PhantomHour

The wider context here is that the TikTok ban had significant support on the grounds of, what politicians and Zionist lobbying groups called, "anti-Israel bias" and "support for Hamas". Not just the explicitly stated "China Bad" motivations.

https://www.nbcnews.com/tech/social-media/tiktok-ban-israel-... (Note: Article from November 2023)

The opposition to TikTok on grounds of it's Chinese ownership had been on a slow burn right up to October 2023, when it picked up steam in the wake of the early response to the Gaza War. US politicians were furious that the youth weren't buying the Bipartisan Approved Position(TM) on Israel.

Considering that major world organisations, even holocaust remembrance ones, are now calling Israel's actions genocide, that fury has aged like the finest bottle of raw milk.

Hence:

and the implication that their intent surely must be to issue pro-Israel propaganda

That is indeed the implication made, for the reasons above, I don't think it is unwarranted.

Waymo doesn't gain anything. Google i.e. Alphabet Inc, does.

Especially these days. Every scrap of news that could pump the stock price is publicized aggressively.

And this makes the absence of such actions suspicious.

Not really sure I understand the economics here

There is nothing to understand. The point of such subsidies is to turn OPEX into a green line on the stock market.

Especially as Microsoft is currently also in a fight with OpenAI.

You'd be surprised. A lot of sellers don't "cash out" from paypal all that often, letting tens of thousands pile up. (And inevitably, some of them get hit with arbitrary account closures and have that money seized)

This is a baseless assertion of emergent behaviour.

Every time a bot searches

We are talking about LLMs by themselves, not larger systems using them.

LLMs ability to separate facts from expression is quite well developed

It is not. Whether you ask an LLM for an excerpt of the bible, or an excerpt of The Lord of the Rings, the LLM does not distinguish. It has no concept of what is, and what is not, under copyright.

Well, all a judge can/should do is to apply current law to the case before them

This is true, and I do not mean to suggest it is bad. But rather, that it leaves uncertainty. These cases can all be struck down without reducing the possibility that if one does stick, the entire industry is at stake.

Copyright infringment is all about having published something based on someone else's work - AFAIK it doesn't have anything to say about someone/something having the potential to infringe (e.g. training an AI) if they haven't actually done it. It has to be about the generated artifact.

A notable problem here is that AI models are not "standalone products" but tools provided as a service. This complicates the situation.

Take Disney/Universal's case against Midjourney, which is both about the models but also the provision of services.

Even if only the latter gets deemed illegal, that's ruinous for the big AI companies. What good is OpenAI if they can't provide ChatGPT? Who would license a LLM if the act of using it creates constant legal risks?

It is not.

While this likely has no legal weight

I wouldn't be quite so sure about that. The AI industry has entirely relied on 'move fast and break things' and 'old fart judges who don't understand the tech' as their legal strategy.

The idea that AI training is fair use isn't so obvious, and quite frankly is entirely ridiculous in a world where AI companies pay for the data. If it's not fair use to take reddit's data, it's not fair use to take mine either.

On a technological level the difference to prior ML is straightforward: A classical classifier system is simply incapable of emitting any copyrighted work it was trained on. The very architecture of the system guarantees it to produce new information derived from the training data rather than the training data itself.

LLMs and similar generative AI do not have that safeguard. To be practically useful they have to be capable of emitting facts from training data, but have no architectural mechanism to separate facts from expressions. For them to be capable of emitting facts they must also be capable of emitting expressions, and thus, copyright violation.

Add in how GenAI tends to directly compete with the market of the works used as training data in ways that prior "fair use" systems did not and things become sketchy quickly.

Every major AI company knows this, as they have rushed to implement copyright filtering systems once people started pointing out instances of copyrighted expressions being reproduced by AI systems. (There are technical reasons why this isn't a very good solution to curtail copyright infringement by AI)

Observe how all the major copyright victories amount to judges dismissing cases on grounds of "Well you don't have an example specific to your work" rather than addressing whether such uses are acceptable as a collective whole.

Whether or not it's "essential" is kind of irrelevant for the ergonomics; Languages do exist that provide only closures. (And yes, the likes of Java do struggle a bit with function pointer ffi) Similarly, "we need this because low level" has never made C++ more tolerable.

The thing about rust is that it's complexity is self-contained and follows established rules. If one understands data ownership, one only needs to be told what a closure is for all these edge cases and problems to make sense.

Contrast the complexities of JavaScript, many of which still boil down to "Some doofus 30 years ago didn't do any homework and made bad design decisions for type coercion". Just "fuck you, go memorize these behaviours".

It's not strictly about the money. (Though it is absolutely also about that)

Dealing with unexpected job terminations (fired or laid off) is the problem.

Herein lies the problem. This gives employers absolutely massive leverage over the employees, which lets them coerce things like ridiculous unpaid overtime and downright abuse.

Even if you pay the same nominal salary, the H-1B is "cheaper" if you can force them to work 60-80h whereas a top-class American is just going to demand 40h weeks. (Though in practice, those extra hours rarely see increased productivity, so whether it's actually cheaper for outputs obtained is up for debate.)

Contrast: Europe. Tech salaries are low by US standards, but you don't see as much of the outsourcing & migrant worker hype around it. European labour laws mean you can't set up a sweatshop in your branch office, and European migrants to the US won't put up with labour abuses as much.

But every article I've read on this makes no statement about whether any of these hundreds of people were actually working without an appropriate visa

This is because ICE is being particularly tight lipped about those details.

The New York Times got their hands on the records for 11 detainees. 6 on B1/B2 visas. 4 on 90-day waivers. 1 Unknown.

ICE claims visa violations, but the records do not state what work the detainees were actually doing. This is especially relevant for the B1 visas, which do permit certain business activities (including applicable ones for this situation; Meetings, trainings, "installation, service, and repair of foreign-bought machinery".

Of particular note is that in one case (out of these 11), ICE's records state there was no visa violation. The worker was deported anyway, forced into a "voluntary" departure.

(https://www.nytimes.com/2025/09/12/business/economy/hyundai-...)

Personal opinion: The degree to which hyundai may or may not have violated the law or operated within previously-tolerated gray areas remains to be seen. But the actions of ICE here are not those of a competent government organisation.

There should be clear records, they should be able to readily answer press questions. And yet they don't.

Worse still is that one person deported despite there being, even by ICE's own admission, no visa violation. Hard to assume good faith in incomplete or withheld records with such shit going on.

And what are other foreign companies to do with this? "Move your manufacturing to America! Oh btw even if you follow all laws to the letter a local chud may deport your workers for being not white enough and ruin the entire project" is an interesting sales pitch.

Do they think they're above it? Are they stupid and don't know what they vote for?

They're somewhat out of touch with tech, and caught up in police narratives around encrypted apps blocking their attempts to find pedos. Tech firm lobbyists sell them some lies about the capabilities of these systems.

Ultimately these are politicians stuck in the notion of "but the police can open your [physical] letters, this isn't any different" completely unaware of how times have moved on.

Matters like how people are already being harassed by CSAM being sent to their DMs, how people raid discord servers and try to have them taking down by spamming CSAM, etc, are completely lost on these politicians.

On top of that it's just cowardice. Not daring to be seen as "aiding pedos".

5:10 "475 were illegally present in the United States"

Not quite; The figure includes those with visa.

"Illegally present in the United States or in violation of their presence in the united states, working unlawfully"

This also leaves open the question of what "violation of the visa" entails here. They may well have been working within their visas, only for ICE to arbitrarily rule otherwise.

I hope the recession ends up being smaller in length and magnitude than the 2008-2009 recession

Not a chance. Were it just the tariffs, the recession would be quite small. The tariffs might even be cancelled if the SCOTUS remembers to do it's job.

But it's not just the tariffs.

The US economy is currently being kept afloat by AI R&D and infrastructure spending. It's stock market kept alive by 7 companies who are all neck deep in AI hype.

This not only disguises the malaise in the rest of the US economy, it's a bubble. Everyone knows it. Nvidia's the only one making any money and even they are now relying on vendor financing and other such red flags. Even one who believes that the technology of AI is here to stay, has to face the reality that it's not a golden goose of infinite dollar bills.

We're looking at something that's going to be at least as bad as the Dotcom crash. 'At least' because while the bubble is of only comparable size, other conditions are much worse.

Trump is trying to seize the fed. Big Tech is tearing the copper wiring out of it's own walls to keep AI going a little longer, and their plans for cutting costs is to dramatically increase H1Bs and outsourcing. (One wonders if there might be a non-economic reason behind this, given it's one specific country they're seeking to hire from >.>)

And underneath it all: A timebomb. Much, much, much more of consumer spending in 2025 is from pensioners than it was in 2000. When the stock market eats a 50% loss and stay there for a decade, those pensions will be cut dramatically. This drives down consumer spending, in turn driving down the stock market, a vicious cycle.

That's not the point under contention. Removing the de minimis exemption is a perfectly cromulent policy. It's not even particularly unpopular to remove it.

The problem is that the Trump Administration is plainly incompetent in handling these matters. There wouldn't be this shitshow of sudden haltings in postal services to the US if they'd done the normal thing of announcing the changes with an appropriate lead time for businesses to adjust, rather than suddenly implementing it alongside the constant ping-ponging between yes-tariffs and no-tariffs.

To be slightly rude, there is just a wikipedia article by the name "Dumping"; https://en.wikipedia.org/wiki/Dumping_(pricing_policy)

The actual legal mechanics are complicated; "Illegal under international law" here specifically entails "WTO agreements allow retaliation in response to dumping".

and why I couldn't do it for cars or commodity goods.

Specifically, it's more enforced. Governments care about their conventional industry. The way this'd look is say, China providing state subsidy to certain industries in order to artificially lower the price of those goods, making them cheaper than US-based industry could produce, with the specific intent of driving US industry out of business.

Just googling "predatory pricing" and "dumping" will get you examples.

Also I think that I doubt how enforceable this is in tech industry as for the most part, they are selling a service and each service is different and thus have different price points and therefore the company should have the ability to decide prices technically.

The problem for tech is this difficulty in assessing "real value" and the assumption that running at a loss for extended periods is "normal" for tech companies.

For a clear-cut example, consider Uber, who paid drivers more than they charged the passenger(s). This is obviously predatory. Uber has tricks like moving insurance/maintenance to the driver's wallet, but a taxi can't be cheaper than what they pay the driver.

why does the govt. interfere b/w them? Does this not impact their rights/freedom?

It does impact their freedom, but the reason why the government intervenes is long-term health of the market.

Things like a 'firesale' because you're going out of business, or moving to a new warehouse, etc, are fine. A single store (even a big-box one) going out of business won't crush the entire market and it's only of short duration.

The problem is that dumping/predatory pricing is a strategy to maintain a monopoly. (Or in the cases of extensive investment funding, build one)

Again, consider something like Uber (but the same applies to any "rental"/gig-economy company). They sell rides below cost paid for by their huge pile of investment money, no other taxi company can compete. All the competing taxis go out of business. Uber can now raise the prices to obscene levels and cash in.

Whenever someone tries to start a new taxi company, it'll be small and local, so Uber just lowers their ride prices in that region again until they go out of business. And because they're small they don't have as much money as Uber so they'll go bankrupt first. Uber keeps the monopoly.

Such monopolies are long-term bad for the entire economy.

On an international level, it's China and steel again. China subsidizes their industry, industry in other countries can't compete and goes bankrupt, China can now raise their prices.

"Dumping" in the context of international trade; Predatory pricing.

The standard model for tech firms has been to run at enormous losses to push competition into bankruptcy or steal their users through subsidized service.

No European social media company could compete with e.g. Twitter, running at a loss for TWELVE years.

In more recent years, it's things like Uber. Subsidizing ride costs to crush existing taxi services & European taxi startups.

This is all, ostensibly, illegal under international law. You can't do it for cars or commodity goods. It's just not been enforced on the tech industry.

There are no major tech companies in Europe, which is so insane it's comical. Let that sink in...a continent full of intelligent tech workers has never been able to get a major tech company off the ground.

This is plainly untrue if you're talking about tech beyond the mag-7 sized supergiants.

Regulation may be good, but understand, actually, recognize, that it is also suffocating. People bragging that they have no weeds in their fields, when they have no fresh crops either....

And yet it is the tech giants in the US, oh so praised for their size, that are the "weeds" in many regards.

What good is Google when it's reliant on an advertising monopoly itself built entirely on monopolistic and fraudulent exploitation of the rest of the economy.

What good is Amazon when it's reliant on crushing all other retail and local manufacturing?

The entire idea of "Oh they'll leave" is ridiculous, an empty threat from billionaires who are afraid of regulation.

The EU has 450M (+80M for UK & similar non-eu countries that are likely to follow the EU on such regulations) population to the US' 350M.

The moment the likes of Google, or Meta, or Microsoft, or whomever else leave the EU, they immediately create a market gap. A market gap that will then in short order be filled with a European company that, because of the population sizes, has a notable comparative advantage to the US tech company.

+ As much as HN's readership loathes to admit it, regulations like this are "Good, Actually". Google's monopolist practices are bad for both advertisers and services showing ads. Any would-be competitor that arises from Google leaving the market would, by virtue of being forced by law to not be so shitty, be the better option. (And yes, this does also apply to pretty much all of the other big tech regulations as well.)

Like, c'mon. "Monopolies bad" is capitalism 101. Even the US' regulators thought Google was going too far.

If they all have valid visas and are legal to work, I'm confident they'll be released

The Trump administration has been in court several times already for trying to deport people who, not only are not supposed to be deported, but whose deportation would and is be illegal.

Even if they're released, ICE is known to be abusive and inhumane in it's detention practices.

Your response reeks of "they'll be proven to be illegals soon enough, and then all the abuse is fair game". Disgusting.

Two things to bear in mind here:

1) As much as JD Vance and his puppetmaster wishes they were the president, they are not. The tech fascism isn't happening, just regular kind. What you're seeing here with the "deregulation" is just a big fat scoop of plain corruption. Give Herr Führer a nice golden trinket and he deregulates your industry or tariffs your competitors.

2) While they are deregulating heavily, there's absolutely zero sense of "Small Government" going on. You need only look at the oppression of groups like LGBT people. There's no bigger government than controlling which toilets people use or inspecting the genitals of children if they want to participate in sports.

And outside of LGBT rights, just look at the whole situation around Israel. Literal secret police going around arresting people for non-violent protest and simple thoughtcrime.

They have visas. Ones that permit specific kinds of work.

They are not "illegal aliens".

One particular datapoint to watch is auction prices for production capacity as these are the most immediate market. Those have jumped, with PJM's +800% in 2024 and +22% (22% from 2024->2025, $28.92 -> $329.17 or +1038% total over 2023 -> 2025) https://www.reuters.com/business/energy/biggest-us-power-gri...

We (Well, the article does it for us as well) can identify that this market jump is driven primarily by increased demand rather than lower capacity, as the amount of capacity sold has increased. (Covering ~50% of the new demand)

These prices do not immediately translate into residential fees, there's a lot of financial engineering in between the auction and customer contracts (F in chat for the counterparty of the hedge), but at the end of the day you can't sell $500 of power for $5 so the increased cost will worm their way through to residential consumers eventually. One particularly annoying thing is that this isn't a consistent amount of time by region; Some people are seeing 3x electricity bills already, for others their power company may have more strongly hedged against price increases.

Is that no longer the case?

Generally (but there are exceptions) in the US they have not gotten better. It's just that datacenter demand has eclipsed the effect of that.

My assertion is just that it is misleading to pin them as THE reason costs have gone up. It is far more complicated than that.

It's not so much that they are the sole cause, but that their effects (especially on local scales) are 1-2 orders of magnitude larger than all other causes, so people focus on this matter.

Consider, if people get the new housing developments that they want, that would also add load to the system.

It's a matter of scale and efficiency. Housing also adds (full-price) customers new taxpayers. A major problem with housing development is that low density (e.g. single-family zoned suburbs) are net-negative contributors. (But this effect is primarily seen with infrastructure like roads)

Any model you do that tries to prevent this is essentially rent stabilization for early members.

The problem with datacenters is twofold:

1) It's a demand-shock. Because of the rapid rollout of datacenters, energy production and transmission capacity simply can't scale up. This causes prices to spike locally.

2) 'Industrial' electricity like this tends to pay very cheap rates, leading to residential electricity customers subsidizing them. Especially as such industry tends to be given high tax & other incentives.

Any model you do that tries to prevent this is essentially rent stabilization for early members. And that has a pretty good track record of not being a good idea.

A question here is whether or not we're in a datacenter construction bubble. (To anyone who says we're not: Nvidia's stock price has soared more than Cisco's during the dotcom bubble.)

If you're in charge of a major electricity company, are you going to sign off on major expansion investments right now, knowing that it will take 5-10 years?

A lot of these datacenters are not owned and operated by Big Tech itself, but instead disposable companies like CoreWeave. If there's a bubble, it'll pop, and these companies will just go bankrupt. The power purchase agreement'll be worthless then.

The distinction is that you cannot make the LLM do the drilling. And the way these tools are designed is to train the user to use the LLM rather than their own brain, so they'll never learn it themselves.

A big problem with the "Just read the code" approach is that reading the code at the level deep enough to truly understand it is at minimum equally time-consuming than writing the code in the first place. (And in practice tends to be significantly worse) Anyone who claims they're reading the LLM's code output properly is on some level lying to them.

Human brains are simply bad at consistently monitoring output like that, especially if the output is consistently "good", especially especially when the errors appear to be "good" output on the surface level. This is universal across all fields and tools.

Just beware the “real programmers hand-write assembly” fallacy. It was said that compilers would produce a generation of programmers unable to understand the workings of their programs. In some sense, this is true! But, almost nobody thinks it really matters for the actual project of building things.

One of the other replies alludes to it, but I want to say it explicitly:

The key difference is that you can generally drill down to assembly, there is infinitely precise control to be had.

It'd be a giant pain in the ass, and not particularly fast, but if you want to invoke some assembly code in your Java, you can just do that. You want to see the JIT compiler's assembly? You can just do that. JIT Compiler acting up? Disable it entirely if you wish for more predictable & understandable execution of the code.

And while people used to higher level languages don't know the finer details of assembly or even C's memory management, they can incrementally learn. Assembly programming is hard, but it is still programming and the foundations you learn from other programming do help you there.

Yet AI is corrosive to those foundations.

What if I want to listen to an obscure book? Should I be paying a human narrator to narrate my personal library?

You weren't doing that before AI either, were you?

The practical answer has already been "you pipe an ebook through a narrator/speech synthesizer program".

The copyright holders are incentivized to make money.

Regulations exist. It'd be rather trivial to pass a law mandating every ebook sold to be useable with screen readers. There's already laws for websites, albeit a bit poorly enforced.

I'm sorry but I simply do not believe you that it does handle things like puns and things requiring footnotes, my experience is that LLMs are miserable at this even when directly "instructed" to.

But for what it concerns my previous comment: It doesn't really matter what the "state of the art" AI is because companies simply do not use that. They just pipe it through the easiest & cheapest models, human review (that does not actually get the time to do meaningful review) optional.

There's a big difference in precisely how the technology is applied.

Transformer models making screen readers better is cool. Companies deciding to fire their human voice actors and replacing all audiobooks with slop is decidedly not cool.

You can really see this happening in translation right now. Companies left and right are firing human translators and replacing their work with slop, and it's a huge step down in quality because AI simply cannot do the previous level of quality. (Mr Chad Gippity isn't going to maintain puns or add notes for references that the new audience won't catch.)

And that's in a market where there is commercial pressure to have quality work. Sloppy AI translations are already hurting sales.

In accessibility, it's a legal checkbox. Companies broadly do not care. It's already nearly impossible to get people to do things like use proper aria metadata. "We're a startup, we gotta go fast, ain't got no time for that".

AI is already being used to provide a legally-sufficient but practically garbage level of accessibility. This is bad.