Yes, for some things.
Will scaling work?
111–120 of 289 posts
Re: Will scaling work?
#112I think there’s a huge assumption here that more LLM will lead to AGI. Nothing I’ve seen or learned about LLMs leads me to believe that LLMs are in fact a pathway to AGI. LLMs trained on more data with more efficient algorithms will make for more interesting tools built with LLMs, but I don’t see this technology as a foundation for AGI. LLMs don’t “reason” in any sense of the word that I understand and I think the ab…
Why next-token prediction is enough for AGI - Ilya Sutskever - https://www.youtube.com/watch?v=YEUclZdj_Sc
We need planning. Imagine doing planning like this "drone in a forest" in a different domain like "migrate this project from python to rust".
Re: Will scaling work?
#113I think the more interesting question is how long will people cling to the illusion that LLMs will lead us to AGI? Maintaining the illusion is important to keep the money flowing in.
You say this is as if it's settled and obvious that it it's an illusion and it's only the delusional that believe the opposite.
But if it were so settled and obvious there would be a clear line of reasoning to make that plain. And there is not. Instead, there is a very vibrant debate on the topic with tons of nuance and good faith (and bad) on each side, if we want to talk about sides.
And, of course, one of the implications of this very real and significant inquiry that needs to be made and that requires real contributions from informed individuals, is that whenever anyone is dismissive or reductive regarding the unresolved difficulties, you can be sure they have absolutely no clue what they are talking about.
Re: Will scaling work?
#114Earlier quoted context omitted.
If humans are basically evolved LLMs, which i think is likely; Reasoning will be an emergent property of LLMs within context with appropriate weights.
Why do you think humans are basically evolved LLMs? Honest question, would love to read more about this viewpoint.
My first answer was a bit hasty, let me try again;
We are clearly a product of our past experience (in LLMs this is called our datasets). If you go back to the beginning of our experiences, there is little identity, consciousness, or ability to reason. These things are learned indirectly, (in LLMs this is called an emergent property). We don't learn indiscriminately, evolved instinct, social pressure and culture guide and bias our data consumption (in LLMs this is called our weights).
I can't think of any other way our minds could work, on some level they must function like a LLM, Language perhaps supplemented with general Data, but the principle being the same. Every new idea has been an abstraction or supposition of someones current dataset, which is why technological and general societal advancement has not been linear but closer to exponential.
Re: Will scaling work?
#115Almost everything interesting about AI so far has been unexpected emergent behavior, and huge gains through minor insights. While I don't doubt that the current architecture is likely to have a current ceiling below that of peak human intelligence in certain dimensions, it's already surpassed it in some, and there are still gains to be made in others through things like synthetic data. I also don't understand the cla…
> I also don't understand the claims that it doesn't generalize. I currently use it to solve problems that I can absolutely guarantee were not in its training set, and it generalizes well enough. I also think that one of the easiest ways to get it to generalize better would simply be through giving it synthetic data which demonstrates the process of generalizing.
I don't think what LLMs are currently doing is really generalizing, but rather:
1) Multiple occurrences of something in the dataset are mutually statistically reinforcing. This isn't generalization (abstraction) but rather reinforcement through repetition.
2) Multiple different statistical patterns are being recalled/combined in novel ways such that it seems able to "correctly" respond to things out of dataset, but really this only due to these novel combinations, not due to it having abstracted it's knowledge and applying a more general (or analogical) rule than present in it's individual training points.
Re: Will scaling work?
#116>Furthermore, the fact that LLMs seem to need such a stupendous amount of data to get such mediocre reasoning indicates that they simply are not generalizing. If these models can’t get anywhere close to human level performance with the data a human would see in 20,000 years, we should entertain the possibility that 2,000,000,000 years worth of data will be also be insufficient. There’s no amount of jet fuel you can a…
Demis Hassabis of Deepmind echoes a similar sentiment[0]: > I still think there are missing things with the current systems. […] I regard it a bit like the Industrial Revolution where there was all these amazing new ideas about energy and power and so on, but it was fueled by the fact that there were dead dinosaurs, and coal and oil just lying in the ground. Imagine how much harder the Industrial Revolution would hav…
His view of the Industrial Revolution is completely wrong.
Societies pre-IR had multiple periods where energy usage increased significantly, some of them based specifically around coal. No IR.
Early IR was largely based around the usage of water power, not coal. IR was pure innovation, people being able to imagine and create the impossible, it was going straight to nuclear already.
Ironically, someone who is an innovator believes the very anti-innovation narrative of the IR (very roughly, this is the anti-Eurocentric stuff that began appearing in the 2000s...the world has moved on since then as these theories are obviously wrong). Nothing tells you more about how busted modern universities are than this fact.
Re: Will scaling work?
#117> ‘5 OOMs off’ I think Google, Microsoft and facebook could easily have 5 OOM data than the entire public web combined if we just count text. Majority of people don't have any content on public web except for personal photos. A minority has few public social media posts and it is rare for people to write blog or research paper etc. And almost everyone has some content written in mail or docs or messaging.
Maybe, and certainly with the current trend of synthetic data they can also create it, but I don't think quantity of data beyond what something like GPT-4 has been trained on will in of itself change much other than reducing brittleness by providing coverage of remaining knowledge gaps.
Quality of data (which I believe is at least part of why synthetic data is being used) can perhaps make more of a difference and perhaps at least partly compensate in a crude way for these models lack of outlier rejection and any generalization prediction-feedback loop. Just feed them consistent correct data in the first place.
Re: Will scaling work?
#118I think there’s a huge assumption here that more LLM will lead to AGI. Nothing I’ve seen or learned about LLMs leads me to believe that LLMs are in fact a pathway to AGI. LLMs trained on more data with more efficient algorithms will make for more interesting tools built with LLMs, but I don’t see this technology as a foundation for AGI. LLMs don’t “reason” in any sense of the word that I understand and I think the ab…
We do have systems that reason. Prolog comes to mind. It's a niche tool, used in isolated cases by relatively few people. I think that the other candidates are similar: proof assistants, physics simulators, computational chemistry and biology workflows, CAD, etc. When we get to the point where LLMs are able to invoke these tools for a user, even if that user has no knowledge of them, and are able to translate the res…
> We do have systems that reason. Prolog comes to mind. It's a niche tool, used in isolated cases by relatively few people. I think that the other candidates are similar: proof assistants, physics simulators, computational chemistry and biology workflows, CAD, etc.
I think OP meant other definition of reason, because by your definition calculator can also reason. These are tools created by humans, that help them to reason about stuff by offloading calculations for some of the tasks. They do not reason on their own and they can't extrapolate. They are expert systems.
Re: Will scaling work?
#119I think there's a need to separate knowledge from learning algorithm. There's need to be a latent representation of knowledge that models attend to but the way it's done right now (with my limited understanding) doesn't seem to be it. Transformers seems to only attend to previous text in the context but not to the whole knowledge they posses which is obvious limitation IMO. Human brain probably also doesn't attend to whole knowledge but loads something into context so maybe it's fixable without changing architecture.
LLMs can work as data extraction already, so one can build some prolog DB and update it as it consumes data. Then translate any logic problems into prolog queries. I want to see this in practice.
Similar with usage of logic engines and computation/programs.
I also think that RL can come up with better training function for LLMs. In the programming domain for example one could ask LLM to think about all possible test for given code and evaluate them automatically.
I was also thinking about using diffusER pattern where programming rules are kinda hardcoded (similar to add/replace/delete but instead algebra on functions/variables). Thats probably not AGI path but could be good for producing programs.
Re: Will scaling work?
#120I think there’s a huge assumption here that more LLM will lead to AGI. Nothing I’ve seen or learned about LLMs leads me to believe that LLMs are in fact a pathway to AGI. LLMs trained on more data with more efficient algorithms will make for more interesting tools built with LLMs, but I don’t see this technology as a foundation for AGI. LLMs don’t “reason” in any sense of the word that I understand and I think the ab…
If humans are basically evolved LLMs, which i think is likely; Reasoning will be an emergent property of LLMs within context with appropriate weights.
Reasoning and intelligence exists without language.