To add to that, the ruling intentionally did not make a decision on two important distinctions:
1. Forcing you to unlock - in this case, Alisigwe unlocked their phone willingly after being asked.
2. Forensic extraction - the ruling does not give authorities the power to run a sophisticated off-site break-in on your phone without reasonable suspicion.
So if you are a US citizen, and you are willingly unlocking your phone at the border (or you don't have a passcode), the Fourth Amendment no longer protect you from a suspicionless manual search of the phone at the border
> That is not explained by increased deficit, but by net increase in demand.
A yield going up means you pay more for the same thing. So if the US wants to issue more debt, they can. The fact that more debt was bought but the yield went to means the supply grew faster than the demand. So an absolute increase in demand, but a net decrease, thus a higher price as shown by the yield
That is not true. Yield is going up globally, so you need to adjust for the difference in yield.
For example, the US and Euro (average) yield have gone up by almost the same amount in that period, and other currencies like Japan and Australia have experienced an even larger increase.
I’m not sure why the level of discussion in this post is so poor.
I’ll just note that NVidia’s moat has never been inference, and there are many chips used at a much larger scale than Chinese chips for inference like TPUs and AMD chips.
The other part is that it’s a bit of a meme here to say that the chip restriction is actually helping China (or shall I say, coordinated effort?). For once, we know that China has put a lot of pressure on the US to relax these controls multiple times. In addition to large chip smuggling networks (e.g. 22% of NVidia’s worldwide revenue magically comes from Singapore, and the ratio has been growing).
Lastly, assuming acceleration in AI (which we ARE seeing), there might not be time to China to catch up. The best estimate right now is that the first EUV chips from China will not come out before 2030. By that time who knows how powerful AI will be.
All I’m saying is that the discussion is so one sided and a bit baselesss with no nuance, that it seems either a meme/groupthink in the community or coordinated. If anything, the data suggests that the US should increase its export controls and better track the tech supply chain if it wants to further curb Chinese progress.
When you say AI will be powerful, what do you mean? Like in terms of national power, will AI let the US fight and win a war against China despite China's size, manufacturing base and possession of nuclear weapons? Will AI let the US beat the Chinese in manufacturing costs and scale despite China's greater adoption of industrial robots and larger number of skilled workers? Or will AI just be really good at writing software?
Because if the AI is just really good at writing software, I am not sure why China has to "catch up". Seems to me China can just treat AI like other technologies. Let the US pay most of the R&D costs, then come after and treat the technology as a commodity. Sell it better and cheaper and at scale.
I am not sure why it's a big deal for China if China is a few months behind the US in terms of the very frontier AI. Just like I'm not sure if it's a big deal which country has the biggest super computer in the world.
If AI is super human, it can beat any nation with today’s technology. It can recursively self improve while suppressing any other lab (think about how good it is in hacking computer systems).
I am not inventing a winner take it all scenario, that is what driving the race.
But, like, how does that smarts lead to beating nations with today's technology? How would an AI stop an incoming nuclear ballistic missile? I've heard on podcasts with MIT professor Ted Postol that based in the laws of physics intercepting these is basically impossible, because of the speeds and physics involved. Can AI discover new physics to overcome those limitations?
If AI cannot stop nuclear ballistic missiles, then how does AI give the US power over China? Seems to me that after AI we have the same mutually assured destruction we have today. In which case America cannot really stop China from progressing or threaten China too much.
Seems to me being super human intelligent does not inherently guarantee power. Just like the smartest person in the world does not always win in a fist fight or a gun fight or war. I.e. Stephen Hawking might have been smart, but he did not have physical strength and power. Power and smarts are not always related.
I just gave you a scenario above - that it hacks every computer system in China, including those controlling nuclear missiles. But a super intelligent AI can invent scenarios that almost by definition we cannot think of.
I would assume that the computer systems in China that control nuclear missiles, as well as our computer systems that control nuclear missiles, are air gapped and therefore not hackable no matter how smart the AI is.
This only demonstrates imposed limits you are able to imagine, not what’s possible in all scenarios. A super intelligent AI can bribe, social engineer and do anything it wants. No system is foolproof.
Historically civilizations that have been only at a slight technological advantage have completely dominated those that are slightly behind.
And we already see AI’s super human ability in computer systems. It will be able to be Omni present and completely control every computer in any place, and bribe and social engineer and mimic any person digitally. All while doing in an instant what a team of humans would require years.
I’m really surprised by the lack of logical thinking in HN when it comes to China.
Why do you think the Chinese are not using their ai chips, which afaics are from huawei, to train their models? If that is true, which I think it is, we don't have to wait until 2030 to see if they're gonna match the performance of nvidia GPU. The evidence is already here - glm is among the most competitive models out there.
Look up reports by The Information, DeepSeek’s training of its big models are all done on NVidia’s chips (more than that - on smuggled Blackwell as well). There have been attempts by many Chinese labs to train on Huawei h chips, but they have all been limited to more minor models or distillations of the bigger ones.
Unless you have the source for the latter it's nothing more than propaganda. Also, DeepSeek report is 2-3 years old, and at that time there were no huawei chips, at least not known to the public. This is a different model, in different age where huawei chips are already delivered into the production as we see
The idea that China wasn’t working hard to build their own chips is a political talking point period. Nothing has been sped up here, China was always going to do this and the backlash you see from people who are against it is only proof that restrictions do hurt them
I do believe your first sentence, but the second one ... doesn't really follow.
The unnatural and sudden nature of the US chip embargo has galvanized the Chinese companies to speed up development. It makes economic sense to do so rather than pay inflated rates to essentially smuggle a necessary good, while never knowing if the embargo will be tightened and those under-the-table supply chains dry up as well.
Yes, there was a plan, but what was planned for 5 years from now is being done today, because if you wait 5 years, you'll be so far behind thanks to the effect of the embargo.
It's an implied corollary: when folks say that "blocking US Chips is what led China to build it's own" they are implying it is US export restrictions that pushed China to innovate. OP is saying this isn't necessarily true. I don't think it is either: China seems to have succeeded due to a centrally planned economy that values independence, with high amounts of strategic government led technology investment.
> All I’m saying is that the discussion is so one sided and a bit baselesss with no nuance, that it seems either a meme/groupthink in the community or coordinated. If anything, the data suggests that the US should increase its export controls and better track the tech supply chain if it wants to further curb Chinese progress.
Or, to put it bluntly, you are a proponent of the trade war.
It’s very likely they all add svg generation into the training data. It’s part of the reason it’s no longer a good benchmark (unless you need to generate SVGs).
It's about more than just training data. Every model since GPT-2 had SVGs in the training data, because they all used a scrape of the Web and the Web is full of SVGs. What's unique about Gemini is that they actively worked to get better results for their SVGs - maybe RLHF, maybe RLVF of some sort.
Shipbuilding is but one proxy for China’s dominance in industrialization. A more relevant one for AI is: China has built more energy capacity in the last 4 years than the USA has built in the last 150. AI is increasingly limited by power. For this and other reasons, IMO the US will inevitably be surpassed on AI.
Last I've seen, it was compute and not solar panels that is limiting AI. And the US share of worldwide compute capacity has increased from 40% to 70% over the last 4 years, while China has shrinked to just 15% from 35% over that period.
> The coverage of Chinese companies is particularly poor. Our average coverage of 8 major companies is 15%.
The bigger issue is that if you're interested in frontier training, the national aggregate doesn't tell you much. The question is whether your top labs can get to the chip scale necessary to train a frontier model and China's labs have already demonstrated they can. See https://newsletter.semianalysis.com/p/deepseek-debates
But it's not even clear that training frontier models is where the real innovation occurs. When it comes to fine-tuning, deploying, and building AI-based products, inference is the name of the game and the chip quality is less important. Huawei's Ascend is already pretty competitive here and then you're also discounting China's open-weight access to good models, a huge base of engineering talent and a massive domestic market to iterate against.
And unless Anthropic, OpenAI et. al. are willing to invest billions of dollars training models that will only be used by a small number of clients (like the government), there's no way to fight distillation. Protecting any model from distillation means forgoing its full revenue potential, improvement via feedback and all the benefits AI is supposed to bring to the broad US economy. And if you restrict your most advanced models, you just push people to Chinese open-weight models, which is what we're already seeing.
Being sarcastic isn't going to solve the problem but you're obviously free to believe that the US is still in a position to call all the shots when it's evident to the rest of the world this isn't the case.
The only thing ridiculous here is thinking that the US government is in control to the level you think it is.
Spending $500+ billion/year (and close to a trillion now) on defense for decades and in ~6 months it has depleted stocks of critical weapons trying unsuccessfully to defeat a third-rate military power of a country that has been under sanctions for almost 50 years. And it can't even replenish them without Chinese raw materials and components.
Your sarcastic refutation of the allegedly ridiculous assertion was devoid of any counterpoints on how the US government can stop China's progress in AI and what you think they are going to do about it. Just repeating it doesn't make it any more or less ridiculous.
I think that the main thing people in this thread are missing, is that it's not about math. AI progression is very likely to affect every single thing humans can do today. Mathematicians are feeling the blow this week, especially as there was wide spread denial in that math community over the capabilities of AI over the last few years, but it's the same problem everywhere.
I work at a top tier academic institution and they are completely asleep at the wheel on this topic. They just blindly assume people will keep taking on a potential lifetime of debt when there is a genuine existential (for ed, at the very least… ignoring humanity for the time being) crisis looming
Yep. Perhaps humanity would be better off if we instituted and enforced the notion of "Thou shalt not make a machine in the likeness of a human mind." It's worth thinking about; just because we can build AI systems doesn't mean we should.
The problem with technology is that the people who have it conquer and kill all the people who don't, essentially ensuring a destructive race to the bottom. You would need a mutual international alliance of nations who not only renounce the technology but also agree to invade anyone who attempts to develop it. Our closest precedent is nuclear weapons, but rather than renouncing them, the US's position is that "only we should have them" and that other people acquiring the same tech they have are valid targets for invasion, and they will obviously apply that notion to any new technology that can be weaponised as well.
Google which countries have nuclear weapons. Hint: It’s not just the US, it’s also its allies and frenemies who’ve not been invaded. It’s only nutball countries like Iran that get invaded when on the cusp of acquiring it and for good reason.
I would remind you that the fictional society that did this was also described as a feudal monarchy in which a tiny class of nobles exploits the economic output of trillions of people reduced to serfs by controlling all means of interstellar trade. Oh, and they also have legal slavery.
During the Butlerian Jihad, the people involved in it didn't know that ten thousand years of slavery would be what they are choosing, so we don't know how bad it actually was.
The way Jihad presented in the oblique references in books, anyway, it sounds a great deal like a fanatical fundamentalist religious movement to me. Those are not known for their fine reasoning.
I’ve gotta say ultimately I’m not convinced Herb put more thought into it beyond its purpose justifying human labour in his universe despite the massively advanced technology. There’s a similar shortcut in the Warhammer backstory.
Herbert did genuinely believe that people overly relying on machines is problematic insofar as it "trains the mind" to think like a machine, and in the process loses the indeterminacy that he believed to be an inherent and valuable trait of human thinking that machines cannot replicate.
(Coincidentally, this is also the main threat in the original books, just manifesting in a different form, with prescient individuals "locking in" the rest of humanity.)
Honest question, is an oppressed human society better than no human society at all? Isn’t there at least a hope for humans to free themselves from human oppressors?
I'll remind you that a fictional product that provided longevity and woo woo powers of prescience and was only available on one planet in the entire universe was the key context that enabled such a centralized society. Dropping thinking machines and their masters like a bad habit was only part of the cause, as it eliminated one means of safe interstellar travel in that reality. I can take ideas from this plot without bringing in the whole ridiculous mess if I like, just like I can walk and chew bubble gum at the same time, thank you very much.
Of course they don't, but if you choose to use a work of fiction as an illustration, it comes as a coherent package (assuming the author did a good job at worldbuilding) - you can't just take the bits you like and ignore the bits you don't and still have something coherent.
I mean, we have a Millenium problem that is most likely solved. Either by OpenAI on its own, OpenAI stealing another's work and finishing it or a researcher using AI to iterate more rapidly than a human alone could.
> We should be happy that they’re stealing people’s work?
No. We should credit that work is being done that otherwise may not be. I'm not defending OpenAI. I'm defending AI in general, something both OpenAI and Buckmaster (and his co-author) used.
No, I did not. If someone says either X or Y happened, they didn't say X happened.
My exact statement: "Either by OpenAI on its own, OpenAI stealing another's work and finishing it or a researcher using AI to iterate more rapidly than a human alone could" [1]. (Either X, Y [a and b] or Z. An Oxford comma could have helped.)
So if Y happened (they stole it), why should we be happy about that? Why should we give them credit in the situation where they stole it? Why is that good for society?
Yes, let's lie and call something stealing, then pontificate about morality.
At worst what OpenAI did was inconsiderate, but not referencing prior work is a minor academic sin. This is especially the case when AI is not playing the vanity game that seems to so inflame some mathematicians.
Sure, we can show anything if we're allowed to make stuff up from personal outrage. Why should anyone else consider your personal opinion on this to have any weight?
And you seem to not realize that your unhinged argument impeaches whatever conclusion you're trying to justify. Good points don't require this sort of slimy misrepresentation.
The technology is fine, great even. It’s the rabid greed of venture capital and sociopathic CEOs that try to integrate themselves into every aspect of our lives for profit, that’s the problem.
“For 19.99 a month you too can be a world class mathematician!” Meanwhile they are extinct.
I am sure it will lead to more material wealth, but I can’t help but feel pessimistic about the political economy of making millions of people redundant. What leverage will knowledge workers have in the new world?
Well, yeah, while that is true, my impression is that mathematicians (as a community, even though there were outliers) didn't care - not even the slightest - about the impact of LLMs and generative AI, other type of AI and machine learning until it started impacting them. So I have no sympathy for their community at all. They brought all this AI dystopia on us, so they should have voiced their concerns much louder and much much earlier. They should also have shamed their peers who helped these companies much louder and they should have lobbied their institutions and their governments much earlier and much stronger. They should also have protested the direction these companies took at the exact moment when they started using AI for profit.
We can look at the past and wish we were more clear-eyed about what was happening, just as we can look to the present and ask that we have more empathy for each other going forward. But I don't think it helps anybody to blame the mathematics community for not organizing until they were directly impacted. This is a tidal wave that has caught everyone off guard in <3 years.
> mathematicians (as a community, even though there were outliers) didn't care - not even the slightest - about the impact of LLMs and generative AI, other type of AI and machine learning until it started impacting them. So I have no sympathy for their community at all. They brought all this AI dystopia on us
Wait wait wait. The research mathematicians brought this AI dystopia on us. Not the techbros, not Silicon Valley, not the global financial system and its tentacles, not surveillance capitalism... No, it was the postdocs working on ultrafilters for C(2,3) epimorphisms on symplectic manifolds. That's certainly a take.
Residential proxies are achieved in many, many ways. The most common is simply people installing software and leaving some checkbox checked, or agreeing to it unknowingly. In the same way that smart TV downloadable apps and games tend to include them.
If this was a regular topic on an American company so blatantly violating user data and lying, people would be pissed.
Unfortunately, hacker news is being so heavily astroturfed that I counted dozens of comments just justifying this behavior appearing very quickly as this post appeared on the front page, together with its position being clearly suppressed.
Mainly because it's not very helpful for a useful discussion. If it's true, the discussion is already dead, if it's false, all it does is annoy the accused.
For a specific post it’s always a guess. When it’s as recurring and consistent as it is here across all posts, it has crossed the point of preponderance of evidence.
And highlighting when it happens is exactly what helps people realize it’s not a one off.
Terminal Bench 4.0 did not introduce new questions. All tasks were public for a while. If you look at GLM 5.2, which is using the same base model as 5.3, but was released prior to most tasks, it does extremely horribly on terminal Bench 3.0 (4 to 8 times worse than every other model) - source: https://benchlm.ai/benchmarks/terminal-bench-3
1. Forcing you to unlock - in this case, Alisigwe unlocked their phone willingly after being asked.
2. Forensic extraction - the ruling does not give authorities the power to run a sophisticated off-site break-in on your phone without reasonable suspicion.
So if you are a US citizen, and you are willingly unlocking your phone at the border (or you don't have a passcode), the Fourth Amendment no longer protect you from a suspicionless manual search of the phone at the border
reply