 80% rest-of-the-work; how do you make AI useful to the average person? How can we get inference on edge devices as cheap and efficient as possible? Models are boring. Everyone fully expects that we'll see an N% increase in intelligence every six months now. Yawn. The exciting thing now is: What are we using AI for?
Spending a couple billion dollars also doesn't make an actual billion dollar company. It's yet to be proven that all this spend on LLM training and running can actually get translated into an actual profit.
1: Take over Taiwan to get TSML under their control
2: Find a way to make a deal with ASML for the needed lithography machines
3: Somehow aquire the knowhow that Nvidia has
Taiwan would require a conflict with the US. ASML is a dutch company but seems to be somehow under US control. I have not yet figured out the exact setup there. And Nvidia is a US company.
They are trying, and will eventually catch up, the same way they have in software and in hardware in many other spaces. Maybe it will take another 10 years. Maybe another 10 months.
This idea that EUVL is—and always will be—outside of the reach of China is, frankly, silly. It's a silly strategy to maintain dominance. They will straight up steal the technology if needed.
We should stop pretending like we can roadblock the technological development of the largest country. It's just going to make the fall that much harder. Once they do attain the ability to manufactuer EUVL domestically, capital is going to flood out of US tech stocks like no tomorrow.
You haven't seen anything that hints at it, because most news about what happens in China doesn't leave China. The propoganda machine is hard at work reminding you that China is terrible and eveyrone in China is poor. You'll learn about China catching up only after it's already happened.
With the combination of very real geopolitical risk (which was a topic of considerable discussion at a tech conference I attended late last fall) and the current political climate, there's a significant mindset that the US should be pulling back a lot of things to its own borders where practical even if not optimal at the moment.
Tesla makes a lot of promises that it can’t keep and losing money and market share globally.
I doubt Waymo is going to be a big deal in much of the US over the next decade. Even if they do figure out all of the technical issues. People will accept hundreds dying from car crashes. But not one dying from autonomous cars.
It won’t happen. The supply chain is far too complex. Not to mention that the labor market in the US is not willing to do a lot of the work that you see in China and isn’t large enough even if there were enough willing people.
And then you have the rare earth minerals that aren’t available here.
I don't get why folks keep saying x86 linux servers here for AI, if anything it'd be M series arm based servers, either running linux or macos. Realistically I'd imagine a set of scaled up mac mini arm servers for running inference or fine tuning on them as more likely being the "ai servers" than x86 based anything. Power is the key thing that they'll be optimizing for, and that's where ARM shines.
Don't they need gpus (for training)? Apple already did a footshoot wrt gpus in the apple ecosystem. unless they have some sort of apple-internal ai chips ready to train models.
Apple's Private Cloud Compute is on racks of M4 chips which have NPUs and GPUs on-die and unified memory access to however much RAM they want to put on them. All of a sudden they're competitive with NVIDIA, but they don't let anyone else use that platform.
Nvidia has proven the space is incredibly lucrative and Apple is best equipped for high end chip designs. Remember 10 years ago it was unthinkable for an ARM chip to compete with x86.
First Apple has to prove they have competitive designs. Apple Silicon GPUs simply do not compete with the efficiency of Nvidia's GPU compute architecture: https://browser.geekbench.com/opencl-benchmarks
Apple's obsessive focus with raster efficiency really shot their GPU designs in the foot. It will be interesting to see if they adopt Nvidia-style designs or spend more time trying to force NPU hardware to work.
I think performance per watt is way in Apple's favor, but raw performance is not.
That said, an M4 Ultra (extrapolating from Max and Pro) would likely compete with my 3090, and with 192GB of memory (for 10x the amount it should cost) will out perform my 3x3090 AI server. And honestly, cost less than my 3 3090s + rest of the computer + electricity.
It won't outperform a bunch of A/H 100s (or even a single one, or any other cards in the enterprise realm) though, but it will cost an order of magnitude less than a single card.
Careful when comparing performance and efficiency. As a rough factor power increases quadratically as you increase clocks on a design, so you can quite easily make a high performance design low power by under-clocking it. The same is not true for the reverse.
That's not necessarily news, unless I am missing something. Craig made an indirect mention of this during last year's WWDC regarding the private cloud compute.
Does this mean competing with Asahi to run a Linux kernel, or will this be an attempt to run AI workloads on XNU?
Consider the cost of GPUs, losing what could be double digit percents on overhead might not make this very competitive. The macOS microkernel can still beat NT in some situations (like not having filters slowing filesystem down to a crawl), but it lags significantly behind the investment in Linux performance over the years by every other major company.
I believe the intention is to use their own M-Series CPUs - to get what they call "Private Cloud Compute". The cpu on your phone will encrypt data and a request, send it over the network to am M-series CPU which will decrypt and process/send back an encrypted response.
The idea being there's no VMware, kernel or piece of hardware that can have backdoors built into unless someone files off the top of the chip and somehow probes the silicon
> Apple said that it, together with Foxconn Technology Group, will later this year begin producing the servers that power the cloud component of Apple Intelligence — a system called Private Cloud Compute — in Houston. That marks a relocation, at least for some production, from overseas. Next year, it says a 250,000-square-foot facility for such manufacturing will open in the city.
> The Private Cloud Compute servers use advanced M-series chips already found in the company’s Mac computers. Those chips themselves, however, continue to be produced in Taiwan.
I don't think it's AI servers for Apple silicon. I think it's just regular x86 Linux servers to power Apple's AI cloud services. It's a commitment to internal investment rather than a product.
Apple doesn't like B2B and Steve Jobs was very vocal on this (there are various videos where he explains why). Ever since they can afford to, they reduced their B2B to the bare minimum needed. So don't expect anything server-like from Apple.
I'm not sure. Because the moment you enter the system room, the ecosystem is a completely different universe.
You can't easily sell "Good / Better / Best" version of a single model, and tell "These are the options, take it or leave it". Servers are customized to the screws they come with and are expanded throughout over the years. So, the logistics are somewhat different for these kinds of devices.
Plus, macOS is not a CLI first operating system for server operations, and macOS Server is not updated for some years. Allowing Linux would be a different offering, and allowing macOS to work with all kinds of hardware from ordinary Ethernet to 100G+ Ethernet and 400gbps Infiniband (plus all the other interconnects) will be a fun exercise in testing flexibility of both macOS and Apple development teams.
So it's quite complicated. All servers are built to order SKUs. Dell keeps configurations "per server" in their databases, for example. If you have a Dell server, enter its service tag to support site, and you'll get the configuration of the device as it left the factory.
They don't have to go all the way though and fully compete with Dell EMC/HPE. I'm not sure what the original commenter was thinking but in my mind they could simply sell a Mac Studio variant with dual PSUs, better networking, a rackmount chassis, etc. Basically have their existing consumer machine placed in a more datacenter friendly factor.
I mean places like Github and AWS are painfully racking up Mac Minis for their deployments and this theoretical server model would simplify everything. It also becomes an option for on premises AI inference using MLX, especially if they manage to get ANE support working in conjunction with the GPU for faster prefill.
The support and software stack for the server model would be the exact same as the consumer variant and they certainly wouldn't have special Linux offerings, Infiniband, and all that. If there's networking beyond their existing 10G it's going to be built into the board and they aren't going to support random 3rd party cards. The unit also doesn't need to be upgradable either.
"Four years ago, a few months after President Joseph R. Biden Jr.’s inauguration, Apple announced an “acceleration” of its U.S. investments, pledging to spend $430 billion and add 20,000 jobs over five years. In January 2018, during Mr. Trump’s first term, the company said that its “direct contribution to the U.S. economy” would be $350 billion over five years and that it planned to create 20,000 jobs over that period.
Apple did not immediately respond to a request for comment."
Is this just lip service? What happened to those previous investments?
This happens all the time with large investments in $X that make for impressive sounding press releases. If you (can) dig into the details, invariably a lot of the money is in previously committed/spent allocations in a whole bunch of different buckets (or, per bombcar's sibling comment, money that may never be spent at all).
I can't tell why you're being downvoted, we have quite a laundry list of protracted investments at Apple leading to nothing worth shipping. Given how much of a nothingburger AI has been up to this point, I have zero hope that Apple will succeed at commoditizing a zero-demand service.
This is an insane amount of cash to throw at any problem. They could have Apple rockets mining asteroids with $500,000,000,000. There is no way all this cash goes into AI. What will actually happen is they will take 1/10th this cash to an over-valued startup and acquire them.
Between land, hardware, having to build multiple power plants, the labor costs involved in all of this, and setting money aside to run all of those for however long, then yeah I can see where it gets up to that price. 20k engineers is easily 5 billion per year in salaries, probably more.
Assume the human resources at $100,000 per head, and you get $2B/yr. Four years comes to just $8B for human resources. Assume land costs $10B. Assume construction costs $100B.
> Four years ago, a few months after Mr. Biden’s inauguration, Apple announced an “acceleration” of its U.S. investments, pledging to spend $430 billion and add 20,000 jobs over five years. In January 2018, during Mr. Trump’s first term, the company said that its “direct contribution to the U.S. economy” would be $350 billion over five years and that it planned to create 20,000 jobs over that period
Anyways, the land (obviously in Texas) is already purchased and has been sitting empty. The unbuilt factory keeps getting more expensive though.
Do we think Apple will once again sell servers to customers?
I guess they could sell servers to customers who want to run the latest Apple Intelligence models on-prem, even though that probably wouldn't make much of a difference, since you probably still have to trust Apple.
> Apple said that it, together with Foxconn Technology Group, will later this year begin producing the servers that power the cloud component of Apple Intelligence — a system called Private Cloud Compute — in Houston. That marks a relocation, at least for some production, from overseas. Next year, it says a 250,000-square-foot facility for such manufacturing will open in the city.
[...]
> Apple will also expand data center capacity in Arizona, Oregon, Iowa, Nevada and North Carolina, all states with existing Apple capacity. The company confirmed that mass production of chips started at a Taiwan Semiconductor Manufacturing Co. facility in Arizona last month. Bloomberg News recently reported that plant is building chips for some Apple Watches and iPads.
I think they might build a cloud offering. Something like Cloudflare workers but AI centric, perhaps running Swift on the Apple equivalent of V8 isolates.
Makes sense from a business perspective - there's significant growth potential for them as their presence in web tech is approximately nil.
It would make sense for Apple to fork their next highest-end Mac Studio motherboard, make relatively minor changes to it (e.g. add a higher bandwidth NIC and strip out unnecessary I/O) then wrap multiples of those into a rack mount chassis, with commodity-grade cooling and power supply solutions appropriate for the context.
Combined with a properly headless fork of their OS stack (think Darwin, not OS X Server) they could spin up a highly competitive solution using entirely "B-team" resources.
...then it would be piped through their design-council, run through 5 more iterations to get a unique unibody case for it, accompanied by an optional proprietary Apple rack and a price-tag triple of the competition.
That's along the lines of how it usually rolls whenever Apple tries to make something purely utilitarian, it's the most considerate and "fresh" look at a product, but ultimately designed to be used and then disposed when finished.
A purely utilitarian IT-appliance without a individual end-user doesn't seem to be possible in their product pipeline, you usually end up with something "Prosumer": Impressive on its own, yet of degraded maintainability and scalability.
It's like asking Bugatti to design a public transport bus. It would surely be an impressive bus, but not one you would want to maintain over years at a scale of hundreds.
Oh god. Hopefully they’re lighter than the old xserves. We had one still running up until a few years ago when we finally removed it. You could break a toe if it dropped while pulling it out of the rack. People are still selling them on eBay.
I suspect that was the RAID drive bay (ridiculous item). That was a 3-4U monster.
The Xserve, itself, was a 1U unit that was pretty much the same (or lighter than) any other 1U server (we also had HP and Dell servers that were heavy). The weight distribution could be weird.
That stupid drive bay was a proprietary nightmare. The disks cost a fortune.
However, from what I can see, this will be for "internal-use" servers. I don't think they will be selling iron; just services run on the iron.
I don't think it's AI servers for Apple to sell. I think it's just regular x86 Linux servers to power Apple's AI cloud services. It's a commitment to internal investment rather than a product.
They have hiring positions for running a Darwin-based server OS, and their private cloud compute is on Apple Silicon. I doubt it's going to be swaths of x86.
I'm curious what they will look like, given that these are not for anyone else to buy. Maybe Apple made a different form factor/configuration that suits their datacenters better?
expect of course they won’t… few months into every new administration (see 2017, 2021…) they’ll make a splash announcement like this… and then wait for the next administration to make it again
A strong majority of Apple's revenue comes from outside the US. If Apple is tied to US protectionism, it is precisely the sort of org that suffers.
Now to be real, Apple has been announcing these multiple hundred billion dollar investments every four years like clockwork. They did it early in Trump's first term, they did it again early in Biden's term, now they're doing it again. But for all of the "Yeah, this is what tariffs are for! Hoo ya!" rhetoric, note that companies like Apple have far, far more to lose than what America "gains" by acts like this.
Texas is big tech's choice to skirt employee protections. I'm sure these are the type of jobs, similar to Foxconn per the article, that Americans are looking forward to.
I know a lot of tech workers in Texas, specifically in the Austin area. They seem to be doing very well. I'm quite proud of America's working conditions. A lot of workers in other countries would marvel at our opportunities and be grateful these investments are happening here as opposed to elsewhere.
California's housing crisis is a result of small-c conservatives wanting their property values to rise forever. Prop 13 and it's consequences have been a disaster for the state.
While that’s certainly a key component, Texas is also home to the largest potential solar and wind capacity in the country. There’s also a ton of land to build on.
Unlikely. Texas is the largest exporter of crude oil and natural gas, the largest in capacity for refining petroleum, and huge exporter of petrochemicals in the US. More solar and wind means more oil for refinement and export.
I do find it slightly offensive that you would insinuate that hiring in Texas is solely about less worker protections and not that we have plenty of skilled workers in one place and with a lower cost of living.
I'm a worker not an owner,and I prefer living in Texas at this stage in my life and have turned down offers to move back to California.
Working in tech in a big Texas city easily puts you in the top 10% of the cushiest jobs in the US. Based on how I've been treated here, I really doubt that worker protections (or lack thereof) is the real driving force behind more tech jobs moving here. We are far from being oppressed here.
The most likely driving force of tech moving to Texas is that mid career professionals like myself don't see a future in California due to the insane cost and bad vibe of raising a family there. It's a great place for people just out of college, but Texas is a better place to settle down unless you are pulling an outrageous salary. The other big advantage California has is VC and startup networks being located there, which is also something that primarily benefits early career people rather than those of us that need a stable job at an established company.
It's also worth pointing out that Texas has long had a large technology industry presence. The dominance that California experienced during the early 2000s through to mid 2010s is an outlier and it shouldnt surprise anybody that things are evening out.
IMO we've really got to start pressuring our own governments to take control of their networks, as well as the companies the population is going through (not just Facebook, but even international contractors for services). Letting a foreign government have this much control over the data of your populace and the ability to feed whatever algorithmic message they like is a path bound for disaster in the long run. The powers of the world are way too consolidated as is, and a company can turn into a state actor at the drop of a hat. I don't think we can maintain kayfabe about the country/corporate divide. I also think this can be done without impacting freedom of speech for your population, as long as you don't consider corporations people.
Most countries don't have the resources to do much, but even then they can try their hardest not to be beholden to any single foreign country coughChina, Americacough.
There's a lot of understandable skepticism about this announcement because there are so many PR announcements. I want to temper that with an alternative perspective.
I actually think Apple is a potential dark horse when it comes to AI hardware. What we have now is essentially a three-layered monopoly: ASML, TSMC, NVidia. This has been incredibly lucrative for NVidia. But, as we know, Apple doesn't like to rely on third-party hardware. They've invested heavily in ARM going back to buying PA Semi [1]. Apple replaced Intel chips (which originally replaced Power chips) with the M series in recent years. Apple is in the process of replacing Qualcomm modems in iPhones, which is not only a technical feat but a legal one given Qualcomm's patent dominance over 4G/5G.
Apple has the resolve for long-term initiatives that few other companies have. Apple Pay continues to chip away and get slowly better in a way that, if it were a Google product, would've been cancelled, rebranded, relaunched probably 3-4 times by now (Google Checkout, Google Wallet, Google Pay, Android Pay, etc).
Apple clearly sees AI as a strategic issue. They have loads of cash on hand to finance basically anything they want. And they won't want to be beholden to NVidia.
I expect Apple to have a significant impact here but it won't be tomorrow or even this year. It'll be over the next 5-10 years.
Apple's incentives have definitely aligned with replacing Nvidia entirely ever since they ceased diplomatic relations. But Nvidia also knows this, which is why they invest heavily in things Apple will never do. They write the official Linux drivers Apple wouldn't get caught dead supporting. They give users and integrators freedom to choose their OS, software and library stacks to better suit their application. They sell individual GPUs and unlocked edge compute hardware with no distribution terms or $99/year "developer license" bullshit. Nvidia is a hardware company in places where Apple tries shipping services instead.
Then there's also the software issues. Nvidia has invested in GPU-based compute nonstop for the past 10+ years. Apple invested in Nvidia, then invested in OpenCL after abandoning Nvidia, then abandoned OpenCL for Metal compute which would eventually become the proprietary Accelerate framework. Nvidia's eggs are all organized in one, valuable basket. Apple's investments are spread out all over the place, with much of the time and money going into projects that don't even exist anymore.
Apple has the TSMC advantage, but that's just about it. Their GPU designs aren't comparably efficient or compute-oriented to what Nvidia ships today. Additionally, Nvidia will continue investing in places that Apple principally refuses to support. Unless a serious tide change occurs at Apple, they aren't going to get a fair competition with Nvidia.
Some part of me thinks they are billing an overbudget here to report that they actually didn't need to spend that much so that they can beat lower guidance (same play for MS, Google, Meta). We've heard that actually training these models doesn't cost even a billion dollars.
There's a lot of noise I can see behind the scenes on investor confidence. Noise as in "everything is fucked" sort of level of noise. Thus I expect this is being said to try and stop the AAPL stock collapsing in the upcoming recession that the analysts are predicting more than a tangible expansion and recruitment goal.
I also take issue with their being 20,000 people on the market who are still able to contribute something useful. They will be culled quickly and quietly down the line in the annual corporate lay offs.
It is not the time to make grand gestures unless you're trying to gain political favour, at which point any respect I have at least is gone.
Reverse technology transfer from countries like China is kind of fair. But EU companies should be very wary of Trump tariff blackmail that forces them to build production lines in the US.
Not a lot of them, but Trump accuse EU of tariffs because it has VAT. That’s not the same thing because VAT apply to domestic companies as well so there’s no real unfair advantage.
You could make the claim that it’s still worse for foreign goods as VAT can be offset with VAT from supplier purchases, but that only works if you produce goods in the same country as you sell them. But US companies have even larger advantage already as they don’t have any VAT, so in the end the playing field is kind of level.
This doesn’t stop Trump from pressuring and squeezing and making false claims to justify his standing point.
Expect an unstable trade environment due to tariffs and retaliation. This will hurt small businesses that have even more complex environment to operate in.
>But EU companies should be very wary of Trump tariff blackmail that forces them to build production lines in the US.
That's the whole point of tariffs, to encourage domestic production.
Put another way, what is the difference between what you wrote and
>But US companies should be very wary of EU tariff blackmail that forces them to build production lines in the EU.
?
Whether a particular tariff is economically viable is a reasonable debate. Calling Trump's tariffs "blackmail" without assigning that epithet to all tariffs from whatever source is not.
Yeah, the real stock market mover is the mouse cursor. That wiggly thing they did that grows the mouse cursor really send ripples throughout Wall Street.
He wasn’t stupid. He’d observe their own developers and see how they rely on the terminal and command line for their work. He’d ask pointed questions and demand thoughtful answers.
Then he’d find a way to make it the #1 AI developer platform or distort reality until it is.
Or you can take the extra minute to install them yourself with brew, this a complete non issue for anyone that understands the command line in the 21st century using MacOS. Also, I would never build anything against macos userland because it's almost never the target.
So what? Do you complain when apt, yum, dnf, pacman, ports, or any other package management system does a download? I bet you don't, so it's not really a usable argument. Secondly, yeah, not tainting my systems OS and system paths is a good thing and opt/ from the filesystem perspective is the absolute right place to put add on packages.
"The /opt/ directory is normally reserved for software and add-on packages that are not part of the default installation"
Desktop Linux is right here, desktops Linux has always been here!
I don't understand why people complain about Apple neglecting developers when desktop Linux provides a superior experience aside from the rare times you need to compile an iOS/Mac specific application.
Edit: In case it's not clear from my initial, gut-driven snark: I definitely think if you use a reasonably popular distro (commercially backed or not) in 2025, you should never have any trouble connecting peripherals to it, with the possible exception of Bluetooth, which I hear also applies to macOS.
I'm on Windows, and my work laptop has no audio via its headphone jack.
I have no idea how this is supposed to even work, and since it's not my computer I don't mess around trying to install drivers. I just use my phone to call in for Teams.
Things happen, let's not act like any OS is perfect.
MacOS is definitely not perfect. I'm being snarky. But it has been my anecdotal experience as both a user and observing colleagues that MacOS is more reliable and stable for desktop use than Linux. This is unsurprising since it's easier to build a stable walled garden than an open ecosystem.
Macs are generally more reliable, but if you buy a year old ThinkPad Linux will be just as stable .
The only issue Linux really has is when new chipsets come out you might need to wait 6 months or so for the drivers to be updated. But to be completely fair, on one of my laptops I had no webcam support for like six or seven months until Windows update decided to finally install it for me.
If you need a significant amount of hard drive space, Macs are almost always exorbitantly expensive. I make music so I find myself dual booting between windows and Linux. I don't want to speed 3k+ on a MacBook just to get a 4TB SSD I can add to any Windows PC for 200$.
Plus on Linux you can customize your personal experience to a much greater level. If you dislike X,Y,Z you can disable it or find an alternative.
Both OSX and Windows are cramming so much monetization into the OS, there's a very real feeling that I'm just sharing my computer with a giant corporation rather than actually owning it.
It’s less convenient when you are on the go, but you can pack an external SSD and offload stuff to it. A friend of mine had one velcroed to the back of the screen.
It's actually cheaper to own a MacBook Air for things that need to work 100%, like a coding interview, and then a secondary laptop when you're playing video games or making music .
That's basically what I do now, my old M1 MacBook air is more than good enough for LeetCode and I'm more or less know it's never going to fail.
I haven't personally experienced that problem. Updates on Mac have always been smooth for me. But I'm a sample of one and it's probably workflow dependent.
Last year I was using a Windows laptop for work and Teams was very unreliable with audio and video. And don’t even think of using the nice camera on top of the expensive video conferencing monitor on my desk.
Presumably these users have audio in other contexts? Are they running the web app version of teams? Do other web apps play audio? From 10000 feet up, I wouldn’t start by blaming Linux here (even as a non-Linux-desktop user,)
This isn't a hill I want to die on, but isn't it the case that even if the problem is in MS software compatibility with Linux that still results in desktop Linux being a less reliable platform for day to day use?
Tons of peripheral devices do not work well or reliably on Linux, and I literally cannot remember the last time I have had ANY issue with Bluetooth on macOS. Certainly not in the last decade.
The last time I did was this morning. I get dropped connections constantly, microphone not working in Teams (solved by reboot), pegged connections preventing handoff, etc.
I went through 5 distros a month ago dealing with fractional scaling issues on my 4K monitors. Decided it is not worth dealing with a went back to macOS so… No.
I very recently tried again to adopt Linux on the desktop. I'm really sick of feeling like a frog in a pot of water. It's becoming harder and harder to bypass their literal gatekeeping of which applications I can run on my computer, and with every new version of macOS the temperature in the pot keeps rising.
The main problem I have with living in a Gnome desktop environment, is with the keyboard. I'm not willing to abandon my use of Emacs control+meta sequences for cursor and editing movements everywhere in the GUI. On macOS, this works because the command (super/Win on Linux/Windows) key is used for common shortcuts and the control key is free for editing shortcuts.
I spent a day or so hacking around with kanata[0], which is a kernel level keyboard remapping tool, that lets you define keyboard mapping layers in a similar way you might with QMK firmware. When I press the 'super/win/cmd' it activates a layer which maps certain sequences to their control equivalents, so I can create tabs, close windows, copy and paste (and many more) like my macOS muscle memory wants to do. Other super key sequences (like Super-L for lock desktop or Super-Tab for window cycling) are unchanged. Furthermore, when I hit the control or meta/alt/option key, it activates a layer where Emacs editing keys are emulated using the Gnome equivalents. For example, C-a and C-e are mapped to home/end, etc.
The only problem is, this is not the behavior I want in terminals or in GNU/Emacs itself. So I installed a Gnome shell extension[1] that exports information about the active window state to a DBUS endpoint. That let me write a small python daemon (managed by a systemd user service) which wakes up whenever the active window changes. Based on this info, I send a message to the TCP server that kanata (also managed by a systemd user service) provides for remote control to switch to the appropriate layer.
After doing this, and tweaking my Gnome setup for another day or so, I am just as comfortable on my Linux machine as I am on my Mac. My main applications are Emacs, Firefox, Mattermost, Slack, ChatGPT, Discord, Kitty, and Steam. My Linux box was previously my Windows gaming box (don't get me started about frog boiling on Windows) and I'm amazed that I can play all my favorite titles (Manor Lords, Hell Let Loose, Foundation) on Linux with Proton.
I don't get it either, it takes twenty minutes to burn a USB stick and run the installer. It takes me longer to remove the bloatware and set up my preferred settings on a proprietary OS than it does to install Linux nowadays, and that's been true for a decade at least now.
It's just not hard! It's not more work! And yet the meme about it being more trouble just. won't. die.
You people are supposed to be technologists! Why won't you spend 20 minutes of one time setup to get a better experience?
Not all hardware works? I don't see anybody complaining about having a limited set of hardware options when they buy Apple! Canonical maintains a list of fully compatible computers; just pick one, buy it, and you wind up with a computer just as easy to use as Mac OS but without the endless paper cuts of using a system that has no respect for you at all and thinks it knows better
Most of us have .dotfiles, I can snap any macos installation into my preferred configuration in about 5-10 minutes unattended depending on internet speeds. I do most of my work in a terminal, as long as that works, I'm good on Linux, MacOS, and BSD's. They all have pros and cons.
I use Apple laptops primarily for the hardware. But Linux has never really been a great experience on Mac laptops when it comes to battery life, reliable suspend/resume, etc. etc. I used to use Yellow Dog Linux on a G4 PowerBook way back in the day, but I haven’t had much luck with Linux on Mac hardware since then.
I had an old i5 Mac mini laying about I wanted to use desktop Linux on the other day. The last time I tried, was about 20 years ago. I note nothing has changed since.
I made this. It makes it easy to use all of the most common gnu tools via brew, without having to do gsed for sed, etc... all with working man pages. It also lets you switch back easily in a shell session if you need the mac native ones for some godforsaken reason:
So are the BSD tools by some definition of "behind". Another way to look at this is to say that GNU tools as typically seen in modern Linux are bloated (I know, Linux and "bloat" are kind of a meme, but it is generically true for the most part when it comes to the command line utilities feature creep over the years, so it's a useful and descriptive word).
I have to work with old machines and legacy operating systems quite a bit in my day to day and I always am going to prefer something lighter and with less ways to shoot myself in the foot w.r.t. POSIX compliance. MacOS is Unix certified so I appreciate them being somewhat reserved in the features they add on top of POSIX.
Modern GNU userland utils are nice and fun but if you are looking for compatibility it's best not to use them. Consequently, the MacOS situation doesn't bother me especially given you can install more up to date tools if you want. I think keeping the defaults older and more compatible is a good thing.
There is poor cross-UNIX compatibility if you're doing anything complicated, anyway. I maintained a large test suite for about a year that was written in POSIX sh and targeted Linux, macos, {Free,Open,Net}BSD. It wasn't fun because every program behaved in slightly different ways, half of them undocumented (for example, I remember having lots of pain with how different versions of tail handled SIGPIPE).
In the end it was was easier to rewrite in Perl than to keep maintaining that thing, struggling for hours to find ways of implementing every little bit of functionality that worked reliably on every OS. You'd add or fix something, and the tests would break on FreeBSD. You would fix it there and it would stop working on NetBSD. And so it goes.
I prefer a lot of the BSD variants of the typical POSIX tools (i.e. bsdtar vs GNU tar, ksh or similar instead of bash, etc etc). Usually because they add less extensions on top of what is required by POSIX, but are still easily acquired in a modern Linux distribution. I mostly just alias them.
If I write a script using BSD esque tools I can be reasonably sure they will work on any Unix-like, whereas if I write/test my script on a machine using GNU utils, I'm fairly likely to accidentally use a GNU extension that would cause the script to fail on an older Unix-like OS. For instance, I do a lot of work migrating code off of AIX,and I need the scripts I develop to work on AIX when I'm gathering environment information from customers. I can't just assume they will have a ~2020+ implementation of Unix userland tools with all the GNU extensions and nice features. Sometimes the machines have been sitting quietly in the back of a data center not being updated for quite a while and will have more "90s style" of Unix tools.
It’s like having the tools in a different computer. You can mount your local filesystem onto the container, but if feels like WSL - there’s always an “impedance mismatch” between the two sides.
I prefer to use the tools running locally on the same OS I’m working with. For that, MacPorts is great.
Funny I turned Siri off because i didn't want apple intelligence running amok. The follow-on problem --> lack of Siri killed my Carplay because Siri is required (also use itf for setting alarms/timers). The kicker? I can't seem to turn Siri back on after look through all the menus.
I.e. My preference for apple CarPlay supersedes my concerns on GPT running over my contents. Though the UI/UX has made it next to impossible to turn it back on.
Apple Intelligence and Siri are still separate (though Apple like to make it look like they are fundamentally intertwined). You can turn Apple Intelligence off and leave Siri on for CarPlay.
How did you turn Siri off in the first place? That's where I'd start...
The part of Siri that causes the most trouble is the speech recognition - which uses a voice recognition model that we now colloquially refer to as "AI." The part that works reliably, the part that sets your alarm or sends the message, is an action that's hardcoded.
IMO, moving towards AI just leads to increased uncertainty and undesirable outcomes, which is something several journalists reviewing Apple Intelligence have attested to.
If your phone is new enough for Apple Intelligence, Siri is now under that umbrella. There's no "just Siri" option anymore, unless you're rockin an iPhone 14 or older.
When I put in timers -- for some reason my timer frequently/randomly just sets to 79 hours and a random assortment of minutes and seconds. I have no clue why. I always have to double check otherwise I might be waiting awhile.
It feels like it was a residual timer or something but I have never set anything like that - it is quite strange.
I don't know if this is an actual problem you have, but since Siri appears to be composed of independent voice-to-text and text-to-action systems, you can say "start a one three minute timer".
The problem is AI current best use case is creative work, art, music, programming, but skilled creative professionals is a/the core userbase for Apple products.
Apple is stuck and it’s AI will never be good enough until those creatives embrace it. Right now it’s disdain when mentioned.
AutoCAD came to the Mac when Intel was shitting the bed (with aggressive OEM contracts for first party system integrators that prevented AMD adoption across HP/Dell/Lenovo-lines) and Windows 11 was being forced on users.
WINTEL played the monopoly game too hard and is starting to lose ground.
An oft-cited quote goes something like this: "we wanted robots/AI to automate boring, routine, meaningless jobs to let people be free to pursue arts, music, creativity. It's a sad state of affairs that AI is taking over arts/music/creativity stranding people with boring, routine, meaningless jobs"
When Siri first debuted it would automatically beep, so I could immediately tell if the phone did not recognize recognize "Hey Siri" (just "Siri" didn't work). A couple of iOS updates later this went away, which means I can't tell without actually picking up the phone and looking at it whether the command was accepted.
Even more annoyingly, sometimes there is a beep! ¯\_(ツ)_/¯
Whoa, whoa, whoa, on occasion, if the planets align just right, I can also get Siri to set a reminder (and at least half the time Siri gets it 80% right).
If you haven't tried OpenAI's advanced voice mode, it's a mind blowing version of exactly what things like Siri really ought to become with a little more development. If that's what you mean by LLM Siri, I totally agree.
Being able to chat casually with low latency, correct yourself, switch languages mid-sentence, incorporate context throughout a back-and-forth conversation etc. turns talking to these kinds of systems from a painful chore into something that can actually add value.
It's the other way around. The model is impeccable at "understanding text." It's a gigantic mathematical spreadsheet that quantifies meaning. The model probably "understands" better than any human ever could. Running that backwards into producing new text is where it gets hand-wavy & it becomes unclear if the generative algorithms are really progressing on the same track that humans are on, or just some parallel track that diverges or even terminates early.
Only if you wildly oversimply to the level of being misleading.
The precise mechanism LLMs use for reaching their probability distributions is why they are able to pass most undergraduate level exams, whereas the Markov chain projects I made 15-20 years ago were not.
Even as an intermediary, word2vec had to build a space in which the concept of "gender" exists such that "man" -> "woman" ~= "king" -> "queen".
3 lines? That's still going to be oversimplifed to the point of being wrong, but OK.
Make a bunch of neural nets to recognise every concept, the same way you would make them to recognise numbers or letters in handwiting recognition. Glue them together with more neural nets. Put another on the end to turn concepts back into words.
... Oh interesting. And those concepts are hand picked or generated automatically somehow?
> For a less wrong but still introductory summary that still glosses over stuff, about 1.5 hours of 3blue1brown videos
Sorry, my religion forbids me from watching talking heads. I'll have to live with your summary for now. Until I run into someone who condensed those 1.5 hours in text that takes at most 30 min to read...
> Oh interesting. And those concepts are hand picked or generated automatically somehow?
Fully automated.
> Sorry, my religion forbids me from watching talking heads.
What about professional maths communicators who created their own open sourced python library for creating video content and doesn't even show their face on most videos?
You're unlikely to get a better time-quality trade-off on any maths topic than a 3blue1brown video.
He's the kind of presenter that others try to mimic because he's so good at what he does — you may recognise the visuals from elsewhere because of the library he created[0] in order to visualise the topics he was discussing.
Simplifying to that point is more of what a Markov chain is. LLMs are able to generalize a lot more than that, and it's sufficient to "understand text" on a decent level. Even a relatively small model can take, e.g. even this poorly prompted request:
"The user has requested 'remind me to pay my bills 8 PM tomorrow'. The current date is 2025-02-24. Your available commands are 'set_reminder' (time, description), 'set_alarm' (time), 'send_email' (to, subject, content). Respond with the command and its inputs."
And the most likely response will be what the user wanted.
A Markov chain (only using the probabilities of word orders from sentences in its training set) could never output a command that wasn't stitched together from existing ones (i.e. it would always output a valid command name, but if no one had requested a reminder for a date in 2026 before it was trained, it would never output that year). No amount of documents saying "2026 is the year after 2025" would make a Markov chain understand that fact, but LLMs are able to "understand" that.
I’m confident that LLM’s will not have hallucination problems in the type of requests that I send to Siri.
I don’t ask Siri for facts (just like I don’t ask LLM’s for facts). As long as it can correctly, understand what and when I ask to be reminded about something, that would be a huge improvement for me.
That and being able to map “Bedroom Fan”/“Bedroom Fan Light” to “Bedroom Fan Lights” without having to specify aliases (and even then it hearing me wrong).
I’ve see Home Assistant working with LLMs and it can understand groupings that I never explicitly defined which is very nice. I can say “Turn off all overhead lights” and it will find all my overhead lights and turn them off. Siri/Alexa can’t handle those tasks currently.
Over on Android it's the opposite situation. The voice interface to Google Assistant was very reliable for simple things like reminders and appointments, and even for general knowledge questions. It was part of why I didn't switch to an iPhone. Then Gemini came along, and that core functionality got a lot worse.
...that will grind your request to set email Vacation Mode through the world's worst speech-to-text, jam the text into Chat GPT, and spend the next three minutes reading you an uninterruptible 3 minute essay about violence.
I'm a formerly non-mac guy who finally bought a brand new iPad. I got bored with the wallpaper but couldn't figure out how to change it. "Hey Siri, how do I change the wallpaper?"..."Sorry, I can't help with that". Tried a couple more questions and all it did was Google it for me. This is the latest M4 that was around $2k.
This is what our "AI accelerated" chips give us in return? What a disgrace
I detached this comment from https://news.ycombinator.com/item?id=43158188 so we could pin the latter to the top. What you posted was fine! I just don't want to take up extra real estate at the top of the thread.
We used to call this an "Industrial policy" or an "Economic development policy". Back in the golden era when a strong labor movement coexisted with a Red Scare. 78 years after Taft-Hartley and 44 years since PATCO, not so much.
We have maybe fifty or a hundred million people rotting away in areas where jobs are scarce and housing is plentiful, because we used government policy to shut them out of areas where jobs are plentiful and housing is scarce. We systematically exported jobs from places that aren't big cities because they can be performed overseas and our aristocracy can still profit from them by owning those people overseas.
I don't know if returning to a little more deliberate of an economy is even a partial salve for the place we've found ourselves, but I don't think this laissez faire thing is sustainable for a whole lot longer. We are overleveraged, and arrogantly delusional about our sway at the moment; "Ownership" is not some valuable skill. The fall of an anchor currency and global conversion to an alternate financial network would be a spectacular thing, an astroid striking terrain, which might leave craters on entire other continents from secondary ejecta. World wars have been fought over less.
"We" wasn't big government. It was a million homeowners who decided that the neighborhood they moved into should be frozen in amber forever. Everyone wants housing to be cheap but also for their property values to rise onto infinity. They push back against any attempt to change this and then complain about the inevitable results.
Say what you want about Trump, this is the kind of deal that wouldn't have happened with any of the previous administrations, both Democrat or Republican. It's the kind of deal that keeps MAGA loyal to him, despite all the noise about DOGE.
https://archive.ph/EfwDl
Some useful context: this is almost certainly being driven by Apple’s Private Cloud Compute architecture and not tariffs, as an investment of this magnitude is not planned overnight.
Why is PCC driving Apple to spend billions to build servers in the states? Because it is insane from a security standpoint (insanely awesome).
PCC is an order of magnitude more secure server platform than has ever been deployed for consumer use at planet scale. Secure and private enough to literally send your data and have it processed server side instead of on device without having to trust the host (Apple).[1] Until now the only way to do that was on device. If you sent your data for cloud processing, outside of something exotic like homomorphic encryption[2], you’d still have to trust that the host did a good job protecting your data, using it responsibly, and wasn’t compromised. Not the case with PCC.
To accomplish this Apple uses its own custom chips with Secure Enclaves that provide a trust foundation for the whole system, ultimately cryptographically guaranteeing that the binaries processing your data have been publicly audited by independent security auditors. This is the so called hardware root of trust.
It is essential then that the hardware deployed in data centers has not been physically tampered with. Without that the whole thing falls apart. So Apple has a whole section in their security white paper detailing an audited process for deploying data center hardware and ensuring supply chain integrity.[3]
You can imagine how that is the weak point in the system made more robust by managing it in the US. Tighter supply chain control.
[1] https://security.apple.com/blog/private-cloud-compute/
[2] Fun fact, Apple also just deployed a homomorphic encryption powered search engine! It’s also insane!
[3] https://security.apple.com/documentation/private-cloud-compu...
Trusting Secure Enclaves custom chips over processing locally is going to be a hard to impossible sell for those who truly care about privacy.
Thankfully for Apple that's a very low number in a world where people demand tiktok remain legal when shown how their data is being used by foreign actors. People only care about privacy when it's local (don't want mother to find out, neighbours to talk, friend to think a certain way about you or classmate stalking) and that's why ai fakes are much more concern then a company knowing everything you do.
But this product is great for fortune 500 businesses.
I think this is a level of security Apple is providing at additional cost to themselves that only a tiny fraction of consumers would even pay an extra cent for.
From that perspective I really appreciate this effort by Apple.
Did they basically just recycle this?
https://www.apple.com/newsroom/2021/04/apple-commits-430-bil...
The 2021 announcement works out to $86B per year. The 2025 announcement works out to $125B per year.
In my mind that’s a pretty substantial increase.
IIUC, that $86B in 2021 plus inflation works out to ~$100B in 2025. So it's a 25% increase then?
They paused the last one they announced, so it's an infinite increase if it happens.
But i expect, once the media cycle dies down, it'll get paused too, and then ignored because can't admit that something didn't work out!
Does anyone have a list of the times Apple has lied about bringing jobs back to the US (or keeping them here)?
I recall Apple making a lot of noise about Macbooks being manufactured here, but that they eventually got shipped off to China.
No dog in this fight, and I agree with the premise, however there was never a time Apple made a ton of noise about MacBooks being manufactured in the US.
There was a ton of noise about Mac Pros being manufactured in the US, but sadly, I am not nearly as familiar with Apple after, say 2018*. Not even sure if they have a Mac Pro anymore. :X and if they do, I assume it's not the same model (the black trashcan), so it makes me wonder if they bothered retooling here, or quietly moved it somewhere else
* TL;Dr at some point it became clear to me Cook is Sculley 2.0. I date it to around walking around NYC and seeing an absurd amount of Apple News bus-stop ads. Services! (TM)
Yep. With Trump, people make promises, he forgets, they forget. But the base get their little thrill of the day.
Much as it might be pandering to Trump's nationalist (America first / American only) policies, or simply an action to avoid some of the effects of tariffs that might be imposed, this time around, I see no such connection for the 2021 announcement. Unless they are connected as they are a more generic "pleasing the incoming administration" to try curry favour for when decisions that might affect the company are being made.
[dead]
About half of that is just inflation; $86B is $104B now.
They are focused on sustainability, why not extend it to recycling investments.
(This comment was originally posted to https://news.ycombinator.com/item?id=43158187, so "they" was Apple. We merged that thread hither.)
"Hither" sadly underutilized word. Doubly true for "thither"
Follow those two up with a `yon` and buddy, you got a stew brewin'
Yep, just with less Austin Texas.
Yup.
"Apple’s most recent announcement on US investment was a 2021 promise to spend $430 billion over the following five years, including a 3,000-employee campus in North Carolina, though development on that project has since paused."
https://www.theverge.com/news/618172/apple-500-billion-us-in...
Some new things in the article
- a larger investment number in a previously announced Austin campus
- new factory in Houston "which will create thousands of jobs"
- "doubling its $5 billion US Advanced Manufacturing Fund to $10 billion"
- "It will also open an Apple Manufacturing Academy in Detroit in which Apple engineers and other experts will offer consultations to local businesses on “implementing AI and smart manufacturing techniques,” along with free classes for workers."
Coincidentally, construction isn’t set to start until late November 2028—convenient timing. If this mess blows over, they can quietly backpedal and carry on like nothing happened.
The NYT says the Texas facility will open in 2026: https://www.nytimes.com/2025/02/24/business/apple-tariffs-jo...
Source on the construction start?
Every major business leader is stuck saying "do what we have to" a lot.
The difference between the great leaders and the crap leaders is all in the details.
Good timing because Trump should be significantly weaker, and it'll be clear where Trumpism is headed in the culture, but also it will be more clear where AI will end up.
Even if the move forward with investment, they will be a bit of a 'late' mover, but will have had a chance to see what is working and what isn't working for everyone else.
Trump might be weaker but if Vance were to win the 2028 election he'd just continue the same policies.
Maybe, but that's a huge IF. It's assuming
1. JD Vance independent of Trump will have the same policies.
2. JD Vance will have enough popularity for a serious 2028 run. He might fall out favour with Trump as Trump tries to mount a bid for a third term, Trumpism might just generally lose popularity if policies lead to bad outcomes.
3. Dems don't figure their shit out. They should be able to take back some control in mid-terms, and then start to push their own policies, or at least credibly show that the most extreme policies from the exec branch don't have teeth anymore.
And you're assuming that the 2028 election (and 2026 for that matter) will be business as usual elections, against all evidence to the contrary staring us directly in the face.
It’s incredible this is a possibility at this point.
The federal government doesn't control elections in the US so they don't have much power here. Also, firing all the FBI agents is a bad first step to using them for interference. They have no clue how to be authoritarians; to do that, you need to be popular and have the security forces like you.
Going to be watching closely - but cynically, a promise of investment (for avoidance of tariffs) only needs to last one news cycle until tariffs are no longer top of mind. Then it can be walked back without tariffs being imposed.
Maybe instead of saying the t-word tariff, US gov can charge Apple a special fee on each iPhone. They can call it something catchy, like say, a Core Technology Fee.
Sounds like a throwback lol
https://edition.cnn.com/2019/09/23/tech/apple-mac-pro-united...
This should be the top comment. Apple are doing business the way business is done, just like last time. Results don’t matter, it’s economic policy via press release. Form over substance.
Results matter, it's not hard to imagine that Apple considers the real risk of its promise and market position of being the privacy option being undermined by their supply chain risks, and leverage being used against them by privacy unfriendly actors.
What's the added risk here? It's fine to "risk" almost the entire iPhone itself to be manufactured in China but the servers for some random AI features need to be pure?
Sounds more like technical marketing and the company will treat any decisions around it as a marketing exercise.
Apple's commented previously on why they build in China, and it's beyond just the pricing - the supply chain for every single part they use is in China and mostly in the same geographic region, so there's a level of flexibility there they couldn't get in the US. It wouldn't surprise me if it was genuinely a goal for Apple to manufacture more in the US - they're a notoriously privacy-focused (corporate, not end-user) company, and China's known for IP wandering its way off campus. They're not going to sacrifice the iPhone economics until the US option is actually viable, but I'm not surprised they keep kicking the tires on US manufacturing.
> "the supply chain for every single part they use is in China"
Not entirely true. Some of the highest value components in an iPhone, including the CPU/SoC, baseband, and the majority of OLED displays, are sourced from countries that are not mainland China.
You clipped off
> and mostly in the same geographic region
> They're not going to sacrifice the iPhone economics until the US option is actually viable, but I'm not surprised they keep kicking the tires on US manufacturing.
Apple could, with its immense cash hoard and cash flow, _make_ the US viable, but it chooses not to because it'd rather take the easy way out and have China or India or $COUNTRY fund it and return money to shareholders. They've returned money to shareholders rather than invest it in US operations, by design.
This is a classic feint to protect Tim Cook's entire raison detre. He built his career on super high efficiency operations by outsourcing to cheap labor countries. It relies on the low-to-no tariff access to US consumer money.
And I don't care that it's better for their stock price; that's Apple's problem not mine as a US citizen. And even as an Apple investor I would rather the money be spent on US on-shore operations.
It's mostly about cost and market access to China.
Most smartphone supply-chain for Samsung and Apple exist outside China -- primarily in Japan (camera, sensors), South Korea (DRAM/NAND, OLED), and the US (various ICs fabbed at TSMC in Taiwan). There are quite a few reliable estimates/teardowns showing that these three countries account for close to about 90% of iPhone BOM (bill of materials). That's one reason why Samsung's smartphone unit was able to pull out of China without much disruption back in 2019 -- ie, low dependence on China.
I feel that Apple has pushed this misleading narrative a bit too long to defend their massive China outsourcing.
They've actually been diversifying iPhone manufacturing away from China for a few years already. As of April 2024, 14% of all iPhones were already manufactured in India. That's around 30 million phones per year. And Apple plans to double their India manufacturing again by 2028.
So they shouldn't reduce risk anywhere, if they're currently unable to do it on the iPhone?
Results don't matter as much as PR, this is time when this is unfortunately valid. Just look at US elections.
Measurable results affect rational aspects of our minds, PR attempts to attacks directly emotions bypassing the former, ie to induce impulsive shopping.
Also, what actual security? Apple is as vulnerable as cheap chinese phones against state actors using 0days. Apple devices are still being stolen for spare parts, Apple doesnt secure each component AFAIK and thieves know this (very recent case with friend of a friend, they even knew how to bypass that built in airtag tracking). I haven't seen anything but very well crafted PR statements on this topic. All money-accessing apps on absolutely any phone are a security risk.
But folks love convenience above all.
> This should be the top comment. Apple are doing business the way business is done, just like last time. Results don’t matter, it’s economic policy via press release. Form over substance.
If the Trump administration has any competence, they will rub those old promises in Apple's face until Cook actually does something meaningful.
The whole Trump administration is all about form over substance, though. I would not expect Trump to do anything actually productive about it, as long as Tim Cook sings his praise (and pays his dues).
[dead]
Gruber suggests Apple already had these plans and simply packaged the up as a win for the current administration:
https://daringfireball.net/linked/2025/02/24/mission-accompl...
I don't pay enough attention to Apple's plans to judge if he is right.
This is inline with what other entities (Canadian and Mexican governments) have done when threatened with tariffs.
[dead]
I don't understand. Can companies curry government favor to get tariff exceptions? Aren't the tariffs in place already?
Charitably, tariffs exist so POTUS can either lower taxes or increase jobs in US, but both would take time to pan out assuming things go well. So if a company is willing to onshore money or jobs, its achieving its intended purpose in their eyes.
I believe the stated purpose of the latest round had to do with fentanyl precursor manufacturing.
I believe that's the Mexico/Canada Tariffs, but it's really hard to keep it straight, by design.
No I think Mexico/Canada were largely about stopping immigrants and fentanyl smuggling. But China was targeted for not doing enough to stop manufacturing of fentanyl precursors.
Sorry to say but that's already been walked back on after Canada committed $1 billion dollars for extra northern border security and it made no difference in the tariffs discussions.
https://www.cbc.ca/news/politics/rcmp-border-patrol-canada-u... <- discussion of commitment from Canada
https://financialpost.com/news/trump-says-tariffs-on-canada-... <- as of today (sorry for FT source but this is literally all across google you can just googla USA tarrifs)
The idea that the Canadian tariffs were ever really about Fentanyl is patently absurd.
0.2% of all US border Fentanyl seizures were on the Canadian border. That's almost literally nothing.
The Fentanyl numbers are nothing in comparison to Mexico, but that doesn't mean it's not a problem.
Also it's not only about drugs, but also humans smuggling (https://www.cbc.ca/news/canada/fentanyl-dr-smuggler-1.737348...), and overall border security.
- According to CIS, the number of Canadian crime groups producing synthetic drugs doubled between 2023 and 2024 - There's a lack of Canadian agents who are tasked at preventing this and current legislations make it very inefficient between federal and provincial law agents - There's an upward trend in Fentanyl seizures in Canada the last 2 years - Fentanyl is now being produced domestically in Canada
All of that is within the control of Canada with better policies.
Let’s put it into perspective, because those numbers don’t give a baseline for what the problem is. Also they don’t necessarily have anything to do with trafficking.
Last year there were 45 lbs of fentanyl intercepted crossing into the US from Canada. Thats a backpack. There’s 500x as much coming from Mexico.
It’s unrealistic to expect that zero fentanyl will come into the US from Canada, and until that happens we will tariff all trade with them.
That's kind of the point of tariffs
Yes this is one reason tariffs are so valuable to a corrupt POTUS. They have essentially unilateral and very fine-grained control over them, down to exempting specific companies or products outright.
Congress needs to step up on this, honestly. The entire idea that the President can unilaterally implement trade policy is as plain a violation of separation of powers I can think of, and SCOTUS is a fan of non-delegation doctrine.
Legislators step up when enough of their voting constituencies make it clear that they value something as a non-negotiable (assuming votes still matter).
Which means those who care about this are back to not only contacting legislators but also persuading a lot fellow voters that separation of powers is crucial and worth prioritizing over familiar well-handled and loved heuristics.
Yes a million times. For all the rhetoric about authoritarians, the Democrats never seem to want regin in Executive power when they are they majority. It is like a game of chicken where America winds up with a populist dictator from either the left or right.
Yes because populism is a reaction to government being generally unresponsive to people’s needs.
Congress has become increasingly unproductive and unresponsive. There are many popular policies that Congress essentially ignores, and many problems that go unsolved. So trust in government dwindles and people crave strongman solutions.
I’m not sure there is a solution. There are so many interlocking problems gumming up the process that any “we just need to fix X” solutions (where X is gerrymandering, money in elections, lobbying, the two party system, first past the post, corruption, income inequality, the electoral college, the slow death of journalism, consolidation of industries, etc) are nearly impossible and also probably insufficient because they all feed back into one another: they are both causes and effects.
So when people are mad about a downstream effect like the price of eggs and digging any deeper touches one of the topics above (“to fix egg price gouging you need to reinvent the political system” sounds a lot like “to make an omelette first you need to create the universe”), it’s really easy to throw your hands up.
We'll never know for sure, but I don't think Harris would have ended with delusions of being populist dictator.
> Congress needs to step up on this, honestly.
Congress seems pretty unwilling to do much of anything right now.
Sorry they are right now cutting Medicare/Medicaid/snap right now to justify billionaire tax cuts.
https://www.newsweek.com/nearly-all-republicans-vote-against...
tl;dr the New Deal.
The Reciprocal Trade Agreements Act by Congress gave FDR sweeping authority over tariffs without direct congressional approval.
Then later the Trade Expansion Act under Kennedy.
Historically, these have served to decrease tarrifs.
[flagged]
We handed Congress to a complicit party. This is a full bunker situation until the midterms.
Clearly a non-sequitur, but I'll bite anyways. Paul Pelosi is a VC, and I'm sure that most of the Pelosi net worth is due to his income, not hers. Since the laws preventing members of Congress from trading on information they receive as part of their duties, you can't say that the Pelosi's have violated any securities laws.
And both sides of the aisle benefit from this. Whether it's legalized insider trading or jumping to corporate jobs when out of office, it's a corrupting influence. All members of Congress, SCOTUS and POTUS should have to place their assets into blind trusts. That won't stop this corrosive influence, but it is the bare minimum.
> Since the laws preventing members of Congress from trading on information they receive as part of their duties, you can't say that the Pelosi's have violated any securities laws.
Oh yes, you can:
"Paul Pelosi, 83, sold 30,000 shares of Google (GOOGL) stock in December 2022, just one month before the tech giant was sued over alleged antitrust violations." https://finance.yahoo.com/news/deceptive-tactic-nancy-pelosi...
> Since the laws preventing members of Congress from trading on information they receive as part of their duties, you can't say that the Pelosi's have violated any securities laws
The feds can and do go after people for using family members and friends to execute trades. So if Nancy Pelosi told her husband some material non-public info, and Paul Pelosi traded on it, that would still be insider trading.
There was a guy at Microsoft who was caught once using a friend to place trades. He said he talked himself past his ethical concerns by reasoning that members of Congress do it.
https://www.businessinsider.com/sec-alleges-insider-trading-...
Edit: to be clear, the absence of a prosecution does not mean that the Pelosis did not insider trade. Nor that they did. We can't tell from this distance, only speculate.
"Paul Pelosi, 83, sold 30,000 shares of Google (GOOGL) stock in December 2022, just one month before the tech giant was sued over alleged antitrust violations."
https://finance.yahoo.com/news/deceptive-tactic-nancy-pelosi...
[flagged]
Who is advocating for that?
iirc as soon as anything becomes beyond the border the President holds the keys for various reasons including the ever-vague “national security” but also due to being prescribed as the primary negotiator https://en.m.wikipedia.org/wiki/Treaty_Clause
Tariffs are not treaties, they're taxes/duties - only Congress has the power to raise them. Article I is extremely clear on that. Historically, tariffs were always raised by acts of Congress, not by Presidential fiat.
Trump is using extremely misguided legislation from the 1960s/70s where Congress allowed the President to enact tariffs for national security and emergencies. There is a very strong argument (in the sense it resonates with the conservative SCOTUS majority) that Congress cannot delegate its fundamental powers to the executive by legislation alone.
I think people are just too cowardly to bring a case in front of the courts to challenge the constitutionality of it all. Non-delegation doctrine is what the Federal Society want to use to kneecap all federal regulation. Trump operates on a spoils system so it's not in the interest of conservatives or businesses to challenge him, for fear of retribution.
Trump is using tariffs not to raise revenue, but rather use it as a stick to force companies to invest in USA.
Previously they were outsourcing and offshoring as much as they could get away with it. Which led to transfer of advanced technologies outside USA and America losing its manufacturing and technology edge
So how's that going? Outsourcing seems to be going strong, the tarriffs instead pissed off allies who are preparing counter-tarriffs, and the CHIPS Act is being dismantle as we speak (there goes our investment.
If TSMC acquires Intel (which is in the works), then chips act (basically a government handout to private corp) wont be needed anymore
Yeah. Just killing the US semiconductor production is the cheapest option.
You should let his Commerce Secretary know (or stop sanewashing)
https://www.axios.com/2025/02/20/commerce-secretary-lutnick-...
Cool, still unconstitutional.
So, for those of us who haven't studied this in depth, why do you say it's unconstitutional? Do tariffs require congressional approval? Or what?
I have already explained my thinking up this comment chain. I'm mostly replying to GP who misunderstands that the intent of the tariffs is besides the point.
TL;DR read Article I section 8, read up about the Trade Expansion act of 1962 and Trade act of 1974, and "non-delegation doctrine", and you can trivially find legal debate about the constitutionality of IEEPA. Rather than listen to random nerds on HN you should seek out this information yourself.
Tariffs are a form of taxation. If I want to import say tea, and the government is placing a tariff on that imported tea, I am effectively taxed by the government. And only Congress can impose new taxes.
The Trade Act of 1974 gives the president power to impose retaliatory tariffs.
Not saying you're wrong, but... I have seen claims that tariffs are a source of government income that Congress doesn't control. You're claiming they do.
I haven't seen a citation from either side. Can you substantiate your position?
You can when the president is corrupt.
Its not corruption when the president creates US jobs
It is corruption when president outsources and offshores US jobs, though
There is no jobs problem in the US though. Unemployment is at 4% which is mostly just job churn. Long term unemployment is only 1%.
US consumers, that’s all of you, are being hammered with taxes on imported goods most of which can’t realistically be produced in the US anyway, to solve a problem you don’t have.
A commitment like this takes years to plan. It can’t possibly be a response to tariffs announced weeks ago. This is all optics.
Can you tell the tech industry and my temp office that, please? Job market still sucks in Los Angeles.
And I don't trust unemployment in this gig economy. I'm technically not unemployed, but I haven't had a full time job in nearly 2 years.
I fully believe that the real ("main street not wall street") economy is in worse shape than government numbers on unemployment suggest and both sides are to blame for different aspects of this problem.
But nothing Trump is doing is going to fix your situation.
In fact he (or rather Elon/doge) is very actively making things worse for you with the massive government layoffs, flooding the market with even more people to compete with you for jobs making finding work more difficult and also eventually dropping all of our wages.
>But nothing Trump is doing is going to fix your situation.
I'm aware. I'm sure he's responsible for at least 3 job freezes I ran into mid-interview this year. He's literally costing me job opportunities because no one can budget around this chaotic government.
4% too many and probably understated. The BLS repeatedly underestimated unemployment during the previous administration. Also the labor participation rate, which is harder to game, still hasn't reached pre-Covid levels yet: https://www.bls.gov/charts/employment-situation/civilian-lab...
> https://www.bls.gov/charts/employment-situation/civilian-lab...
In February 2020 it was 63.3% and in January 2025 it was 62.6%, for a difference of 0.7%. Also note the steady decline post-2008 and the multi-year plateau that jitters around 63%.
Having the plateau change from ~63% to ~62.5% isn't an unreasonable scenario.
* https://fred.stlouisfed.org/series/CIVPART
The labor participation rate is going to be declining for the foreseable future because the population is aging.
If you look at the prime-age employment rate, we're almost up to the record high levels: https://fred.stlouisfed.org/series/LNU01300060
> Also the labor participation rate, which is harder to game, still hasn't reached pre-Covid levels yet
That's because Americans got /richer/ and retired. You want to look at the prime-age rate, which the boomers have moved out of.
Right. One reason we're in this mess is that voters don't even notice unemployment but are extremely sensitive to inflation.
False, USA has a big problem with manufacturing. All US jobs are service jobs to prop up consumer economy, that have no strategic benefit.
A lot of fake employment and low productivity jobs are in the government/NGO sector, paper pushers, DEI jobs, law/compliance type jobs - that should have been manufacturing jobs instead.
USA has no shipyards and infrastructure is crumbling precisely because of misallocation of resources and labor
> USA has no shipyards and infrastructure is crumbling precisely because of misallocation of resources and labor
That's because of the Jones Act and other poorly designed protectionism.
DEI jobs and the _total_ NGOs employment are basically nothing in the overall employment. Less than 1%.
Heck, the _whole_ Federal government workforce is less than 2% of the total workforce.
> USA has no shipyards and infrastructure is crumbling precisely because of misallocation of resources and labor
Yes, and you have tariffs to thank for this. The Jones Act requires US-built ships for any ship traffic within the US waters.
So the shipbuilders were insulated from competition and just degraded into Defense contract moochers.
Depends how you count it. I wouldn’t count most gig app workers as employed. But the government does.
The government determines the employment rate via surveys, i.e. they just go and ask people if they're employed. It's not a calculation from taxes or from employers or anything.
So it's up to the gig workers if they think they're employed or not. Presumably this depends on how often they do it.
I also wouldnt count fake bullshit government/corp/college administration jobs (email jobs) with negative productivity
I just gotta pay rent dude. I'm not judging how you do it.
Neither of those fit the definition of corruption.
Apparently, yes. I saw mention of discussion around the Trump administration potentially giving Apple a tariff waiver. And I believe in Trump’s last term, Apple did have some sort of waiver.
I’m on mobile but Googling for “Apple tariff waiver” and “Apple tariff exemption” will point you to several news items.
A couple (from a pool of many) citations:
From the 45 presidency: https://www.marketwatch.com/story/apple-escaped-tariffs-last...
Now: https://9to5mac.com/2024/12/17/analyst-trump-will-waive-appl...
I am a free trader in principle. However you have a country (China) with an authoritarian government that makes favored industries subsidized.
Of course the standard economic argument is that China using its GDP to make goods cheaper for our own citizens to purchase is better for us - they are subsidizing our economy. However it ignores the strategic disadvantage by our country losing its manufacturing capabilities.
The graphs may show economic advantage. It’s hard to quantify the long term strategic and militaristic disadvantage to not being able to make anything yourself if a world war occurs.
> However you have a country (China) with an authoritarian government that makes favored industries subsidized.
This is overlooking the forest for one tree. The thing is, mean chinese manufacturing wages are $25k/year (purchasing parity adjusted! $15k unadjusted) for a 49h week.
That is the reason that so much manufacturing/industry has shifted there, not some nebulous "Chinese government subsidies" (not saying those are not a thing, just that they don't really matter all that much).
> It’s hard to quantify the long term strategic and militaristic disadvantage to not being able to make anything yourself if a world war occurs.
Certainly. But forcing low-skill industry to stay at a relevant size in a high-wage country is expensive business (compare agriculture, which is subsidized basically for exactly this reason) and not straightforward (see Jones act).
Presenting tariffs as a viable alternative to taxation is just beyond ridicule, but that has not stopped people so far either...
Salaries are just a small part of the reason industry works in China.
The bigger picture is that China invests in the development of an industrial chain. This has many aspects: infrastructure, education, training, housing, and of course tax incentives. The USA decided to stop investing in practically all of these. Even scientific research, the last area in which the US used to lead, is now in jeopardy from both sides: competition from China and internal cuts.
They could move to Bangladesh or Africa and pay $3k / year. They aren't. China has many advantages beyond cheap wages.
Absolutely. You do need a minimum baseline for infrastructure, government stability and workforce.
Most of Africa is just starting to slowly get there, Bangladesh is already very relevant for textile production.
I would expect the same basic trend to repeat that we saw with electronics manufacturing in 90s Japan: First cheap products move (very wage sensitive), then the local sector expands, wages rise with the whole local industry moving up the value chain, then at some point local wages become high enough for the whole process to repeat with the next low-wage country...
I think trying to block this trend off with tariffs is a futile waste of taxpayer money which american consumers are gonna pay for.
Spending tax money to keep some degree of self-sufficiency in critical industries (like with agriculture) can be a solid idea if done sparingly and cleverly, but that is not how the current US admin has approached this...
[dead]
>However you have a country (China) with an authoritarian government that makes favored industries subsidized.
If things keep going the way they are going, that could describe the US just as well in a few short years.
https://en.wikipedia.org/wiki/United_Fruit_Company
"The company responded by intensively lobbying the U.S. government to intervene and mounting a misinformation campaign to portray the Guatemalan government as communist.[18] In 1954, the U.S. Central Intelligence Agency armed, funded, and trained a military force that deposed the democratically elected government of Guatemala and installed a pro-business military dictatorship.[19]"
This is not a one-time aberration.
> If things keep going the way they are going, that could describe the US just as well in a few short years.
Why do you think so?
Sour grapes. Most economists were just happy with this situation until recently. What I mean is, the current situation arises by the desire of Western businesses of getting hid of productive investments and concentrating only on capital investments. It has nothing to do with trading with an authoritarian government or not, which almost everyone believed was Ok until recently.
The US subsidizes a hell of a lot themselves already.
What about Canada? We have even higher tarrifs than China and in the last 100 years are the USA's closest ally.
"I am a free trader in principle. However you have a country (China)"
just think of china as another trader with more capital than you and pull yowrself up by your bootstraps.
> The graphs may show economic advantage. It’s hard to quantify the long term strategic and militaristic disadvantage to not being able to make anything yourself if a world war occurs.
Is the United States at risk of not being able to make anything ourselves? We have the second largest manufacturing output in the world.
https://en.wikipedia.org/wiki/Manufacturing#List_of_countrie...
there’s also the “national security aspect”, government often wants certain hardware not to be made offshore at all
Maybe the next administration should keep up the tariffs (as Biden did to a degree). Cheap trade with China distorts the tech sector too. Jobs and Wozniak were the products of a system in which americans had to build products at home. Tim Cook is the product of a system where you can become a trillion company by hyper-optimizing foreign supply chains. Which is better?
When did isolationism become cool? Isn't this why we declared independence in the first place? To get away from the British restricting free trade?
You’re incorrect about history. Mercantilism not only restricted foreign trade, but restricted domestic industrial development by requiring the colonies to sell raw materials to Britain and buy finished goods from the Britain. Tariffs were a core pillar of the Lincoln Republican Party.
There’s been an isolationist wing in tech as long as I’ve been in it (early 2000s). I remember chatting with someone at Cisco/Juniper in the late aughts about Huawei ripping off their router designs down to the silk screening. Of course today Huawei makes their own state of the art routers with their own silicon, and some lower-end Cisco/Juniper gear is white boxed foreign equipment. And of course tech folks were complaining about immigration and outsourcing back in the early 2000s when Republicans were enthusiastically supporting both.
When the people with the money decided it was better spent in places that weren't their own country.
Having access to cheap oversea steel allows Americans to focus on building companies with significantly higher value-add. Onshoring low-value industries is a massive human capital waste and an easy way to depress wages.
It allows a very small portion of Americans to build companies with significantly higher value-add.
It destroyed the futures of a larger number of Americans.
Then again, why do we make the distinction "American"? If you have people who became unfathomably wealthy by shipping off strategic industries to the lowest bidder regardless of geopolitical implications, does nationality matter anymore?
No, the analysis (and it’s not exactly rocket science) says just the opposite: Way more downstream manufacturing jobs that rely on steel as input are lost, vs. domestic steel production jobs gained.
https://www.pbs.org/newshour/economy/making-sense/steel-tari...
Here's another idea: tax those people appropriately and pump that money back into the economy...
Way easier and more globally optimal than just saying we're going to do absolutely everything (even the shitty jobs) here in the US.
If there's one thing those people hate more than paying Americans to do labor, it's paying taxes.
And...? Everyone hates paying taxes. Normal Americans pay them anyway.
Indeed, but the hallowed Job Creators have the means to influence the people in power to make the taxes go away.
And congress is trying to kowtow to it as we speak
https://www.newsweek.com/nearly-all-republicans-vote-against...
If we don't put pressure, those people will get their way.
Do "normal Americans" pay taxes? From the numbers I've seen, ~1/3 - ~1/2 of tax filers receive more money from the government than they pay. To them, "refund season" is a cause for celebration rather than a stressful event.
> Do "normal Americans" pay taxes?
Yes.
The on-average crossover between negative and positive net total federal income (individuals will differ because of individual circumstance beyond just income level) tax when taking into account refundable credits (most notably, but not exclusively, EITC) is a bit below the median personal income but not that far below it, so certainly lots of individual "normal" (by most reasonable definitions) Americans do not pay net federal income tax .
But even if they don't pay net federal income taxes, they probably still pay a net positive amount in a variety of state taxes, federal payroll taxes, and federal consumption taxes (e.g., gas tax.)
I'm a normal American and owe 6000 this year to the Feds, so yes.
W-2 get refunds because the Feds took out too much from their paycheck beforehand.
I'm a normal American and have only once paid more than I've received for federal taxes. Withholding has nothing to do with it.
The problem with taxes is that it's a prisoner's dilemma. You need global cooperation at some base level of taxes, otherwise companies move to more favorable tax jurisdictions in the long term and offshore from there, which would hurt the US even more. It doesn't have to be all-or-nothing, but any marginal dollar of increased taxes in one place will have some non-zero effect of encouraging the next investment dollar to be spent elsewhere.
To be clear, I do think capital gains taxes are criminally low in the US relative to income tax, so I'm not arguing in _favor_ of lower taxes. I'm just saying why raising taxes isn't a panacea.
You can raise US company taxes and capital gains tax a lot before the US stops being a low-tax country.
You’re not wrong, of course, about how every tax percentage point matters. But Americans arguing that their taxes are too high is never not hilarious.
Creating an underclass that relies on economic elites paying taxes rather than being economically independent because you want to optimize for "high value add industries" is a terrible long term strategy.
So, like a tariff?
> tax those people appropriately and pump that money back into the economy
So make the US to be like a far less successful country? Kill your economy by increasing taxes? The US economy is singularly successful because it has incentives to build businesses - see YC.
Have you tried living in a country that doesn't encourage businesses? They are often great tourist destinations. I'm in New Zealand and too many ambitious young people leave here: we have an emigration problem because our economy sucks. The government fixes the economy with 30% immigrants (disclaimer: I love immigrants). I have many friends that are never coming back here except for holidays. I hate the New Zealand government incentives for businesses (taxation and regulation) and I can see no way to fix them. Even our "business" political party ACT is completely fucked (latest story - they will be selling everything profitable to overseas "investors" - destroying the economy).
Taxation incentives matter to businesses. Be careful what you ask for because the majority have little understanding and vote for the wrong incentives.
Even business owners don't seem to understand incentive systems that well. Perhaps game designers do?
Yeah, screw that. Capital taxes are at record lows and they want to make it lower at the cost of Medicare and Medicaid.
They are parasites at this point. If they think they can find lower taxes than 22% they are happy to leave. As if they aren't already avoiding taxes.
I will respond by https://en.wikipedia.org/wiki/Splitting_(psychology) myself.
Do Medicare and Medicaid exist without businesses? I'm from New Zealand and our society causes problems for our socialised healthcare.
Businesses are symbionts: productive societies accept some costs from businesses so long as the society get more gains.
Why do you look at money as though that is all that matters?
Who measures the benefits we get from modern society?
Finding downsides and complaining about them is too easy. Looking for upsides is less popular.
Every poor person I've met avoids taxes.
Do income taxes on the 60th percentile earner completely kill their incentive to earn an income?
Then why would ensuring the same effective tax rate on the 99th percentile kill their incentive?
The ultra-wealthy appear as toxic to me too.
However I believe that incentives need to be marginal. If you already have a lot perhaps you need a big carrot as your incentive? I don't know any billionaires that I can ask how they feel about taxation incentives: I reckon you are making assumptions about what you think they should feel.
What makes Tim Cook make the US more money?
Taxation cliffs are shit. In New Zealand our Green party decided that 1 million was enough. Why would you bother growing a business after you reached 1 million? Retirement? A business is defined as being about making money (albeit some people do run "businesses" for other outcomes - why is Warren Buffett still working?).
High marginal taxation is also shit IMHO.
The hard part is to design the incentives so that productive people build your economy for the benefit of everybody.
If a government discourages business then the economy is crap and everybody suffers. See other economies.
Few people understand the incentives of others, and few people understand how wealth is created for all: the hoi polloi dismiss the wealthy as vampiric money grubbers. Anyone who uses the word capitalist in a derogatory way has been brainwashed. Most everything that makes our economies work is invisible non-monetary rewards. https://news.ycombinator.com/item?id=43162596
I can speak for my own financial incentives. My perception is that I have an effective tax rate of well over 50% in New Zealand (any retirement savings are not safe because our demographics and governments will screw our economy).
I do not feel the incentive to work in a business - My attitude means I now produce marginally less than I could for the New Zealand economy (I still pay taxes so they are advantaged but they could get a lot lot more from me). I now mostly selfishly concentrate on those closest to me. Why should I work if it isn't marginally beneficial enough for me? I'm no more selfish than my retired friends that I know (a wide variety of people from many walks of life).
(Reëdited to expand and clarify).
We can't decide how much is fair. Compare yourself to a dead king - what is fair? We can design systemic incentives so that we each make the world better for everyone. Not that that it is easy... Trite thoughtless dismissals of the most productive members of society are not helpful.
Edit 2: I guess this discussion is as close to work as it gets for me. Too much adulting. Should I get into politics? Are morals an impediment to helping others? There are too few politicians I admire, and too many I wouldn't want to shake hands with or be associated with. Every idiot has political opinions - how much of an idiot am I? Every politician is smart enough to win an election - they are not stupid yet they make too many horrific mistakes. What about the cryptically smart ones? I see how systems affect people that join a system. What would I become if I join our political system? Understanding our different systems is hard because they grow so weirdly with vestigial complexities due to history, complex interactions, and reflexivity.
>Onshoring low-value industries is a massive human capital waste and an easy way to depress wages.
Compared to now where there's just no full time jobs and also wage suppression?
I get the value add argument, but lots of people just need income to pay for living expenses. Without an income, those people become disaffected and sometimes violent. Then they embrace right-wing protectionism because, while their gadgets are cheaper, they have no income to buy cheap gadgets.
Nor can they move to these offshore places (where the cost of living is lower) because immigration laws exist in part to control worker mobility.
> Having access to cheap oversea steel allows Americans to focus on building companies with significantly higher value-add. Onshoring low-value industries is a massive human capital waste and an easy way to depress wages.
That's the talking point, but it's bullshit. A lot of those "low-value industries" are fundamental capabilities, and China sure as hell isn't going to let the US own the "higher value-add" areas. They dominate those next, and the US free-trade business elites will be fine with it as long as they get to make some money for themselves.
Being a high value-add area is endogenous to how hard it is for others i.e. China to reproduce. In other words, if it were easy to make GPUs they wouldn't be so damn expensive.
The newborn US imposed a ton of tariffs specifically to escape British control of industry
Indeed. I think we've come very far 250 years later. But we lowered those 90% tarriffs 100 years ago for a reason.
It's always been cool. It's why you can't buy a Kei truck or a BYD car
> When did isolationism become cool?
Ask maybe China.
China's economic power is certainly not rooted in their isolationist social policies. They're just as bullish about foreign investment as the US was at the height of the free trade era.
>> Isn't this why we declared independence in the first place? To get away from the British restricting free trade?
No. I'm not sure where you got that idea. If you look at something like the Boston Tea Party, it wasn't high taxes on tea that were being protested against, it was lowered taxes on tea that undercut the smuggling operations of people like Sam Adams and John Hancock. "No taxation without representation" makes better press than "No undercutting my smuggling operation" though.
In the early years of the US, between 80 and 90 percent of federal revenue came from tariffs. Not exactly free trade.
> In the early years of the US, between 80 and 90 percent of federal revenue came from tariffs...
To be fair, the Federal Budget back then was 2%-ish of GDP. And their political consensus gave the Federal Gov't very few things that it had the power to tax.
For readers looking for context, Google tells me it was ~24% of GDP as of 2024
https://fred.stlouisfed.org/series/FYONGDA188S
Is a better graph
That seems impractical and unsustainable.
Based on what? Even adding in state spending, the US government spending share of GDP isn't high by world standards.
So that means the government is taking more than 1/5 of all the money generated by the US. That’s crazy and no wonder the nation is going bankrupt
It reminds me of the Portuguese court taking 1/5 of all the gold mined from colonial Brazil:
https://pt-m-wikipedia-org.translate.goog/wiki/Quinto_do_our...
State and local governments also take their cut. I believe the total is about 35%.
Important to note that outlays do not equal federal taxes, because spending exceeds revenue by a substantial margin.
I want to say “that’s not what isolationism means”, but I realize it starts to feel vague just like the word “fascism”, used when convenient but varies wildly in rhetorical meaning… to be more specific is better, I like what George Washington had to say about it in his farewell address because it shows the nuance of the topic across the spectrum, it’s not as simple as isolation good vs bad:
The great rule of conduct for us in regard to foreign nations is, in extending our commercial relations to have with them as little political connection as possible. So far as we have already formed engagements let them be fulfilled with perfect good faith. Here let us stop.
Europe has a set of primary interests which to us have none or a very remote relation. Hence she must be engaged in frequent controversies, the causes of which are essentially foreign to our concerns. Hence, therefore, it must be unwise in us to implicate ourselves by artificial ties in the ordinary vicissitudes of her politics or the ordinary combinations and collisions of her friendships or enmities.
Our detached and distant situation invites and enables us to pursue a different course. If we remain one people, under an efficient government, the period is not far off when we may defy material injury from external annoyance; when we may take such an attitude as will cause the neutrality we may at any time resolve upon to be scrupulously respected; when belligerent nations, under the impossibility of making acquisitions upon us, will not lightly hazard the giving us provocation; when we may choose peace or war, as our interest, guided by justice, shall counsel.
Why forego the advantages of so peculiar a situation? Why quit our own to stand upon foreign ground? Why, by interweaving our destiny with that of any part of Europe, entangle our peace and prosperity in the toils of European ambition, rivalship, interest, humor, or caprice?
It is our true policy to steer clear of permanent alliances with any portion of the foreign world, so far, I mean, as we are now at liberty to do it; for let me not be understood as capable of patronizing infidelity to existing engagements. I hold the maxim no less applicable to public than to private affairs that honesty is always the best policy. I repeat, therefore, let those engagements be observed in their genuine sense. But in my opinion it is unnecessary and would be unwise to extend them.
>when belligerent nations, under the impossibility of making acquisitions upon us, will not lightly hazard the giving us provocation; when we may choose peace or war, as our interest, guided by justice, shall counsel.
Ironic that as a Canadian, the US is moving from the nation that would be guided by Justice into the belligerent nation in this situation.
It also serves as a lesson to us that we should have learned from you and George Washington, and stood on our own first and ensured our own security before cooperating with others. We have a long way to go to get back there now, unfortunately under the position of potentially our closest ally and economic partner being belligerent, untrustworthy and unreliable.
Apple went bankrupt under Jobs and Wozniak and was saved by hyper optimizing foreign supply chain company Microsoft only to rise 10 years later by focusing on hyper optimizing foreign supply.
There was a lot more than that going on and I think you've pretty generally mischaracterized the main problem with the mid-80's era Apple—which had nothing to do with domestic manufacturing and everything to do with not delivering new products that people wanted, at a reasonable price. You can claim overseas manufacturing solved the pricing component of that, but that's not at all clear: other companies were manufacturing in the US at the time and still out-competing Apple.
All of the big PC companies had factories in Texas in the 80's and early 90's, didn't they?
And Dell became a case study of outsourcing everything (and sending your stock and profits soaring the whole time), until you have nothing.
I don't know about Microsoft, but I'm very clear that the "miracle" operated by Apple was exactly to perfect foreign supply chain at a time when Intel/Dell/HP and others were still heavily focused on the US. The quality of Apple products was already there since the beginning, but they had no way to compete with the PC market until they figured out Asian supply chains.
A bit revisionist here.
If Trump is our McKinley or Hoover, I'm excited for our next FDR.
My first thought was payment to avoid sanctions for being "woke" (read: anti-discrimination)
[dead]
Let's go with 'Genius Tax'. Your average Apple fanboi will be lining up around the block for the prestige to pay that twice :D.
Anyone keeping count of how many trillions in hypothetical investments and millions of jobs large American corporations have promised in the next 3-5 years?
[flagged]
While I have no real opinion on this, I do have questions:
* What type of jobs?
* Does the US have the required people, in terms of numbers and skills?
* Does this mean moving to US based fabs for the M-series chips?
* Is this actually profitable, or is this just a political move?
> Does this mean moving to US based fabs for the M-series chips?
This is not really a practical option. A big part of the M-series success is TSMC's lead in cutting edge process nodes. And Taiwan does not allow export of technology for the latest nodes. It is available only there.
* What type of jobs? - "The 20,000 additional jobs, Apple said, will focus on research and development, silicon engineering and AI."
* Does the US have the required people, in terms of numbers and skills? - "The company is opening up what it calls a manufacturing academy in Detroit, where it will help smaller companies with manufacturing. It already operates an academy for app developers in the city. It’s also doubling its manufacturing fund in the US to $10 billion." - Sounds like they are upskilling, and will count the employees of companies joining the academy as "jobs created"
* Does this mean moving to US based fabs for the M-series chips? - "[M-Series] chips themselves, however, continue to be produced in Taiwan.
* Is this actually profitable, or is this just a political move? - Define profitable. It is cheaper than paying tariffs.
I might be reading it wrong, but that's the 20,000 ADDITIONAL jobs, which is going to be R&D, engineering and "AI".
Those 20,000 people won't be staffing the production lines. So how many manufacturing jobs, especially low skill, entry level with decent pay, will this create? The whole thing is framed in a way that makes it sound like Apple is creating thousands of manufacturing jobs.
data labeling
I seriously doubt it's cheaper than tariffs.
> Does the US have the required people, in terms of numbers and skills?
For 30 years, IT managers at blue chip US corporations have exploited the H1-B visa program by saying, "No," and then hiring a never-ending stream of barely-capable Java coders from programmer mills in India, take 5 times longer to make an app than it should have taken, get promoted, and leave everyone holding the bag with shitty web app that we all hate because it's too slow, too bloated, and doesn't work like it needs to. And the companies who can't get enough of that bullshit in-house just hire it out to sub-sub-contractors that do the same thing. Can we not invest in our native population and education systems this time around? I'm so tired of the fact that 90% of the IT staff in my Fortune 250 is Indian, and I know people who would be better at their jobs living in my home town. It hurts our community and our country, in the long run, and by the VERY same logic as re-homing our chip production.
It sounds like you should be directing more of your anger to the C-suite than the people they’re hiring. If they couldn’t get even cheaper Indian immigrants you’d be complaining about code boot camp hires instead - what you need is a tech union which would give you the ability to push back against short-sighted decisions which make your life worse cleaning up messes.
Well, those Indians living in the US will have families of their own, and over time become part of the community you claim to be a part of. Very much like your ancestors did, except they likely didn't face the arbitrary constraints on immigration that Indians (and any other nationality) face today.
I would find it hard to believe that there weren’t racial prejudices involved at literally every point of immigration in American history
The same thing that happened in the UK will happen in America.
People in the UK who are against immigration are often talking about Poles who moved to the UK after the EU and not Indian families who have lived in those neighborhoods for generations.
The crazy thing is, it's not that long ago that Irish and Italian immigrants were not discriminated against. They didn't even consider Italian immigrants to be white.
And one more:
* Will this raise prices for customers outside the US for no justifiable reason?
And related to your last bullet point: will it actually happen or is floating this just a political move...
It will actually happen because it’s nothing new. The 500b is almost all wages for existing US-based employees. They are looking for a carve out from the new China tariffs (same as last time). Note - they made a very similar announcement 4 years ago https://www.apple.com/newsroom/2021/04/apple-commits-430-bil...
The required people can just be imported from China?
> Does the US have the required people, in terms of numbers and skills?
doesn't matter because we have work visas
Sounds like ireland or wherever their tax haven is might make some real savings from this
Thanks to ireland all big US corporations saved hundreds of billions of dollars past few years so now they can get back to US with this massive cash for anything they want (ofc. nothing will get back to EU as long as they ignore tax heavens)
Looks like the US is moving fast to bring the whole supply chain for everything and anything into its own borders.
I wonder where this will lead to.
Depression.
The US is a great place to have your headquarters and a terrible place to have your not-so-cheap labor.
Their actions will drive prices higher and, indirectly, wage higher. Businesses without a war chest will not be able to keep going and fold, the labor market will collapse.
The rest of the world will trade amongst each other and I suspect to save themselves, some big tech companies will relocate their headquarters.
As an ex Apple engineer, i think you're overlooking that a huge fraction of labor is robotic, and has been for at least 15 years.
(even in the famous contract manufacturers used by Apple and Dell, etc.)
The point is that what you counted as not cheap labor probably is largely capex.
Labor is only costly when humans do the work.
But that will be less and less the case.
Waymo is doing 150,000 autonomous rides per week now.
That could also lead to a depression. I haven't heard a lot of politicians here (Andrew Yang in 2020? does he even count as "a politician"?) with good plans for what to do when automation hits jobs even harder.
A failed politician is still a politician.
In three cities with good weather and they don’t do highways.
And a year ago they did less than 15,000!
That is not even one tenth!
They clearly will not get anywhere with this.
Yes, just like every startup pitch deck.
“We grew from 10 customers to 100 customers in a year. At this rate we will have 20% of the world’s population in a decade!!!”
The first cohort of customers of any company is always the easiest to obtain with the lowest acquisition cost. You solve the easiest problems first.
This is Cohort Analysis 101. Not to mention Waymo still hasn’t shown to be able to operate in less than ideal weather conditions or proven that the unit economics will make sense or be economical especially taking into account maintenance, or utilization ratios.
It’s been operating safely in each market they’re in. The AI keeps getting better. They have no competition (please don’t bother mentioning Tesla vapor ware). Path to high growth seems pretty sure at this point.
Which is very very few markets, and all of them share weather patterns that are very similar.
When Waymo can demonstrate reliably going from Chicago to Ann Harbor in the middle of a snow storm thats when we can start talking about how its good enough.
It reminds me of the advanced ED209 robot in Robocop that was taken out by an inability to go down a flight of stairs
https://youtu.be/_MS4sLlBvbE
And the markets they are in are low hanging fruit with good weather. I’m not saying Waymo is less safe than human drivers. I am saying that it will only take one fatal accident by any self driving car for people to lose confidence, investigations to start, rollouts to be paused etc. I’m also not saying that is a logical response.
> I am saying that it will only take one fatal accident by any self driving car for people to lose confidence, investigations to start, rollouts to be paused etc.
Uber and Cruise are both great examples of this, but it seems like the effect is mostly localized to the company itself that has the issue.
Uber hit and killed a jaywalking pedestrian, resulting in their self driving tech being sold to Aurora. [0]
Cruise hit a pedestrian that was flung into the cars path that a human driver hit previously. This resulted in GM completely abandoning Cruise and their future seems foggy at best. [1]
[0]: https://en.wikipedia.org/wiki/Death_of_Elaine_Herzberg
[1]: https://en.wikipedia.org/wiki/Cruise_(autonomous_vehicle)#Su...
None of these products were as good when they started as when they had a billion users.
https://www.scribbr.com/research-bias/survivorship-bias/#:~:...
Is that really what you want to argue in front of investors? That you are going to be the next Facebook, Google or Apple?
There was also MySpace, Friendster, Altavista, InfoSeek, RIM, Nokia…
Facebook was the next Facebook with Instagram.
Apple was the next Apple with the iMac. Then with the iPod. Then with the iPhone.
Google was the next Google with Android. With Chrome. With Gmail. With Google Maps. With YouTube.
Google becoming the next Google with Waymo would not be a black swan event.
Google also had literally hundreds of failures and Android is not an amazing financial success by any means and Google still ends up paying Apple over $20 billion a year because people with money buy iPhones.
Google is not exactly known for its success rate getting products out of the door that aren’t ad related.
Give stuff for free, ???, profit. It's kinda worked for them, but not as well as Apple's strategy.
You can’t give hardware away for free profitably.
You also have hundreds of failures
https://killedbygoogle.com/
Including Google Fiber.
https://arstechnica.com/information-technology/2019/02/googl...
Google Stadia was disaster.
In the phone market. The Motorola acquisition was a major failure and Pixels aren’t taking the world by storm.
The entire “Other bets” haven’t led to any major successes.
There are only two tech companies that have shown any ability to do hardware at scale as mass consumer products in the last 25 years - Apple and Tesla.
This is all intended policy to benefit Trumps super donors. They can then scoop up marketshare and competition for pennies, then lobby to get the tariffs lowered or removed, but the higher prices - that we will be used to paying at the point this all comes together - will not go down.
All else being equal, companies are going to use the source of labor that results in the cheapest product they can produce. No one is forcing companies to move this kind of manufacturing to the United States. A 10% (let me reiterate that: TEN PERCENT) tariff on incoming goods is inflationary, but by very little, and quickly absorbed by companies and consumers. No one is moving their labor supply from China to the US to avoid a ten percent tariff; US labor is more expensive than that, and there are fifty other places around the planet you could find cheap low-skill labor that aren't on Trump's shitlist.
But you won't believe any of that, because you want all this to happen. You're a doomer; doomers and preppers secretly want the doom they predict to happen, even if they won't admit it to themselves.
For now they are going with the tide, because they don’t want to relocate head office.
If trump’s terrible plans are rolled back there’s no harm.
If they aren’t rolled back they will undergo the costly move to relocate.
The world is a bigger market than USA and just about every other country has cheaper labor and no tariffs.
It’s just logic, I have no emotion tied to this.
Your "logic" (masking) conveniently avoids the point I raised: That these tariffs are being enforced to the tune of 10%. That isn't enough to alone justify this level of investment, or relocation of significant production capacity. Obviously, Apple agrees with this, because the investments they're making aren't as far as I can tell a relocation of capacity from China to the United States, but rather greenfield investment in high-skill research and development. Apple has also made significant investment into advanced silicon manufacturing in the United States; something they also did not rely on China for previously.
> The world is a bigger market than USA and just about every other country has cheaper labor and no tariffs.
Have you done zero research into this? The EU imposes a tariff on Chinese EVs. India imposes insane tariffs on all imported electronics. China tariffs Australian wine. Russia tariffs agricultural products from the EU. Brazil tariffs all imported automobiles. The list goes on. Tariffs are everywhere, everyone uses them for something.
Until things are actually built, I take press releases like these with a grain of salt. Similar to the stories about Mark Zuckerberg removing tampons from men's washrooms the week before the Presidential inauguration, I believe that a lot of these stories are intended for an audience of one.
Will they remain competitive with other big players producing in cheaper countries?
Who are those "other big players" when it comes to building a complete AI stack?
Who can compete with Tesla and Waymo without having to buy hardware and software from the US?
Who can compete with OpenAI, Google, Grok, Perplexity without US hardware and software?
Deepseek seems to prove there's no super secret sauce that makes these models irreproducible outside the US and that the companies here are suffering a bit from the glut of cash/credits leading them to burn tons of extra processing power that could have been optimized away.
DeepSeek has been trained on US hardware.
US hardware in what sense? My understanding was it was trained on spare compute owned by the parent company but I didn't bother looking that deep.
A nice model, does not a billion dollar company make. The hard part of AI is not the model; Apple needs people to do the 80% rest-of-the-work; how do you make AI useful to the average person? How can we get inference on edge devices as cheap and efficient as possible? Models are boring. Everyone fully expects that we'll see an N% increase in intelligence every six months now. Yawn. The exciting thing now is: What are we using AI for?
Spending a couple billion dollars also doesn't make an actual billion dollar company. It's yet to be proven that all this spend on LLM training and running can actually get translated into an actual profit.
The Chinese?
It sounds flippant but the US hasn't got some unassailable moat around chip production or their application.
It's not a small feat. China would have to:
1: Take over Taiwan to get TSML under their control
2: Find a way to make a deal with ASML for the needed lithography machines
3: Somehow aquire the knowhow that Nvidia has
Taiwan would require a conflict with the US. ASML is a dutch company but seems to be somehow under US control. I have not yet figured out the exact setup there. And Nvidia is a US company.
China has steadily been developing its own processes and the laser tech needed for smaller sizes.
It is very clear that both US and China are racing to onshore and establish sovereign control of chip production.
Sucks to be Taiwanese right now. The "silicon shield" is evaporating.
Citation needed.
I have yet to see anything that hints at China being able to catch up with ASML+TSMC+Nvidia.
They are trying, and will eventually catch up, the same way they have in software and in hardware in many other spaces. Maybe it will take another 10 years. Maybe another 10 months.
This idea that EUVL is—and always will be—outside of the reach of China is, frankly, silly. It's a silly strategy to maintain dominance. They will straight up steal the technology if needed.
We should stop pretending like we can roadblock the technological development of the largest country. It's just going to make the fall that much harder. Once they do attain the ability to manufactuer EUVL domestically, capital is going to flood out of US tech stocks like no tomorrow.
You haven't seen anything that hints at it, because most news about what happens in China doesn't leave China. The propoganda machine is hard at work reminding you that China is terrible and eveyrone in China is poor. You'll learn about China catching up only after it's already happened.
With the combination of very real geopolitical risk (which was a topic of considerable discussion at a tech conference I attended late last fall) and the current political climate, there's a significant mindset that the US should be pulling back a lot of things to its own borders where practical even if not optimal at the moment.
> Who can compete with Tesla
BYD says hi.
Tesla makes a lot of promises that it can’t keep and losing money and market share globally.
I doubt Waymo is going to be a big deal in much of the US over the next decade. Even if they do figure out all of the technical issues. People will accept hundreds dying from car crashes. But not one dying from autonomous cars.
That what tariffs will do
20K people is a drop in the bucket.
It won’t happen. The supply chain is far too complex. Not to mention that the labor market in the US is not willing to do a lot of the work that you see in China and isn’t large enough even if there were enough willing people.
And then you have the rare earth minerals that aren’t available here.
Work will be increasingly done by robots.
Which raw earth materials could be a show stopper for the US?
That was a typo above (since edited). I meant rare earth materials.
And why hasnt that happen yet instead of Foxconn employing 770K people?
These are the raw materials used by the iPhone
https://www.cnet.com/tech/mobile/the-metals-inside-your-ipho...
I don't get why folks keep saying x86 linux servers here for AI, if anything it'd be M series arm based servers, either running linux or macos. Realistically I'd imagine a set of scaled up mac mini arm servers for running inference or fine tuning on them as more likely being the "ai servers" than x86 based anything. Power is the key thing that they'll be optimizing for, and that's where ARM shines.
Don't they need gpus (for training)? Apple already did a footshoot wrt gpus in the apple ecosystem. unless they have some sort of apple-internal ai chips ready to train models.
Apple's Private Cloud Compute is on racks of M4 chips which have NPUs and GPUs on-die and unified memory access to however much RAM they want to put on them. All of a sudden they're competitive with NVIDIA, but they don't let anyone else use that platform.
Apparently right inline with their plans: https://daringfireball.net/linked/2025/02/24/apple-500-billi...
Something buried in this, Apple are starting their own server manufacturing.
I would assume it is for their own use, not for people wanting the XServe to come back with M4 Ultra (i.e. me...).
Nvidia has proven the space is incredibly lucrative and Apple is best equipped for high end chip designs. Remember 10 years ago it was unthinkable for an ARM chip to compete with x86.
First Apple has to prove they have competitive designs. Apple Silicon GPUs simply do not compete with the efficiency of Nvidia's GPU compute architecture: https://browser.geekbench.com/opencl-benchmarks
Apple's obsessive focus with raster efficiency really shot their GPU designs in the foot. It will be interesting to see if they adopt Nvidia-style designs or spend more time trying to force NPU hardware to work.
I think performance per watt is way in Apple's favor, but raw performance is not.
That said, an M4 Ultra (extrapolating from Max and Pro) would likely compete with my 3090, and with 192GB of memory (for 10x the amount it should cost) will out perform my 3x3090 AI server. And honestly, cost less than my 3 3090s + rest of the computer + electricity.
It won't outperform a bunch of A/H 100s (or even a single one, or any other cards in the enterprise realm) though, but it will cost an order of magnitude less than a single card.
Careful when comparing performance and efficiency. As a rough factor power increases quadratically as you increase clocks on a design, so you can quite easily make a high performance design low power by under-clocking it. The same is not true for the reverse.
I think you are comparing apples and oranges.
inference is not the same as training.
Agreed, but I wonder--given investors demands for continued growth--if they're considering going up against NVIDIA.
Nah, surely 80% margins for matrix multiplication on the latest TSMC node will last forever.
Such a shame. I'd be interested in a personal server.
Why would Apple want you to have one though?
That's not necessarily news, unless I am missing something. Craig made an indirect mention of this during last year's WWDC regarding the private cloud compute.
https://security.apple.com/blog/private-cloud-compute/
That’s not completely news: https://www.macrumors.com/2024/05/06/apple-building-m2-ultra...
(This comment was originally posted to https://news.ycombinator.com/item?id=43158187, so "this" was a different article. We merged that thread hither.)
And in the United States, too. (Houston)
Does this mean competing with Asahi to run a Linux kernel, or will this be an attempt to run AI workloads on XNU?
Consider the cost of GPUs, losing what could be double digit percents on overhead might not make this very competitive. The macOS microkernel can still beat NT in some situations (like not having filters slowing filesystem down to a crawl), but it lags significantly behind the investment in Linux performance over the years by every other major company.
I believe the intention is to use their own M-Series CPUs - to get what they call "Private Cloud Compute". The cpu on your phone will encrypt data and a request, send it over the network to am M-series CPU which will decrypt and process/send back an encrypted response.
The idea being there's no VMware, kernel or piece of hardware that can have backdoors built into unless someone files off the top of the chip and somehow probes the silicon
> Apple said that it, together with Foxconn Technology Group, will later this year begin producing the servers that power the cloud component of Apple Intelligence — a system called Private Cloud Compute — in Houston. That marks a relocation, at least for some production, from overseas. Next year, it says a 250,000-square-foot facility for such manufacturing will open in the city.
> The Private Cloud Compute servers use advanced M-series chips already found in the company’s Mac computers. Those chips themselves, however, continue to be produced in Taiwan.
I don't think it's AI servers for Apple silicon. I think it's just regular x86 Linux servers to power Apple's AI cloud services. It's a commitment to internal investment rather than a product.
FWIW, Apple’s AI “private cloud compute” is Apple Silicon-based, and it’s a core part of some of the security guarantees offered. See [0].
[0]: https://security.apple.com/blog/private-cloud-compute/
I'll honestly love to see an Apple Silicon based XServe and comeback of "macOS $VERSION Server" add-on, and maybe an XSAN box.
Doing a Mac Mini sized version for SOHO would be great, too.
One can dream, I guess...
Apple doesn't like B2B and Steve Jobs was very vocal on this (there are various videos where he explains why). Ever since they can afford to, they reduced their B2B to the bare minimum needed. So don't expect anything server-like from Apple.
Yeah, I know, this is why I added "One can dream, I guess". :)
Funny, because I think Apple could do quite well in this realm.
I'm not sure. Because the moment you enter the system room, the ecosystem is a completely different universe.
You can't easily sell "Good / Better / Best" version of a single model, and tell "These are the options, take it or leave it". Servers are customized to the screws they come with and are expanded throughout over the years. So, the logistics are somewhat different for these kinds of devices.
Plus, macOS is not a CLI first operating system for server operations, and macOS Server is not updated for some years. Allowing Linux would be a different offering, and allowing macOS to work with all kinds of hardware from ordinary Ethernet to 100G+ Ethernet and 400gbps Infiniband (plus all the other interconnects) will be a fun exercise in testing flexibility of both macOS and Apple development teams.
So it's quite complicated. All servers are built to order SKUs. Dell keeps configurations "per server" in their databases, for example. If you have a Dell server, enter its service tag to support site, and you'll get the configuration of the device as it left the factory.
They don't have to go all the way though and fully compete with Dell EMC/HPE. I'm not sure what the original commenter was thinking but in my mind they could simply sell a Mac Studio variant with dual PSUs, better networking, a rackmount chassis, etc. Basically have their existing consumer machine placed in a more datacenter friendly factor.
I mean places like Github and AWS are painfully racking up Mac Minis for their deployments and this theoretical server model would simplify everything. It also becomes an option for on premises AI inference using MLX, especially if they manage to get ANE support working in conjunction with the GPU for faster prefill.
The support and software stack for the server model would be the exact same as the consumer variant and they certainly wouldn't have special Linux offerings, Infiniband, and all that. If there's networking beyond their existing 10G it's going to be built into the board and they aren't going to support random 3rd party cards. The unit also doesn't need to be upgradable either.
Quite a lot of cash to generate AI stickers in messenger.
this is it. siri will finally be reliable
"Four years ago, a few months after President Joseph R. Biden Jr.’s inauguration, Apple announced an “acceleration” of its U.S. investments, pledging to spend $430 billion and add 20,000 jobs over five years. In January 2018, during Mr. Trump’s first term, the company said that its “direct contribution to the U.S. economy” would be $350 billion over five years and that it planned to create 20,000 jobs over that period. Apple did not immediately respond to a request for comment."
Is this just lip service? What happened to those previous investments?
they are all counted as part of this 500bn?
This happens all the time with large investments in $X that make for impressive sounding press releases. If you (can) dig into the details, invariably a lot of the money is in previously committed/spent allocations in a whole bunch of different buckets (or, per bombcar's sibling comment, money that may never be spent at all).
I think the question is "have they done anything" or do they just keep promising to do something someday?
Will it be collocated with the promised Foxconn plant?
Mark my words: it will end up like the Apple Car.
Apple “Intelligence” has been a flop.
I can't tell why you're being downvoted, we have quite a laundry list of protracted investments at Apple leading to nothing worth shipping. Given how much of a nothingburger AI has been up to this point, I have zero hope that Apple will succeed at commoditizing a zero-demand service.
This is an insane amount of cash to throw at any problem. They could have Apple rockets mining asteroids with $500,000,000,000. There is no way all this cash goes into AI. What will actually happen is they will take 1/10th this cash to an over-valued startup and acquire them.
Apple is just giving a PR win to the administration like they did last time their preferred party was in power.
Between land, hardware, having to build multiple power plants, the labor costs involved in all of this, and setting money aside to run all of those for however long, then yeah I can see where it gets up to that price. 20k engineers is easily 5 billion per year in salaries, probably more.
The maths are not adding up.
Assume the human resources at $100,000 per head, and you get $2B/yr. Four years comes to just $8B for human resources. Assume land costs $10B. Assume construction costs $100B.
100+10+8 is $118B
Where is the money actually going?
Are they going to abandon H-1B and focus on hiring Americans for this project?
It’s a con to con the con …
> Four years ago, a few months after Mr. Biden’s inauguration, Apple announced an “acceleration” of its U.S. investments, pledging to spend $430 billion and add 20,000 jobs over five years. In January 2018, during Mr. Trump’s first term, the company said that its “direct contribution to the U.S. economy” would be $350 billion over five years and that it planned to create 20,000 jobs over that period
Anyways, the land (obviously in Texas) is already purchased and has been sitting empty. The unbuilt factory keeps getting more expensive though.
Do we think Apple will once again sell servers to customers?
I guess they could sell servers to customers who want to run the latest Apple Intelligence models on-prem, even though that probably wouldn't make much of a difference, since you probably still have to trust Apple.
I'm pretty sure this is for their own infra. FTA:
> Apple said that it, together with Foxconn Technology Group, will later this year begin producing the servers that power the cloud component of Apple Intelligence — a system called Private Cloud Compute — in Houston. That marks a relocation, at least for some production, from overseas. Next year, it says a 250,000-square-foot facility for such manufacturing will open in the city.
[...]
> Apple will also expand data center capacity in Arizona, Oregon, Iowa, Nevada and North Carolina, all states with existing Apple capacity. The company confirmed that mass production of chips started at a Taiwan Semiconductor Manufacturing Co. facility in Arizona last month. Bloomberg News recently reported that plant is building chips for some Apple Watches and iPads.
They will sell servers to customers ... as an add-on to iCloud.
Apple now offers access to Apple Intelligence Pro at 9.99$ a month.
I think they might build a cloud offering. Something like Cloudflare workers but AI centric, perhaps running Swift on the Apple equivalent of V8 isolates.
Makes sense from a business perspective - there's significant growth potential for them as their presence in web tech is approximately nil.
The Apple TV box (2021) is $129 with 3 GB RAM and 8 GB flash. Not hard to see where that could go.
1u boxes or Mac Mini "servers"?
It would make sense for Apple to fork their next highest-end Mac Studio motherboard, make relatively minor changes to it (e.g. add a higher bandwidth NIC and strip out unnecessary I/O) then wrap multiples of those into a rack mount chassis, with commodity-grade cooling and power supply solutions appropriate for the context.
Combined with a properly headless fork of their OS stack (think Darwin, not OS X Server) they could spin up a highly competitive solution using entirely "B-team" resources.
...then it would be piped through their design-council, run through 5 more iterations to get a unique unibody case for it, accompanied by an optional proprietary Apple rack and a price-tag triple of the competition.
That's along the lines of how it usually rolls whenever Apple tries to make something purely utilitarian, it's the most considerate and "fresh" look at a product, but ultimately designed to be used and then disposed when finished.
A purely utilitarian IT-appliance without a individual end-user doesn't seem to be possible in their product pipeline, you usually end up with something "Prosumer": Impressive on its own, yet of degraded maintainability and scalability.
It's like asking Bugatti to design a public transport bus. It would surely be an impressive bus, but not one you would want to maintain over years at a scale of hundreds.
Lots of interesting things Apple could do with their resources.
A modern version of the Xserve RAID for high speed flash storage could be very interesting.
The Mac Mini could be used as small blades.
or they could do something really wild, like take the Oxide rack scale approach and make something big for DCs.
But they might also want to get a piece of the prosumer homelab market that Ubiquiti is in?
That is a con to con the con …
:) its like the dot com boom. How many around here remember those days?
Good to see but brings to mind the deal they cut with China in 2016 https://news.ycombinator.com/item?id=29482351
Oh god. Hopefully they’re lighter than the old xserves. We had one still running up until a few years ago when we finally removed it. You could break a toe if it dropped while pulling it out of the rack. People are still selling them on eBay.
I suspect that was the RAID drive bay (ridiculous item). That was a 3-4U monster.
The Xserve, itself, was a 1U unit that was pretty much the same (or lighter than) any other 1U server (we also had HP and Dell servers that were heavy). The weight distribution could be weird.
That stupid drive bay was a proprietary nightmare. The disks cost a fortune.
However, from what I can see, this will be for "internal-use" servers. I don't think they will be selling iron; just services run on the iron.
I don't think it's AI servers for Apple to sell. I think it's just regular x86 Linux servers to power Apple's AI cloud services. It's a commitment to internal investment rather than a product.
They have hiring positions for running a Darwin-based server OS, and their private cloud compute is on Apple Silicon. I doubt it's going to be swaths of x86.
I'm curious what they will look like, given that these are not for anyone else to buy. Maybe Apple made a different form factor/configuration that suits their datacenters better?
Hey quality stuff is heavy!
The desktop case for my 286, I could stand on it and it would not bend!
The sentiment in this thread. Surely Apple has a fiduciary duty and so no reason for FUD.
Reminder: https://www.theverge.com/2019/7/30/20747765/apple-mac-pro-20...
Apple has been sitting on a war chest the size of Jupiter for a long time. Glad to see them putting it to good use.
expect of course they won’t… few months into every new administration (see 2017, 2021…) they’ll make a splash announcement like this… and then wait for the next administration to make it again
A strong majority of Apple's revenue comes from outside the US. If Apple is tied to US protectionism, it is precisely the sort of org that suffers.
Now to be real, Apple has been announcing these multiple hundred billion dollar investments every four years like clockwork. They did it early in Trump's first term, they did it again early in Biden's term, now they're doing it again. But for all of the "Yeah, this is what tariffs are for! Hoo ya!" rhetoric, note that companies like Apple have far, far more to lose than what America "gains" by acts like this.
strong minority you mean :)
https://www.statista.com/statistics/382175/quarterly-revenue...
Texas is big tech's choice to skirt employee protections. I'm sure these are the type of jobs, similar to Foxconn per the article, that Americans are looking forward to.
I know a lot of tech workers in Texas, specifically in the Austin area. They seem to be doing very well. I'm quite proud of America's working conditions. A lot of workers in other countries would marvel at our opportunities and be grateful these investments are happening here as opposed to elsewhere.
>A lot of workers in other countries would marvel at our opportunities and be grateful these investments are happening here as opposed to elsewhere.
"In other coutries"...any particular ones?
I can’t think of any developed country where that’d make sense.
When Texas flips blue like California did way back it's going to be interesting
No thanks
Nah, watch what happens when California goes red, if Dems repeat their 2024 performance trends.
Unlike Texas, California thankfully hasn’t been gerrymandered to death. It still has a functioning democracy.
California's housing crisis is a result of small-c conservatives wanting their property values to rise forever. Prop 13 and it's consequences have been a disaster for the state.
Prop 13 is a disaster, but it was not the product of any specific political persuasion.
It was even well-intentioned, but the unforseen (although predictable in direction if not magnitude) consequences are in fact disastrous.
I have wondered about how to repeal Prop 13, but I can't come up with a repeal that doesn't do equal and opposite damage.
While that’s certainly a key component, Texas is also home to the largest potential solar and wind capacity in the country. There’s also a ton of land to build on.
And Austin was an island of sanity according to my friends who lived there.
At this rate, Texas is going to go the way of it's hat and ban "woke energy" like solar and wind. [0]
[0] https://heatmap.news/plus/the-fight/spotlight/renewable-ener...
Unlikely. Texas is the largest exporter of crude oil and natural gas, the largest in capacity for refining petroleum, and huge exporter of petrochemicals in the US. More solar and wind means more oil for refinement and export.
I'm in Dallas and doing just fine tyvm.
I do find it slightly offensive that you would insinuate that hiring in Texas is solely about less worker protections and not that we have plenty of skilled workers in one place and with a lower cost of living.
I'm a worker not an owner,and I prefer living in Texas at this stage in my life and have turned down offers to move back to California.
Working in tech in a big Texas city easily puts you in the top 10% of the cushiest jobs in the US. Based on how I've been treated here, I really doubt that worker protections (or lack thereof) is the real driving force behind more tech jobs moving here. We are far from being oppressed here.
The most likely driving force of tech moving to Texas is that mid career professionals like myself don't see a future in California due to the insane cost and bad vibe of raising a family there. It's a great place for people just out of college, but Texas is a better place to settle down unless you are pulling an outrageous salary. The other big advantage California has is VC and startup networks being located there, which is also something that primarily benefits early career people rather than those of us that need a stable job at an established company.
It's also worth pointing out that Texas has long had a large technology industry presence. The dominance that California experienced during the early 2000s through to mid 2010s is an outlier and it shouldnt surprise anybody that things are evening out.
what can people outside US and china do?
Depending on where you live, be thankful for a functioning democracy?
IMO we've really got to start pressuring our own governments to take control of their networks, as well as the companies the population is going through (not just Facebook, but even international contractors for services). Letting a foreign government have this much control over the data of your populace and the ability to feed whatever algorithmic message they like is a path bound for disaster in the long run. The powers of the world are way too consolidated as is, and a company can turn into a state actor at the drop of a hat. I don't think we can maintain kayfabe about the country/corporate divide. I also think this can be done without impacting freedom of speech for your population, as long as you don't consider corporations people.
Most countries don't have the resources to do much, but even then they can try their hardest not to be beholden to any single foreign country coughChina, Americacough.
Time to move away from Apple before they train their AI Models on the rest of my data that they left untouched
There's a lot of understandable skepticism about this announcement because there are so many PR announcements. I want to temper that with an alternative perspective.
I actually think Apple is a potential dark horse when it comes to AI hardware. What we have now is essentially a three-layered monopoly: ASML, TSMC, NVidia. This has been incredibly lucrative for NVidia. But, as we know, Apple doesn't like to rely on third-party hardware. They've invested heavily in ARM going back to buying PA Semi [1]. Apple replaced Intel chips (which originally replaced Power chips) with the M series in recent years. Apple is in the process of replacing Qualcomm modems in iPhones, which is not only a technical feat but a legal one given Qualcomm's patent dominance over 4G/5G.
Apple has the resolve for long-term initiatives that few other companies have. Apple Pay continues to chip away and get slowly better in a way that, if it were a Google product, would've been cancelled, rebranded, relaunched probably 3-4 times by now (Google Checkout, Google Wallet, Google Pay, Android Pay, etc).
Apple clearly sees AI as a strategic issue. They have loads of cash on hand to finance basically anything they want. And they won't want to be beholden to NVidia.
I expect Apple to have a significant impact here but it won't be tomorrow or even this year. It'll be over the next 5-10 years.
[1]: https://www.wired.com/2008/04/four-reasons-ap/
Apple's incentives have definitely aligned with replacing Nvidia entirely ever since they ceased diplomatic relations. But Nvidia also knows this, which is why they invest heavily in things Apple will never do. They write the official Linux drivers Apple wouldn't get caught dead supporting. They give users and integrators freedom to choose their OS, software and library stacks to better suit their application. They sell individual GPUs and unlocked edge compute hardware with no distribution terms or $99/year "developer license" bullshit. Nvidia is a hardware company in places where Apple tries shipping services instead.
Then there's also the software issues. Nvidia has invested in GPU-based compute nonstop for the past 10+ years. Apple invested in Nvidia, then invested in OpenCL after abandoning Nvidia, then abandoned OpenCL for Metal compute which would eventually become the proprietary Accelerate framework. Nvidia's eggs are all organized in one, valuable basket. Apple's investments are spread out all over the place, with much of the time and money going into projects that don't even exist anymore.
Apple has the TSMC advantage, but that's just about it. Their GPU designs aren't comparably efficient or compute-oriented to what Nvidia ships today. Additionally, Nvidia will continue investing in places that Apple principally refuses to support. Unless a serious tide change occurs at Apple, they aren't going to get a fair competition with Nvidia.
Some part of me thinks they are billing an overbudget here to report that they actually didn't need to spend that much so that they can beat lower guidance (same play for MS, Google, Meta). We've heard that actually training these models doesn't cost even a billion dollars.
Official release: https://www.apple.com/newsroom/2025/02/apple-will-spend-more... (https://news.ycombinator.com/item?id=43158187)
Cynical take here, but I think realistic.
There's a lot of noise I can see behind the scenes on investor confidence. Noise as in "everything is fucked" sort of level of noise. Thus I expect this is being said to try and stop the AAPL stock collapsing in the upcoming recession that the analysts are predicting more than a tangible expansion and recruitment goal.
I also take issue with their being 20,000 people on the market who are still able to contribute something useful. They will be culled quickly and quietly down the line in the annual corporate lay offs.
It is not the time to make grand gestures unless you're trying to gain political favour, at which point any respect I have at least is gone.
Market opening today should be interesting...
All for the purpose of the AI servers that will power Apple's own private AI models for training and inference running on Apple Silicon.
Still in a AAPL long position since $170 and waiting for the re-introduction of Xserve with a new Apple Silicon-based OS. [0]
Probably "aiOS" (Just guessing).
[0] https://news.ycombinator.com/item?id=40278371
> waiting for the re-introduction of Xserve with a new Apple Silicon-based OS.
I nearly tripped over a skeleton holding up a sign with the same sentiment walking into HN today. I expect you'll be waiting quite some time...
Watch them lay off 5k highly paid employees in order to create 20k low paid jobs.
I’d be happy in the 20k bucket honestly, the job market is terrible
[dead]
Reverse technology transfer from countries like China is kind of fair. But EU companies should be very wary of Trump tariff blackmail that forces them to build production lines in the US.
So the EU has no tariffs?
Not a lot of them, but Trump accuse EU of tariffs because it has VAT. That’s not the same thing because VAT apply to domestic companies as well so there’s no real unfair advantage. You could make the claim that it’s still worse for foreign goods as VAT can be offset with VAT from supplier purchases, but that only works if you produce goods in the same country as you sell them. But US companies have even larger advantage already as they don’t have any VAT, so in the end the playing field is kind of level.
This doesn’t stop Trump from pressuring and squeezing and making false claims to justify his standing point. Expect an unstable trade environment due to tariffs and retaliation. This will hurt small businesses that have even more complex environment to operate in.
> they don’t have any VAT
VAT is what Americans call "sales tax" and US companies do pay and charge that.
>But EU companies should be very wary of Trump tariff blackmail that forces them to build production lines in the US.
That's the whole point of tariffs, to encourage domestic production.
Put another way, what is the difference between what you wrote and
>But US companies should be very wary of EU tariff blackmail that forces them to build production lines in the EU.
?
Whether a particular tariff is economically viable is a reasonable debate. Calling Trump's tariffs "blackmail" without assigning that epithet to all tariffs from whatever source is not.
If only they could spare a few dozen devs to update the command line tools to the 21st century.
Please explain how this will increase the stock price in the next week, otherwise it will not be prioritized
:(
Yeah, the real stock market mover is the mouse cursor. That wiggly thing they did that grows the mouse cursor really send ripples throughout Wall Street.
Given its value for aging users with weakening eyes, that tracks :)
to Tim Cook, yes, but not to Steve Jobs.
I don't think Steve Jobs would care any more about command-line MacOS tools than Tim does
Steve Jobs would lean into making Apple the #1 AI developer platform and showcase at WWDC how the terminal is now obsolete
He wasn’t stupid. He’d observe their own developers and see how they rely on the terminal and command line for their work. He’d ask pointed questions and demand thoughtful answers.
Then he’d find a way to make it the #1 AI developer platform or distort reality until it is.
Or you can take the extra minute to install them yourself with brew, this a complete non issue for anyone that understands the command line in the 21st century using MacOS. Also, I would never build anything against macos userland because it's almost never the target.
Brew takes a minute just to update itself, let alone install anything.
And then everything needs to have the /opt/brewsomethingsomething PATH. /bin:/usr/bin:/usr/local/bin not good enough.
MacPorts is The Way.
In the big '25? Nixpkgs is the New Way.
So what? Do you complain when apt, yum, dnf, pacman, ports, or any other package management system does a download? I bet you don't, so it's not really a usable argument. Secondly, yeah, not tainting my systems OS and system paths is a good thing and opt/ from the filesystem perspective is the absolute right place to put add on packages.
these two are orthogonal to each other.
Desktop Linux is right here, desktops Linux has always been here!
I don't understand why people complain about Apple neglecting developers when desktop Linux provides a superior experience aside from the rare times you need to compile an iOS/Mac specific application.
> desktop Linux provides a superior experience aside from the rare times you need to compile an iOS/Mac specific application
or reliably use peripherals
Wow, did you travel here from 2002?
Edit: In case it's not clear from my initial, gut-driven snark: I definitely think if you use a reasonably popular distro (commercially backed or not) in 2025, you should never have any trouble connecting peripherals to it, with the possible exception of Bluetooth, which I hear also applies to macOS.
Nope, I came directly from an MS Teams meeting where the desktop linux users had no audio.
I'm on Windows, and my work laptop has no audio via its headphone jack.
I have no idea how this is supposed to even work, and since it's not my computer I don't mess around trying to install drivers. I just use my phone to call in for Teams.
Things happen, let's not act like any OS is perfect.
MacOS is definitely not perfect. I'm being snarky. But it has been my anecdotal experience as both a user and observing colleagues that MacOS is more reliable and stable for desktop use than Linux. This is unsurprising since it's easier to build a stable walled garden than an open ecosystem.
Macs are generally more reliable, but if you buy a year old ThinkPad Linux will be just as stable .
The only issue Linux really has is when new chipsets come out you might need to wait 6 months or so for the drivers to be updated. But to be completely fair, on one of my laptops I had no webcam support for like six or seven months until Windows update decided to finally install it for me.
If you need a significant amount of hard drive space, Macs are almost always exorbitantly expensive. I make music so I find myself dual booting between windows and Linux. I don't want to speed 3k+ on a MacBook just to get a 4TB SSD I can add to any Windows PC for 200$.
Plus on Linux you can customize your personal experience to a much greater level. If you dislike X,Y,Z you can disable it or find an alternative.
Both OSX and Windows are cramming so much monetization into the OS, there's a very real feeling that I'm just sharing my computer with a giant corporation rather than actually owning it.
It’s less convenient when you are on the go, but you can pack an external SSD and offload stuff to it. A friend of mine had one velcroed to the back of the screen.
It's actually cheaper to own a MacBook Air for things that need to work 100%, like a coding interview, and then a secondary laptop when you're playing video games or making music .
That's basically what I do now, my old M1 MacBook air is more than good enough for LeetCode and I'm more or less know it's never going to fail.
> But it has been my anecdotal experience as both a user and observing colleagues that MacOS is more reliable and stable for desktop
You mean a few month after a new MacOS version has shipped and they've got time to fix all the bugs it introduced, right?
I haven't personally experienced that problem. Updates on Mac have always been smooth for me. But I'm a sample of one and it's probably workflow dependent.
MS Teams audio issues aren't the best example -- ask anyone forced to use Windows
I came from meet where none of the macs could screen share due to recent OS updates :)
Last year I was using a Windows laptop for work and Teams was very unreliable with audio and video. And don’t even think of using the nice camera on top of the expensive video conferencing monitor on my desk.
Presumably these users have audio in other contexts? Are they running the web app version of teams? Do other web apps play audio? From 10000 feet up, I wouldn’t start by blaming Linux here (even as a non-Linux-desktop user,)
This isn't a hill I want to die on, but isn't it the case that even if the problem is in MS software compatibility with Linux that still results in desktop Linux being a less reliable platform for day to day use?
Depends on the day to day use. I have experienced a lot more pain on Windows than on any other platform. Perhaps HP-UX or AIX 3.x were more painful.
The inability of the people you work with to use their devices means almost nothing. It’s as if you said nothing.
to be fair Teams barely works on Windows
Team barely works, period. It’s almost a feature, actually.
Tons of peripheral devices do not work well or reliably on Linux, and I literally cannot remember the last time I have had ANY issue with Bluetooth on macOS. Certainly not in the last decade.
I don’t remember the last time I had an issue with Bluetooth on Linux either. Most likely before 2010 or so.
The last time I did was this morning. I get dropped connections constantly, microphone not working in Teams (solved by reboot), pegged connections preventing handoff, etc.
> with the possible exception of Bluetooth
Good thing no one uses that!
I went through 5 distros a month ago dealing with fractional scaling issues on my 4K monitors. Decided it is not worth dealing with a went back to macOS so… No.
I feel like I spent a significant chunk of the mid 1990s editing /etc/X11/XF86Config in search of the right incantation...
I very recently tried again to adopt Linux on the desktop. I'm really sick of feeling like a frog in a pot of water. It's becoming harder and harder to bypass their literal gatekeeping of which applications I can run on my computer, and with every new version of macOS the temperature in the pot keeps rising.
The main problem I have with living in a Gnome desktop environment, is with the keyboard. I'm not willing to abandon my use of Emacs control+meta sequences for cursor and editing movements everywhere in the GUI. On macOS, this works because the command (super/Win on Linux/Windows) key is used for common shortcuts and the control key is free for editing shortcuts.
I spent a day or so hacking around with kanata[0], which is a kernel level keyboard remapping tool, that lets you define keyboard mapping layers in a similar way you might with QMK firmware. When I press the 'super/win/cmd' it activates a layer which maps certain sequences to their control equivalents, so I can create tabs, close windows, copy and paste (and many more) like my macOS muscle memory wants to do. Other super key sequences (like Super-L for lock desktop or Super-Tab for window cycling) are unchanged. Furthermore, when I hit the control or meta/alt/option key, it activates a layer where Emacs editing keys are emulated using the Gnome equivalents. For example, C-a and C-e are mapped to home/end, etc.
The only problem is, this is not the behavior I want in terminals or in GNU/Emacs itself. So I installed a Gnome shell extension[1] that exports information about the active window state to a DBUS endpoint. That let me write a small python daemon (managed by a systemd user service) which wakes up whenever the active window changes. Based on this info, I send a message to the TCP server that kanata (also managed by a systemd user service) provides for remote control to switch to the appropriate layer.
After doing this, and tweaking my Gnome setup for another day or so, I am just as comfortable on my Linux machine as I am on my Mac. My main applications are Emacs, Firefox, Mattermost, Slack, ChatGPT, Discord, Kitty, and Steam. My Linux box was previously my Windows gaming box (don't get me started about frog boiling on Windows) and I'm amazed that I can play all my favorite titles (Manor Lords, Hell Let Loose, Foundation) on Linux with Proton.
[0]: https://github.com/jtroo/kanata
[1]: https://github.com/hseliger/window-calls-extended
I don't get it either, it takes twenty minutes to burn a USB stick and run the installer. It takes me longer to remove the bloatware and set up my preferred settings on a proprietary OS than it does to install Linux nowadays, and that's been true for a decade at least now.
It's just not hard! It's not more work! And yet the meme about it being more trouble just. won't. die.
You people are supposed to be technologists! Why won't you spend 20 minutes of one time setup to get a better experience?
Not all hardware works? I don't see anybody complaining about having a limited set of hardware options when they buy Apple! Canonical maintains a list of fully compatible computers; just pick one, buy it, and you wind up with a computer just as easy to use as Mac OS but without the endless paper cuts of using a system that has no respect for you at all and thinks it knows better
Most of us have .dotfiles, I can snap any macos installation into my preferred configuration in about 5-10 minutes unattended depending on internet speeds. I do most of my work in a terminal, as long as that works, I'm good on Linux, MacOS, and BSD's. They all have pros and cons.
I only wish Apple’s terminal supported Tektronix and ReGIS. And provided a sane implementation for sixels.
I use Apple laptops primarily for the hardware. But Linux has never really been a great experience on Mac laptops when it comes to battery life, reliable suspend/resume, etc. etc. I used to use Yellow Dog Linux on a G4 PowerBook way back in the day, but I haven’t had much luck with Linux on Mac hardware since then.
I had an old i5 Mac mini laying about I wanted to use desktop Linux on the other day. The last time I tried, was about 20 years ago. I note nothing has changed since.
I have one of those running the latest Fedora. It’s a great workstation.
I use macOS because it's the only OS to reliably support Apple's Macbook M series.
(Unless someone wants to correct me.)
If you are paying for Apple hardware, it’s silly to use an OS that’ll make it only less reliable and compatible.
Desktop Linux is an oxymoron. i’ve tried it many times and every single time I went back to macOS.
Any particular reasons?
I made this. It makes it easy to use all of the most common gnu tools via brew, without having to do gsed for sed, etc... all with working man pages. It also lets you switch back easily in a shell session if you need the mac native ones for some godforsaken reason:
https://github.com/kilna/gnu-on
You mean how the Unix standard tools on Mac are way behind Linux?
So are the BSD tools by some definition of "behind". Another way to look at this is to say that GNU tools as typically seen in modern Linux are bloated (I know, Linux and "bloat" are kind of a meme, but it is generically true for the most part when it comes to the command line utilities feature creep over the years, so it's a useful and descriptive word).
I have to work with old machines and legacy operating systems quite a bit in my day to day and I always am going to prefer something lighter and with less ways to shoot myself in the foot w.r.t. POSIX compliance. MacOS is Unix certified so I appreciate them being somewhat reserved in the features they add on top of POSIX.
Modern GNU userland utils are nice and fun but if you are looking for compatibility it's best not to use them. Consequently, the MacOS situation doesn't bother me especially given you can install more up to date tools if you want. I think keeping the defaults older and more compatible is a good thing.
There is poor cross-UNIX compatibility if you're doing anything complicated, anyway. I maintained a large test suite for about a year that was written in POSIX sh and targeted Linux, macos, {Free,Open,Net}BSD. It wasn't fun because every program behaved in slightly different ways, half of them undocumented (for example, I remember having lots of pain with how different versions of tail handled SIGPIPE).
In the end it was was easier to rewrite in Perl than to keep maintaining that thing, struggling for hours to find ways of implementing every little bit of functionality that worked reliably on every OS. You'd add or fix something, and the tests would break on FreeBSD. You would fix it there and it would stop working on NetBSD. And so it goes.
This is true about cross Unix compatibility. I can still dream though.
What tools do you prefer?
I prefer a lot of the BSD variants of the typical POSIX tools (i.e. bsdtar vs GNU tar, ksh or similar instead of bash, etc etc). Usually because they add less extensions on top of what is required by POSIX, but are still easily acquired in a modern Linux distribution. I mostly just alias them.
If I write a script using BSD esque tools I can be reasonably sure they will work on any Unix-like, whereas if I write/test my script on a machine using GNU utils, I'm fairly likely to accidentally use a GNU extension that would cause the script to fail on an older Unix-like OS. For instance, I do a lot of work migrating code off of AIX,and I need the scripts I develop to work on AIX when I'm gathering environment information from customers. I can't just assume they will have a ~2020+ implementation of Unix userland tools with all the GNU extensions and nice features. Sometimes the machines have been sitting quietly in the back of a data center not being updated for quite a while and will have more "90s style" of Unix tools.
For a simple version bump, I feel like brew is fine. Or do you have other tooling updates in mind?
Many people here are commenting that brew is a good way to get modern tools. I must say, I prefer Docker by several country miles.
It’s like having the tools in a different computer. You can mount your local filesystem onto the container, but if feels like WSL - there’s always an “impedance mismatch” between the two sides.
I prefer to use the tools running locally on the same OS I’m working with. For that, MacPorts is great.
Apparently there is a 20,000 surplus somewhere waiting to jump in.
I heard they have strong Nordic developers waiting for free markets in Greenland.
I bet the Panamanian Ai industry masses are also sitting on the edge for this.
Maybe $500 billion will finally be enough to make Siri useful for more than just setting an alarm.
Funny I turned Siri off because i didn't want apple intelligence running amok. The follow-on problem --> lack of Siri killed my Carplay because Siri is required (also use itf for setting alarms/timers). The kicker? I can't seem to turn Siri back on after look through all the menus.
I.e. My preference for apple CarPlay supersedes my concerns on GPT running over my contents. Though the UI/UX has made it next to impossible to turn it back on.
What a world to live in.
Apple Intelligence and Siri are still separate (though Apple like to make it look like they are fundamentally intertwined). You can turn Apple Intelligence off and leave Siri on for CarPlay.
How did you turn Siri off in the first place? That's where I'd start...
Settings -> Apple Intelligence & Siri:
Talk and Type to Siri -> turn all this back on
Allow Siri When Locked -> turn this back on
The part of Siri that causes the most trouble is the speech recognition - which uses a voice recognition model that we now colloquially refer to as "AI." The part that works reliably, the part that sets your alarm or sends the message, is an action that's hardcoded.
IMO, moving towards AI just leads to increased uncertainty and undesirable outcomes, which is something several journalists reviewing Apple Intelligence have attested to.
If your phone is new enough for Apple Intelligence, Siri is now under that umbrella. There's no "just Siri" option anymore, unless you're rockin an iPhone 14 or older.
Or a non-Pro iPhone 15.
I would pay $500bn to get siri to distinguish between a 13 minute timer and a 30 minute timer.
I am very resourceful so I got around this by setting a 14 minute timer to outsmart the Ai.
"Your 40 minute timer starts now".
Workarounds: 781 seconds timer, or 29/31 minute timer.
I’m serious, that’s what I use.
When I put in timers -- for some reason my timer frequently/randomly just sets to 79 hours and a random assortment of minutes and seconds. I have no clue why. I always have to double check otherwise I might be waiting awhile.
It feels like it was a residual timer or something but I have never set anything like that - it is quite strange.
I solve this by asking for a 31, 41, or 51 minute timer.
How does this help me se a 13, 14, or 15 minute timer?
Could you ask to set a 780 second timer?
You can, and it works.
"10 minute timer" + "add 4 minutes"
I don't know if this is an actual problem you have, but since Siri appears to be composed of independent voice-to-text and text-to-action systems, you can say "start a one three minute timer".
The problem is AI current best use case is creative work, art, music, programming, but skilled creative professionals is a/the core userbase for Apple products.
Apple is stuck and it’s AI will never be good enough until those creatives embrace it. Right now it’s disdain when mentioned.
> The problem is AI current best use case is creative work, art, music, programming
This is where it’s being pushed and marketed but I’m not actually sure it’s the best use case.
> skilled creative professionals is a/the core userbase for Apple products.
Then why doesn't it have a professional CAD application?
Even Apple doesn't do CAD on OSX. They run Siemens NX on Windows.
But your statement isn't quite right. Fusion 360 runs fine in mac. I'm ex-Apple btw.
Ok, that's good to know, but my colleagues all use SolidWorks so it doesn't change much for me.
It does.
AutoCAD came to the Mac when Intel was shitting the bed (with aggressive OEM contracts for first party system integrators that prevented AMD adoption across HP/Dell/Lenovo-lines) and Windows 11 was being forced on users.
WINTEL played the monopoly game too hard and is starting to lose ground.
You love to see it.
Uhh
> The problem is AI current best use case is creative work, art, music, programming
By “best” do you mean “marketable?”
Seems weird to see a bunch of creatives and creative professionals “disdain” a tool and still say it’s “best” for them…
An oft-cited quote goes something like this: "we wanted robots/AI to automate boring, routine, meaningless jobs to let people be free to pursue arts, music, creativity. It's a sad state of affairs that AI is taking over arts/music/creativity stranding people with boring, routine, meaningless jobs"
> It's a sad state of affairs that AI is taking over arts/music/creativity stranding people with boring, routine, meaningless jobs"
So far it’s not though.
Oh. It already is. Artists are already saying that a lot of commission work is drying out (e.g. illustrations).
[dead]
Even that seems to work only half the time. ~50% she just doesn't respond to a voice command "Siri" or "Hey Siri" for me.
When Siri first debuted it would automatically beep, so I could immediately tell if the phone did not recognize recognize "Hey Siri" (just "Siri" didn't work). A couple of iOS updates later this went away, which means I can't tell without actually picking up the phone and looking at it whether the command was accepted.
Even more annoyingly, sometimes there is a beep! ¯\_(ツ)_/¯
Yup, the UX has gotten so much worse.
Whoa, whoa, whoa, on occasion, if the planets align just right, I can also get Siri to set a reminder (and at least half the time Siri gets it 80% right).
LLM Siri cannot come fast enough.
> LLM Siri cannot come fast enough.
I don't think it's been demonstrated that Apple could make Siri better with an LLM.
If you haven't tried OpenAI's advanced voice mode, it's a mind blowing version of exactly what things like Siri really ought to become with a little more development. If that's what you mean by LLM Siri, I totally agree.
Being able to chat casually with low latency, correct yourself, switch languages mid-sentence, incorporate context throughout a back-and-forth conversation etc. turns talking to these kinds of systems from a painful chore into something that can actually add value.
But does it do actual stuff, like adding a meeting to your calendar, call people, or set a timer or a reminder?
LLMs are there for generating/hallucinating text, not for understanding text.
For natural language processing you need a different kind of neural network don't you?
It's the other way around. The model is impeccable at "understanding text." It's a gigantic mathematical spreadsheet that quantifies meaning. The model probably "understands" better than any human ever could. Running that backwards into producing new text is where it gets hand-wavy & it becomes unclear if the generative algorithms are really progressing on the same track that humans are on, or just some parallel track that diverges or even terminates early.
I thought it quantifies the probability that a certain word (their output) follows a given word sequence (their training corpus and the prompt)?
Only if you wildly oversimply to the level of being misleading.
The precise mechanism LLMs use for reaching their probability distributions is why they are able to pass most undergraduate level exams, whereas the Markov chain projects I made 15-20 years ago were not.
Even as an intermediary, word2vec had to build a space in which the concept of "gender" exists such that "man" -> "woman" ~= "king" -> "queen".
> Only if you wildly oversimply to the level of being misleading.
Maybe I'm asking for an explanation :)
Since you seem to understand the mechanism, can you do a 3 line summary please?
3 lines? That's still going to be oversimplifed to the point of being wrong, but OK.
Make a bunch of neural nets to recognise every concept, the same way you would make them to recognise numbers or letters in handwiting recognition. Glue them together with more neural nets. Put another on the end to turn concepts back into words.
For a less wrong but still introductory summary that still glosses over stuff, about 1.5 hours of 3blue1brown videos, #4-#8 in this playlist: https://youtube.com/playlist?list=PLZHQObOWTQDNU6R1_67000Dx_...
Thanks!
... Oh interesting. And those concepts are hand picked or generated automatically somehow?
> For a less wrong but still introductory summary that still glosses over stuff, about 1.5 hours of 3blue1brown videos
Sorry, my religion forbids me from watching talking heads. I'll have to live with your summary for now. Until I run into someone who condensed those 1.5 hours in text that takes at most 30 min to read...
> Oh interesting. And those concepts are hand picked or generated automatically somehow?
Fully automated.
> Sorry, my religion forbids me from watching talking heads.
What about professional maths communicators who created their own open sourced python library for creating video content and doesn't even show their face on most videos?
My problem is with the time wasted compared to written info, not with talking heads per se.
He doesn't waste time. No fluff.
You're unlikely to get a better time-quality trade-off on any maths topic than a 3blue1brown video.
He's the kind of presenter that others try to mimic because he's so good at what he does — you may recognise the visuals from elsewhere because of the library he created[0] in order to visualise the topics he was discussing.
[0] https://docs.manim.community/en/stable/faq/installation.html
There's also a playback speed slider in YouTube. I use it a lot.
Simplifying to that point is more of what a Markov chain is. LLMs are able to generalize a lot more than that, and it's sufficient to "understand text" on a decent level. Even a relatively small model can take, e.g. even this poorly prompted request:
And the most likely response will be what the user wanted.A Markov chain (only using the probabilities of word orders from sentences in its training set) could never output a command that wasn't stitched together from existing ones (i.e. it would always output a valid command name, but if no one had requested a reminder for a date in 2026 before it was trained, it would never output that year). No amount of documents saying "2026 is the year after 2025" would make a Markov chain understand that fact, but LLMs are able to "understand" that.
I’m confident that LLM’s will not have hallucination problems in the type of requests that I send to Siri.
I don’t ask Siri for facts (just like I don’t ask LLM’s for facts). As long as it can correctly, understand what and when I ask to be reminded about something, that would be a huge improvement for me.
That and being able to map “Bedroom Fan”/“Bedroom Fan Light” to “Bedroom Fan Lights” without having to specify aliases (and even then it hearing me wrong).
I’ve see Home Assistant working with LLMs and it can understand groupings that I never explicitly defined which is very nice. I can say “Turn off all overhead lights” and it will find all my overhead lights and turn them off. Siri/Alexa can’t handle those tasks currently.
What are you talking about? We've invented the Universal Translator already.
Over on Android it's the opposite situation. The voice interface to Google Assistant was very reliable for simple things like reminders and appointments, and even for general knowledge questions. It was part of why I didn't switch to an iPhone. Then Gemini came along, and that core functionality got a lot worse.
Siri already has LLM integration...
...that will grind your request to set email Vacation Mode through the world's worst speech-to-text, jam the text into Chat GPT, and spend the next three minutes reading you an uninterruptible 3 minute essay about violence.
I'm a formerly non-mac guy who finally bought a brand new iPad. I got bored with the wallpaper but couldn't figure out how to change it. "Hey Siri, how do I change the wallpaper?"..."Sorry, I can't help with that". Tried a couple more questions and all it did was Google it for me. This is the latest M4 that was around $2k.
This is what our "AI accelerated" chips give us in return? What a disgrace
hey thats what alexa is for
No, Alexa is only good to interrupt you randomly when you're watching TV.
Apple software is already largely written my Americans.
Somehow I don't think fealty will change its quality.
Doing gods work, thank you.
I detached this comment from https://news.ycombinator.com/item?id=43158188 so we could pin the latter to the top. What you posted was fine! I just don't want to take up extra real estate at the top of the thread.
[dead]
[dead]
[dead]
[dead]
lol. lmao.
After an oligarchy, looks like the us is also getting a planned economy now.
We used to call this an "Industrial policy" or an "Economic development policy". Back in the golden era when a strong labor movement coexisted with a Red Scare. 78 years after Taft-Hartley and 44 years since PATCO, not so much.
We have maybe fifty or a hundred million people rotting away in areas where jobs are scarce and housing is plentiful, because we used government policy to shut them out of areas where jobs are plentiful and housing is scarce. We systematically exported jobs from places that aren't big cities because they can be performed overseas and our aristocracy can still profit from them by owning those people overseas.
I don't know if returning to a little more deliberate of an economy is even a partial salve for the place we've found ourselves, but I don't think this laissez faire thing is sustainable for a whole lot longer. We are overleveraged, and arrogantly delusional about our sway at the moment; "Ownership" is not some valuable skill. The fall of an anchor currency and global conversion to an alternate financial network would be a spectacular thing, an astroid striking terrain, which might leave craters on entire other continents from secondary ejecta. World wars have been fought over less.
"We" wasn't big government. It was a million homeowners who decided that the neighborhood they moved into should be frozen in amber forever. Everyone wants housing to be cheap but also for their property values to rise onto infinity. They push back against any attempt to change this and then complain about the inevitable results.
> now
I guess before is before Christopher Columbus
‘Planned’. A command economy anyway.
?
A planned economy has a plan. I doubt that will exist.
A command economy has different elements of the economy ordered around to do what leadership wants.
That seems a lot more likely.
Is this a joke that I'm not getting? "Planned economy" and "command economy" are synonyms.
Say what you want about Trump, this is the kind of deal that wouldn't have happened with any of the previous administrations, both Democrat or Republican. It's the kind of deal that keeps MAGA loyal to him, despite all the noise about DOGE.
https://www.apple.com/newsroom/2021/04/apple-commits-430-bil...
Never would have happened with any of the previous administrations...
What keeps MAGA loyal is only consuming news that do not bring up when previous administrations did such kind of deals.
They already did it once in the previous administration, so i guess that argument is out the window.
Got any others?
Apple better hope none of their customers realize how comparatively mid of a product iphones are by the time those servers are ready.