The point about agents to conceal access to the model is a good one.
Hopefully we won’t lose all access to models in future
S1: A $6 R1 competitor?
41–50 of 430 posts
Re: S1: A $6 R1 competitor?
#42> having 10,000 H100s just means that you can do 625 times more experiments than s1 did I think the ball is very much in their court to demonstrate they actually are using their massive compute in such a productive fashion. My BigTech experience would tend to suggest that frugality went out the window the day the valuation took off, and they are in fact just burning compute for little gain, because why not...
Mainly it points to a non-scientific "bigger is better" mentality, and the researchers probably didn't mind playing around with the power because "scale" is "cool". Remember that the Lisp AI-labs people were working on non-solved problems on absolute potatoes of computers back in the day, we have a semblance of progress solution but so much of it has been brute-force (even if there has been improvements in the field)…
We are in a phase where costs are really coming down. We had this phase from GPT2 to about GPT4 where the key to building better models was just building bigger models and training them for longer. But since then a lot of work has gone into distillation and other techniques to make smaller models more capable.
If there is another AI winter, it will be more like the dotcom bubble: lots of important work got done in the dotcom bubble, but many of the big tech companies started from the fruits of that labor in the decade after the bubble burst
Re: S1: A $6 R1 competitor?
#43Qwen's QvQ-72B does much more "wait"s than other LLMs with CoT I tried, maybe they've somewhat used that trick already?
Re: S1: A $6 R1 competitor?
#44The part about taking control of a reasoning model's output length using tags is interesting. > In s1, when the LLM tries to stop thinking with " ", they force it to keep going by replacing it with "Wait". I had found a few days ago that this let you 'inject' your own CoT and jailbreak it easier. Maybe these are related? https://pastebin.com/G8Zzn0Lw https://news.ycombinator.com/item?id=42891042#42896498
This even points to a reason why OpenAI hides the "thinking" step: it would be too obvious that the context is being manipulated to induce more thinking.
Re: S1: A $6 R1 competitor?
#45Deepseek R1 uses and wait and you can see it in the thinking tokens second guessing itself. How does the model know when to wait?
These reasoning models are feeding more to OP's last point about NVidia and OpenAI data centers not being wasted since reason models require more tokens and faster tps.
Re: S1: A $6 R1 competitor?
#46> If you believe that AI development is a prime national security advantage, then you absolutely should want even more money poured into AI development, to make it go even faster. This, this is the problem for me with people deep in AI. They think it’s the end all be all for everything. They have the vision of the ‘AI’ they’ve seen in movies in mind, see the current ‘AI’ being used and to them it’s basically almost t…
I can only say that exponential curves grow nominally sublinearly before they take off. AI is not quite at the obvious take off point, but owners of the biggest clusters have seen the extrapolations and it isn't pretty - once your competitor achieves take off and you aren't anywhere close, you're done for. The risk of not participating in that are too great.
Re: S1: A $6 R1 competitor?
#47> If you believe that AI development is a prime national security advantage, then you absolutely should want even more money poured into AI development, to make it go even faster. This, this is the problem for me with people deep in AI. They think it’s the end all be all for everything. They have the vision of the ‘AI’ they’ve seen in movies in mind, see the current ‘AI’ being used and to them it’s basically almost t…
You can choose to be somewhat ignorant of the current state in AI, about which I could also agree that at certain moments it appears totally overhyped, but the reality is that there hasn't been a bigger technology breakthrough probably in the last ~30 years.
This is not "just" machine learning because we have never been able to do things which we are today and this is not only the result of better hardware. Better hardware is actually a byproduct. Why build a PFLOPS GPU when there is nothing that can utilize it?
If you spare yourself some time and read through the actual (scientific) papers of multiple generations of LLM models, the first one being from Google ~~not DeepMind~~ in 2017, you might get to understand that this is no fluff.
And I'm speaking this from a position of a software engineer, without bias.
The reason why all this really took off with so much hi-speed is because of the not quite expected results - early LLM experiments have shown that "knowledge" with current transformers architecture can linearly scale with regards to the amount of compute and training time etc. That was very unexpected and to this day scientists do not have an answer why this even works.
So, after reading bunch of material I am inclined to think that this is something different. The future of loading the codebase into the model and asking the model to explain me the code or fix bugs has never been so close and realistic. For the better or worse.
Re: S1: A $6 R1 competitor?
#48> If you believe that AI development is a prime national security advantage, then you absolutely should want even more money poured into AI development, to make it go even faster. This, this is the problem for me with people deep in AI. They think it’s the end all be all for everything. They have the vision of the ‘AI’ they’ve seen in movies in mind, see the current ‘AI’ being used and to them it’s basically almost t…
> Real AI, artificial intelligence, is a fever dream. This is machine learning except the machines are bigger than ever before. There is no intellect.
Even if the machine learning isn't really intelligent, it is still capable of performing IF..THEN..ELSE operations, which could have detrimental effects for [some subset of] humans.
And even if you argue that such a machine _shouldn't_ be used for whatever doomsday scenario would harm us, rest assured that someone, somewhere, who either does not understand what the machines are designed to do or just pretends that they work like magic, will put the machines in a position to make such a decision.
Re: S1: A $6 R1 competitor?
#49> having 10,000 H100s just means that you can do 625 times more experiments than s1 did I think the ball is very much in their court to demonstrate they actually are using their massive compute in such a productive fashion. My BigTech experience would tend to suggest that frugality went out the window the day the valuation took off, and they are in fact just burning compute for little gain, because why not...
Besides that, AI training (aka gradient descent) is not really an "embarrassingly parallel" problem. At some point, there are diminishing returns on adding more GPUs, even though a lot of effort is going into making it as parallel as possible.
Re: S1: A $6 R1 competitor?
#50Deepseek R1 uses and wait and you can see it in the thinking tokens second guessing itself. How does the model know when to wait? These reasoning models are feeding more to OP's last point about NVidia and OpenAI data centers not being wasted since reason models require more tokens and faster tps.
Probably when it would expect a human to second guess himself, as shown in literature and maybe other sources.