Slashdot is powered by your submissions, so send in your scoop

 


Forgot your password?
Close
typodupeerror
China United States

US's AI Lead Over China Rapidly Shrinking, Stanford Report Says (axios.com) 66

Posted by msmash from the closer-look dept.
The U.S. is still the global leader in state-of-the-art AI, but China has closed the gap considerably, according to a new report from Stanford. Axios: Institutions based in the U.S. produced 40 AI models of note in 2024, compared with 15 from China and three from Europe, according to the eighth edition of Stanford's Artificial Intelligence Index, released on Monday.

However, the report found that Chinese models have rapidly caught up in quality, noting that Chinese models reached near parity on two key benchmarks after being behind leading U.S. models by double digit percentages a year earlier. Plus, it said, China is now leading the U.S. in AI publications and patents.

US's AI Lead Over China Rapidly Shrinking, Stanford Report Says More | Reply

US's AI Lead Over China Rapidly Shrinking, Stanford Report Says

Comments Filter:

  • Researcher asks for research funds (Score:3)

    by will4 ( 7250692 ) on Tuesday April 08, 2025 @02:12AM (#65288899)

    It's somewhat of a self-serviing article.

    Rank what is hard to rank, claim we're behind as a country, extend hand for research funds.

    We see this same exact template article for US and China aerospace research programs recently.

    • Re: (Score:2, Funny)

      by Anonymous Coward

      Yeah, and Soros is now the largest employer in the US.

      Also, denial still not a river in Egypt.

    • Re: (Score:3)

      by gtall ( 79522 )

      Extend hand for research funds? In the current U.S.? What do you not get about the Republican Party not believing in research? Apparently, for them, it grows just because it wants to.

      • Extend hand for research funds? In the current U.S.? What do you not get about the Republican Party not believing in research? Apparently, for them, it grows just because it wants to.

        It's not so much that Trump and his followers have anything against research. They just don't believe in research that contradicts what they already believe. Vaccines work, bad research. Atmospheric carbon causing climate change, bad research. Tariffs impede economic prosperity, bad research. The real problem is that they and DOGE are lazy and don't want to do surgical removal of the research they don't want. So, they've taken the private equity approach with a chain saw to all research funding.

        Well,

  • that's okay (Score:4, Funny)

    by SirSlud ( 67381 ) on Tuesday April 08, 2025 @02:14AM (#65288903) Homepage

    Americans have the most important thing in the world! Confidence, and an upbringing that, you know, stresses that they're the best! American hegemony is forever. It's ordained or something.

    • And look where that got you...
      Maybe there is a downside to this style of raising kids...

    • Re: (Score:3, Insightful)

      by JamesTRexx ( 675890 )

      stresses that they're the best!

      I always feel like that U.S. need to proclaim it anytime and anywhere is actually a sign of inferiority complex.
      A real confident person doesn't need to do that.

      • The US arrogance a sign of a fragile ego? I do not think so. They really believe it. The greatest, the largest, the best,....
        They actually are in a lot of fields where money is involved. Of course, there is a trade off there. Can't be good at everything. That part is hidden under the rug though.

        • Re: (Score:2)

          by gtall ( 79522 )

          U.S. arrogance starts out as a belief among the proles. As you go up the political ladder, it gets used by the pols to fool the proles into believing it. That ends at the current White House where a 5 year is busy claiming to be smartest, the biggest, etc. He needs constant praise as Dear Leader because he has such a fragile ego.

      • If you don't live in the US I can tell you that people here have an enormous collossal inferiority complex. Men whose penisis don't work buying enormous trucks is just a tiny facet of this.

      • Haha! Mod parent insightful. And I always thought it was just because they were just dumb inbred hillbillies. Could be both, if you can take anything away from the news. Apologies for not posting anon.
    • The whole AI craze is in fact an American ruse to trick Chinese into wasting money. And it seems to be working!
      • Chinese have taken the approach of "do more with less," something many Americans could learn a thing or two about. Take the education system for example.
        • Chinese have taken the approach of "do more with less," something many Americans could learn a thing or two about. Take the education system for example.

          Don' worry. Teachers in the U.S. are very familiar with doing more with less. More students per class, more censorship imposed on what they can teach, all for less pay.
  • The problem with using opensource benchmarks is, even I can make an LLM that scores well (or any score I want) if I have the data for the benchmark available.

    • Re: (Score:1)

      by SirSlud ( 67381 )

      I mean, I'd be impressed if you can even tie your own shoes p5, lets start with that

  • No need to worry. (Score:4, Insightful)

    by Rumagent ( 86695 ) on Tuesday April 08, 2025 @02:40AM (#65288941)
    I am sure those big beautiful tariffs will fix this as well. No need to think, just listen to the great leader.

    • Re: (Score:1)

      by tyroxy ( 1291304 )
      One way to analyze a measure like tariffs is to see who will benefit. If you follow Martyanov, he makes a plausible argument that Europe will be the biggest loser -- these are export dominated economies for whom trade with the US is a vital support to their economy. What they depend on, they will now be taxed for.
      • Europe imports a lot of services from the US. In fact, the trade deficit between the US and Europe is insignificant (less than $50B a year).

        • Re: (Score:2)

          by AmiMoJo ( 196126 )

          For now. The great uncoupling has begun. This mistake will not be made twice.

        • Europe imports a lot of services from the US. In fact, the trade deficit between the US and Europe is insignificant (less than $50B a year).

          For whatever reason, the Mad King only considers goods, not services. When you look at goods only, the trade deficit was $235B in 2024.

      • Re: (Score:3)

        by Bumbul ( 7920730 )

        You have it backwards. Tariffs are paid by the importer and very probably that cost is carried over to the consumer.
        And EU exports many items that are quite crucial to US economy, i.e. there are no easy substitutes - thus, this hurts US economy.

        Most exported goods from EU to US? Top of the list:
        1. Medicinal and pharmaceutical products
        2. Medicaments
        3. Motor cars (BMW, Mercedes Benz, Audi, Ferrari, Porche, Lamborghini, etc...)
        4. Aircraft and associated equipment (e.g. Boeing 777/787 and F35 use Rolls-Ro

        • Re: (Score:2)

          by tlhIngan ( 30335 )

          Goods wise, there is a trade deficit as the EU exports more goods to the US. If you take services, there is a trade surplus as the EU imports more services from the US.

          If you take both together, the difference is tiny.

      • Its best for American jobs. Democrats have been begging to return manufacturing to the US for a long time. Democrats got what they want but Bad Orange Man must always be hated, political alliances take precedent over enacting their ideology.

        Personally, I'm a free trade guy, but its clear here who Trump is going to bat for- The American worker. He wants these products companies are buying oversees to be made here in the USA. What better way than hefty tarriffs? If you have a problem with this, you need to s

        • Re: (Score:2)

          by jp10558 ( 748604 )

          Its best for American jobs.

          Is it though? It seems like it's leading to mass layoffs and a recession. We can't just suddenly magically make everything here, and it seems to me that at least in the next few years it'd make sense to keep your competitiveness for "the rest of the world" and just have the US consumer pay a lot more for the stuff they can't do without. This will leave the US consumer just not doing or buying things they can live without and thereby slowing the economy as they'll be spending money on less stuff and services

      • ... what they depend on YOU will be taxed on ... ftfy

      • One way to analyze a measure like tariffs is to see who will benefit.

        For each of Trump's actions and statements, Asking "does it benefit the US?" is often debatable. Asking "Does it benefit Russia?" is easier. It's always Yes.

  • The last few Chinese models were great and the last US model is huge and doesn't seem to beat the much smaller Chinese one in many disciplines. I would say they should first catch up before talking about taking the lead again.

    • Using AI to learn about biology, I found R1 better than anything so far. The US models just parroted the kind of specious answers you'd find on medline. R1 sounded like it was actually looking through peer reviewed journals. The difference is stark. Like Google's main search vs its Google Scholar (R1 beings Scholar).

    • Re: (Score:2)

      by HiThere ( 15173 )

      Sorry, but that's not valid evidence.

      The US companies are not incentivized to release models that allow just anyone to run the AI. And everybody uses distilling to reduce the size of the model after training. (It *has* been demonstrated by US companies, just not very frequently.)

      The real answer is that I expect NOBODY KNOWS who is more advanced, and in what areas. Perhaps the CCP knows, because they do a lot of spying, and they can demand access to "company private" information. Perhaps some intelligen

      • Re: (Score:2)

        by allo ( 1728082 )

        Then let's speculate. Who do you think has the secret super model hidden from the public?
        Meta's last effort didn't look very good. OpenAI struggles to do something new in the LLM area. Maybe xAI? If you ask the Musk fans, xAI surely has AGI somewhere and only doesn't use it because it's too dangerous.

        • Re: (Score:2)

          by HiThere ( 15173 )

          Well, speculation in the absence of good information is fun, and it's necessary if you're going to plan actions based on it, My guess is that the eventual winner will be somebody very few people know about. It will need to have really good funding, but not spectacular. And it will probably be bought up by someone else (perhaps Microsoft, perhaps Amazon).

          OTOH, I wouldn't plan any actions based on that speculation. It's too uncertain.

          • Re: (Score:2)

            by allo ( 1728082 )

            You're the one speculating on secret models. Let's do it with Occams Razor: As long as there is no indication for secret models, we assume there are none and companies show of their best models.

            • Re: (Score:2)

              by HiThere ( 15173 )

              "Secret models" is a default assumption. To assume otherwise is to assume that no research is being done.

  • The wounded animal lashes out as best it can: tariffs, institutional changes, rash measures...

  • I love how we’re calling any AI as “state of the art”. As if the broken conversations we’ve had with that child “mind” are mind-bogglingly insightful to grown-ass educated adults.

    I suppose we’ll start labeling the works of Hunter Biden as “art” now too. I mean c’mon man. They couldn’t possibly be a scam. Neither is selling “intelligence”, right?

    Label me a Troll? Find a brilliant AI and prove me wrong first.

  • Given how rotten software patents can be, and the notoriously poor quality of Chinese papers, are their numbers really a useful measure?

    • Re: (Score:2)

      by HiThere ( 15173 )

      A valid point, but a strongly biased one. You can't trust what the US companies are saying either. If they see keeping something secret as a commercial advantage, they will. (This is also true for the other players, but I think they are estimated to be further behind, so they have less advantage to lose, and perhaps none.)

      • Yes, but keeping information secret and inventing fake information are rather different, no? If anything, you're saying that the US is farther ahead than public data would show.

  • It's well known that adversity breeds strength, excellence and ingenuity. Militaries the world over rely on it, and it's one of the bases of natural selection. So why is it a surprise that embargoes and trade sanctions have given the Chinese motivation to excel, both in spite of and because of those measures?

    Cutting an enemy / competitor off from the most advanced tech can confer some short-term advantages, but in the long term you may actually be strengthening the opposition.

  • Until someone comes up with a use case that can generate billions in revenue to offset the insane cost of running these models at scale, it doesn't matter who "wins" in a race to the bottom.
    • The importance of your point can't be overstated. As such, perhaps what's holding the US back is the need to hype the model to make money because the investors are starting to suspect that they have wasted their money instead of developing the model until it's good and actually useful for something.
  • The entire comment section for this topic is unreadable because initial or secondary posts are off topic, leading to whole threads being a landmine of trolls, idiots, propaganda and brainless cultists.Please. PLEASE stop it. I want to post about news for nerds. I want to read news for nerds. I'm interested in AI, in the dick measuring contest beyond one narrow facet of American politics, in how China improved, in how much TFA proves or doesn't. I'm not interested in reading the fifteenth post about politici
  • Any country can make a breakthrough in any area which then gives them an advantage. It is often not even a linear path to develop a field, it only depends on who comes up with new ideas. Some things like making integrated circuits are hard to master without proprietary knowledge, but an entirely new method of making integrated circuits could be invented and bypass the current chip fabs leads. The idea that science is a race is stupid, it is much more like exploring an unknown forest and you can not keep

Slashdot Top Deals

Real Users never use the Help key.

Close