4 ms·

There’s a Twitter thread making rounds by Dean Ball about deceleration in AI development caused by open models and I can’t understand how people don’t see that

by baq 3mo ago

There’s a Twitter thread making rounds by Dean Ball about deceleration in AI development caused by open models and I can’t understand how people don’t see that it’s true: open models dismantle the frontier lab capex spend potential by reducing the training budget to zero in the limit. Tokens from different providers are not fungible, but customers are nevertheless very price sensitive and close enough is good enough, eg. K3 being opus+ in capability and cheaper than opus per successful task in the long run is an obvious financial decision. No training budget means deceleration, or at least slower acceleration, margin compression and a completely demolished IPO valuation; path to machine god requires dollars and capable open models externalize training costs to true frontier labs parasitically. IMHO humanity has a better chance at not destroying itself due to less than breakneck pace - but there’s a chance frontier models get sponsored by the USG and are never released publicly so they can’t be distilled and then what?

green7ea 3mo ago
I’m not entirely convinced, there are many dimensions to progress. For example, DeepSeek has had a few very impressive innovations that all models could benefit from. There’s also the law of diminishing returns, the US labs have plenty of CAPEX already. Sometimes, constraints, like sanctions, can also be a source if innovation.
anon373839 3mo ago
He recently did a walkback of that post. But ultimately, who cares? If the only way for AI to progress is in the hands of a few closed players, well, I don’t really think humanity needs that. Of course, it’s a preposterous claim in the first place. The ultimate reason deep learning and LLMs have made it as far as they have is the explosion of open research and research artifacts in the last decade.
cherryteastain 3mo ago
> there’s a chance frontier models get sponsored by the USG and are never released publicly so they can’t be distilled and then what? That premise hinges on one implicit assumption: Chinese advances are due to distillation ONLY and that Chinese model providers cannot keep advancing if they do not distill, which is a very big if. If Chinese models keep advancing in such a scenario, and they almost certainly will, they will overtake publically available models by US providers and China will dominate the LLM industry.
weiliddat 3mo ago
I read his followup tweet, and your comment, and I'm not fully convinced that open models are decelerationist. Happy to hear other thoughts on this. Open weight AI is decelerationist from the perspective that all capital should be allocated to a market leaders for training, and that the market leader is fully invested in continuously making the models smarter, cheaper, faster for its users, or that distillation from this market leader is the main way to make progress. We might reach a local optimum/equilibrium faster without open weight models, with leaders capturing more of the market faster to a point where further R&D isn't required due to lack of competition. I also doubt that distillation is the only/main way that open weight models were advancing AI research. We can name a few examples from DeepSeek around reasoning, context optimization, etc. I'm also unconvinced that the overall market capex on AI is lower given more competition (probably less specifically for US market capex, which is decelerationist from only the US perspective).
fidotron 3mo ago
The big decelerationist threat is a sudden reduction in competition. If either OpenAI or Anthropic drop out or the open weights stuff is banned/becomes uncompetitive then the motivation and tolerance for taking risks with the larger training runs tanks. The closest we've seen to this in tech in recent decades was iOS vs Android, where Android only really was competitive for a very short window of time (approx 4.x) and it was during that period that both Android and iOS actually improved dramatically for end users. Once Android lost the plot again, and especially in the US market, all that energy started going in some very silly directions.
- jonners00 3mo ago
  I have to use both big mobile OSs for work and have since 2009. As a result I have been able to be a bit of a gadfly and switch between phone OSs a few times for personal use. I have switched three times to iOS for a year or so, cause I liked the iteration of the iPhone at the time. 4, 6s, X. I have always gone back to Android because it seemed so much better and now I don't plan to switch again. As an end user, Samsung's flavour of Android always seemed better than iOS. I don't know how they compare from an engineer's perspective just from a user perspective. One of my issues with Apple though was hating all their attempts to lock me in, and the lowest common denominator UX (I'm not a power user, but some flexibility is always good). If you're happy with the defaults/a willing hostage, that might make a big difference I guess. Still feel like it's always had feature/spec parity with iOS and iOS devices, and sometimes been ahead. What makes you say Android has only briefly been competitive?
  - buu700 3mo ago
    [dead]
- ahtihn 3mo ago
  > Android only really was competitive for a very short window of time Complete non-sense. iOS and Android are equivalent. Users do not chose Android or iOS because one or the other is better. It's just brand loyalty, status signalling and ecosystem lock-in that creates enough friction that people don't bother.
zozbot234 3mo ago
> There’s a Twitter thread making rounds by Dean Ball about deceleration in AI development caused by open models and I can’t understand how people don’t see that it’s true: open models dismantle the frontier lab capex spend potential by reducing the training budget to zero in the limit. If you're worried about an AGI arms race between the U.S. and China putting AI Safety at risk, then the fact that inherently less knowledgeable/capable models (fewer and more coarsely quantized total parameters than their proprietary competitors according to commonplace rumors) are having a "decelerationist" effect is actually great news. Even better if China is actually "Yann LeCun-pilled" (verbatim from Ball's post) and doesn't really believe in early AGI. So explain to us exactly why we're supposed to ban/discourage use of these open source models? The only way that makes sense is as a transparently self-serving proposal from the chief OpenAI policy lobbyist.
- NiloCK 3mo ago
  The logic, whose premises you can take or leave: Even at the level of, say, Opus 4.5+, open weight models give a quick turnaround to every Joe and Jane on earth having easy access to pretty high quality improvised weapons design, cyber / auto-fraud capabilities, etc. All the existing models (closed and open) put up decent resistance to participating in activities like this, and especially behind API walls with content monitoring and account bans. But the published open-weight models can be fine tuned or abliterated into arbitrarily sharp-edged tools. EG, if it's physically feasible to build a nuke in your garage, it may soon be the case that more or less anyone will have competent guidance to do so.
  - CamperBob2 3mo ago
    You don't need AI to build a nuke in your garage. You need uranium. And if you have uranium, you still don't need AI. You need a pocket calculator, a library card, and a death wish.
    - NiloCK 3mo ago
      Yes - persons with death wishes having arbitrarily powerful consultation is the crux of it. Apologies for the bad example. Replace w/ gain of function / whatever else, or just brainstorm with your local model, ect.
photios 3mo ago
Love the deceleration narrative :) "No, sir, we haven't reached the peak of this tech... It's those open models! Please, keep pumping dollars into the market!"
a34729t 3mo ago
I dunno, it means Anthropic and OpenAi need to get efficient and maybe cannot just expect trillion dollar ipos?
nullc 3mo ago
The large US closed AI companies are decelerationist because their focus is on monopolizing the market. They spend inefficiently in order to lock up the supply of resources and waste money influencing the state to attempt to lock out competition. This strategy has not been successful due to the existence of isolated resource pools they can't monopolize.

by baq 3mo ago

green7ea 3mo ago

I’m not entirely convinced, there are many dimensions to progress. For example, DeepSeek has had a few very impressive innovations that all models could benefit from. There’s also the law of diminishing returns, the US labs have plenty of CAPEX already. Sometimes, constraints, like sanctions, can also be a source if innovation.

anon373839 3mo ago

He recently did a walkback of that post. But ultimately, who cares? If the only way for AI to progress is in the hands of a few closed players, well, I don’t really think humanity needs that. Of course, it’s a preposterous claim in the first place. The ultimate reason deep learning and LLMs have made it as far as they have is the explosion of open research and research artifacts in the last decade.

cherryteastain 3mo ago

> there’s a chance frontier models get sponsored by the USG and are never released publicly so they can’t be distilled and then what? That premise hinges on one implicit assumption: Chinese advances are due to distillation ONLY and that Chinese model providers cannot keep advancing if they do not distill, which is a very big if. If Chinese models keep advancing in such a scenario, and they almost certainly will, they will overtake publically available models by US providers and China will dominate the LLM industry.

weiliddat 3mo ago

I read his followup tweet, and your comment, and I'm not fully convinced that open models are decelerationist. Happy to hear other thoughts on this. Open weight AI is decelerationist from the perspective that all capital should be allocated to a market leaders for training, and that the market leader is fully invested in continuously making the models smarter, cheaper, faster for its users, or that distillation from this market leader is the main way to make progress. We might reach a local optimum/equilibrium faster without open weight models, with leaders capturing more of the market faster to a point where further R&D isn't required due to lack of competition. I also doubt that distillation is the only/main way that open weight models were advancing AI research. We can name a few examples from DeepSeek around reasoning, context optimization, etc. I'm also unconvinced that the overall market capex on AI is lower given more competition (probably less specifically for US market capex, which is decelerationist from only the US perspective).

fidotron 3mo ago

The big decelerationist threat is a sudden reduction in competition. If either OpenAI or Anthropic drop out or the open weights stuff is banned/becomes uncompetitive then the motivation and tolerance for taking risks with the larger training runs tanks. The closest we've seen to this in tech in recent decades was iOS vs Android, where Android only really was competitive for a very short window of time (approx 4.x) and it was during that period that both Android and iOS actually improved dramatically for end users. Once Android lost the plot again, and especially in the US market, all that energy started going in some very silly directions.

jonners00 3mo ago
I have to use both big mobile OSs for work and have since 2009. As a result I have been able to be a bit of a gadfly and switch between phone OSs a few times for personal use. I have switched three times to iOS for a year or so, cause I liked the iteration of the iPhone at the time. 4, 6s, X. I have always gone back to Android because it seemed so much better and now I don't plan to switch again. As an end user, Samsung's flavour of Android always seemed better than iOS. I don't know how they compare from an engineer's perspective just from a user perspective. One of my issues with Apple though was hating all their attempts to lock me in, and the lowest common denominator UX (I'm not a power user, but some flexibility is always good). If you're happy with the defaults/a willing hostage, that might make a big difference I guess. Still feel like it's always had feature/spec parity with iOS and iOS devices, and sometimes been ahead. What makes you say Android has only briefly been competitive?
- buu700 3mo ago
  [dead]
ahtihn 3mo ago
> Android only really was competitive for a very short window of time Complete non-sense. iOS and Android are equivalent. Users do not chose Android or iOS because one or the other is better. It's just brand loyalty, status signalling and ecosystem lock-in that creates enough friction that people don't bother.

zozbot234 3mo ago

> There’s a Twitter thread making rounds by Dean Ball about deceleration in AI development caused by open models and I can’t understand how people don’t see that it’s true: open models dismantle the frontier lab capex spend potential by reducing the training budget to zero in the limit. If you're worried about an AGI arms race between the U.S. and China putting AI Safety at risk, then the fact that inherently less knowledgeable/capable models (fewer and more coarsely quantized total parameters than their proprietary competitors according to commonplace rumors) are having a "decelerationist" effect is actually great news. Even better if China is actually "Yann LeCun-pilled" (verbatim from Ball's post) and doesn't really believe in early AGI. So explain to us exactly why we're supposed to ban/discourage use of these open source models? The only way that makes sense is as a transparently self-serving proposal from the chief OpenAI policy lobbyist.

NiloCK 3mo ago
The logic, whose premises you can take or leave: Even at the level of, say, Opus 4.5+, open weight models give a quick turnaround to every Joe and Jane on earth having easy access to pretty high quality improvised weapons design, cyber / auto-fraud capabilities, etc. All the existing models (closed and open) put up decent resistance to participating in activities like this, and especially behind API walls with content monitoring and account bans. But the published open-weight models can be fine tuned or abliterated into arbitrarily sharp-edged tools. EG, if it's physically feasible to build a nuke in your garage, it may soon be the case that more or less anyone will have competent guidance to do so.
- CamperBob2 3mo ago
  You don't need AI to build a nuke in your garage. You need uranium. And if you have uranium, you still don't need AI. You need a pocket calculator, a library card, and a death wish.
  - NiloCK 3mo ago
    Yes - persons with death wishes having arbitrarily powerful consultation is the crux of it. Apologies for the bad example. Replace w/ gain of function / whatever else, or just brainstorm with your local model, ect.

photios 3mo ago

Love the deceleration narrative :) "No, sir, we haven't reached the peak of this tech... It's those open models! Please, keep pumping dollars into the market!"

a34729t 3mo ago

I dunno, it means Anthropic and OpenAi need to get efficient and maybe cannot just expect trillion dollar ipos?

nullc 3mo ago

The large US closed AI companies are decelerationist because their focus is on monopolizing the market. They spend inefficiently in order to lock up the supply of resources and waste money influencing the state to attempt to lock out competition. This strategy has not been successful due to the existence of isolated resource pools they can't monopolize.