5 ms·

It’s pretty interesting though. Despite all the renewable buildout they still rely very much on coal and other fossil fuels. As it pertains to the conversation

by ericmay 5mo ago

It’s pretty interesting though. Despite all the renewable buildout they still rely very much on coal and other fossil fuels. As it pertains to the conversation, if you are allowing China to make advances and change you also have to allow that for America - in other words we can compete and build out our grid too and compete across the spectrum, low-cost, energy efficient models, or otherwise. On the high end China does not currently have a competitor so America stands to compete on both fronts with a current advantage on the high end which is what matters most. If China commoditizes their models they, well, become commodities which means lots of competition and market entry by competing firms. American firms for example can also create low-cost, low-energy models it’s just that there is no economic model to do so. Racing to the bottom is a fool’s errand.

bigbadfeline 5mo ago
> it’s just that there is no economic model to do so "Just"? Where do I start? That just outweighs all other problems combined, by far.
- ericmay 5mo ago
  I'm not sure it does. China can churn out all the behind the times models that it wants to and make them open source but what people are willing to pay for are genuinely useful and incredible models at the high end. You can go buy an open source Android phone for $100 from Wal-Mart, but most people choose the iPhone because even though it's more expensive it's better and more useful. Sort of a greater than the sum total of the parts for example. So sure you can churn out free/cheap stuff, but that's not a business model, that's desperation. This whole thread and conversation is exactly proving the point. US - gargantuan amount of spend and expansion of compute power and energy build out with the aim to bring about nothing less than the Machine God. Companies are paying fortunes to run the latest and greatest models and seeing productivity and efficiency gains and increases in revenue. China - lots of idle clean energy production potential due to planned economy edicts (where have we seen this before, cough housing cough), reliant on American tech stacks and chips, releases free and open source models for folks to tinker with that are behind and continuously behind American models. Oh and nobody is paying for the Chinese stuff. Here's the real problem: in America kids are booing AI models and demonstrating against tech companies while in China everyone is all-in on AI and fearful of being behind. We're completely and demonstrably capable of dominating this space. Newsflash - America can and does make great stuff. We are incredibly competitive, dynamic, innovative and operate with very deep capital markets. We have world-leading companies and will continue to do so even as China and companies from other countries continue to compete hard too. The risk factor is anti-American propaganda and leftist subversion convincing people that adoption of technology is a bad thing and then we start banning technology or implementing other ass-backwards policies and we wind up behind.
  - verdverm 5mo ago
    > what people are willing to pay for are genuinely useful and incredible models at the high end I do not believe this to be true in the long run. I think American Ai companies understand this. I've already made the switch because the difference is increasingly small. Progress up to now shows us that the small models are about 1 year behind frontier, and at some point will become good enough for the majority of tasks, i.e. you don't always need the biggest or the best. This conversation is already happening, where people are choosing the smaller models from American companies, or moving to open weight models they can run themselves. Heck, Congress even has American companies coming in to testify about their usage of Chinese models because they do not understand the difference between downloading and running it yourself vs making API calls to servers in another country.
    - ericmay 5mo ago
      We'll find out but certainly in the software industry many companies pay for those services I mentioned even when there are cheaper alternatives. It's one thing to download some "small" model and hack around, but that does not translate to efficiency at scale. The "small" models and such haven't proven to be profitable and so you can wind up with some, certainly powerful, small models but they'll be far enough behind that the bleeding edge models will continue to compound games. It's no different than any other positive feedback loop - you see it in chip design and software in general. If all China is doing is releasing models that are 6 months or more behind for free, forget it. Game over. That's 6 months of run time for a better model to exploit vulnerabilities or whatever. I think there's plenty of room for lots of approaches. But similar to how Open Office is nice and all, Microsoft Office makes however many hundreds of millions. I'm not really sure why folks believe that what we're seeing today will be any different. Bleeding edge compounds gains and advantage. American tech companies are run by some of the smartest people on the planet and from all over the planet. They're not spending hundreds of billions of dollars because they're wrong and because some 1-year behind model is a big threat.
      - verdverm 5mo ago
        The big models have not shown to be profitable, likely the small models will first since it's on the order of 100x cheaper There is a likely world (imo) where Big AI produces mega models primarily to distill them down into economically viable sized models China also has mega models, US also has open models. Both groups are doing the same things. AllenAi is the most open but their models have not been competitive

by ericmay 5mo ago

bigbadfeline 5mo ago

> it’s just that there is no economic model to do so "Just"? Where do I start? That just outweighs all other problems combined, by far.

ericmay 5mo ago
I'm not sure it does. China can churn out all the behind the times models that it wants to and make them open source but what people are willing to pay for are genuinely useful and incredible models at the high end. You can go buy an open source Android phone for $100 from Wal-Mart, but most people choose the iPhone because even though it's more expensive it's better and more useful. Sort of a greater than the sum total of the parts for example. So sure you can churn out free/cheap stuff, but that's not a business model, that's desperation. This whole thread and conversation is exactly proving the point. US - gargantuan amount of spend and expansion of compute power and energy build out with the aim to bring about nothing less than the Machine God. Companies are paying fortunes to run the latest and greatest models and seeing productivity and efficiency gains and increases in revenue. China - lots of idle clean energy production potential due to planned economy edicts (where have we seen this before, cough housing cough), reliant on American tech stacks and chips, releases free and open source models for folks to tinker with that are behind and continuously behind American models. Oh and nobody is paying for the Chinese stuff. Here's the real problem: in America kids are booing AI models and demonstrating against tech companies while in China everyone is all-in on AI and fearful of being behind. We're completely and demonstrably capable of dominating this space. Newsflash - America can and does make great stuff. We are incredibly competitive, dynamic, innovative and operate with very deep capital markets. We have world-leading companies and will continue to do so even as China and companies from other countries continue to compete hard too. The risk factor is anti-American propaganda and leftist subversion convincing people that adoption of technology is a bad thing and then we start banning technology or implementing other ass-backwards policies and we wind up behind.
- verdverm 5mo ago
  > what people are willing to pay for are genuinely useful and incredible models at the high end I do not believe this to be true in the long run. I think American Ai companies understand this. I've already made the switch because the difference is increasingly small. Progress up to now shows us that the small models are about 1 year behind frontier, and at some point will become good enough for the majority of tasks, i.e. you don't always need the biggest or the best. This conversation is already happening, where people are choosing the smaller models from American companies, or moving to open weight models they can run themselves. Heck, Congress even has American companies coming in to testify about their usage of Chinese models because they do not understand the difference between downloading and running it yourself vs making API calls to servers in another country.
  - ericmay 5mo ago
    We'll find out but certainly in the software industry many companies pay for those services I mentioned even when there are cheaper alternatives. It's one thing to download some "small" model and hack around, but that does not translate to efficiency at scale. The "small" models and such haven't proven to be profitable and so you can wind up with some, certainly powerful, small models but they'll be far enough behind that the bleeding edge models will continue to compound games. It's no different than any other positive feedback loop - you see it in chip design and software in general. If all China is doing is releasing models that are 6 months or more behind for free, forget it. Game over. That's 6 months of run time for a better model to exploit vulnerabilities or whatever. I think there's plenty of room for lots of approaches. But similar to how Open Office is nice and all, Microsoft Office makes however many hundreds of millions. I'm not really sure why folks believe that what we're seeing today will be any different. Bleeding edge compounds gains and advantage. American tech companies are run by some of the smartest people on the planet and from all over the planet. They're not spending hundreds of billions of dollars because they're wrong and because some 1-year behind model is a big threat.
    - verdverm 5mo ago
      The big models have not shown to be profitable, likely the small models will first since it's on the order of 100x cheaper There is a likely world (imo) where Big AI produces mega models primarily to distill them down into economically viable sized models China also has mega models, US also has open models. Both groups are doing the same things. AllenAi is the most open but their models have not been competitive