Banning self-recursive improvement in AI models?


Some people are suggesting this, including Ezra Klein.

I do not understand how it is supposed to work.  Put aside the issue of foregone innovations, let us say I seek to access an AI from overseas.  Am I allowed to visit their website?  So many sites have AI behind them, including Chinese open source AI.  Do all of those sites get banned?  And how?  Firewall imposed on Americans?  And we cannot put the good open source models on our hard drives?  How enforced?  Does the government have to read the creation logs of the model (how can they?), and then decide whether I can use that model or not?

What if Anthropic licenses its IP to an independent subsidiary in the Cayman Islands and they keep on using code to improve the AI?  Or maybe only foreign nations are allowed to have the best models and the really fast rates of improvement?  How are multinationals supposed to operate across borders?  (Presumably you cannot just use the RSI of your foreign affiliate, but then the whole MNC is crippled and eventually rendered uncompetitive?)  How are we to keep “AI sovereignty” for Americans?

Does this whole thing mean that regular companies can use AI to write code, but the main labs cannot?  Solve for the equilibrium there.  Or the whole code writing function goes away for everybody?

I do not find this to be a workable idea.  I think it is “people wanting to do something that sounds reasonable and less scary,” without thinking through the practical details.  I have made a related point before, but one does best with AI policy when you start with the stuff that otherwise you might put into your last two or three paragraphs of an essay (“It remains to be seen how…but we are all human beings and surely we should not give up hope…”).  Even if you believe in a deal with China, and I do not, we (and others) now know competitors can catch up more quickly than we had thought, at least if they think they have a chance of winning.  They might even buy some space compute from Elon.

C’mon people, I know for many of you this is a difficult take to swallow, because it feels like “we don’t care enough” if we do not take some dramatic steps.  Nonetheless the reality is that, a long time ago, we made a whole series of decisions that imply…we have no other choice than to just see this one through and get it more or less right.  The longer we deny that truth, the harder it will be for all of us.



Source link

  • Related Posts

    CGTN: How head-of-state diplomacy steers China-US ties toward stability

    Elaborating on the new positioning, Zhang Tengjun, associate research fellow at the Institute of American Studies at Chinese Academy of Social Sciences, said it is a new calibration of the…

    Continue reading
    Unaudited interim results for the three-and six-month periods ended 30 June 2026

    Forward-looking statementsCertain statements in this announcement are, or may be deemed to be, forward looking statements. Forward looking statements are identified by their use of terms and phrases such as…

    Continue reading

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You Missed

    Trump and Xi see tensions flare over AI, trade and Iran but still seek stability

    Trump and Xi see tensions flare over AI, trade and Iran but still seek stability

    CGTN: How head-of-state diplomacy steers China-US ties toward stability

    T-Mobile Promo Codes: 25% Off | September 2026

    T-Mobile Promo Codes: 25% Off | September 2026

    Asian Games final – India bat vs Sri Lanka in gold-medal match – both sides unchanged

    Asian Games final – India bat vs Sri Lanka in gold-medal match – both sides unchanged

    Anya Taylor-Joy’s Near-Perfect Breakout Folk Thriller Is Officially Leaving Netflix

    Anya Taylor-Joy’s Near-Perfect Breakout Folk Thriller Is Officially Leaving Netflix

    7 Best Sustainable Trundle Beds For Siblings

    7 Best Sustainable Trundle Beds For Siblings