Comment by YeGoblynQueenne | Hacker News Reader

Comment by YeGoblynQueenne | Hacker News Reader

YeGoblynQueenne

>> Consider STRIPS planning problems.

The thing to keep in mind in all these discussions about LLM planning and reasoning is that, when experts on planning and reasoning say that LLMs can't do planning and reasoning they have a very different, formal definition of those things in mind, than everybody else who hasn't studied them. Even some of the research papers on planning and reasoning with LLMs play very fast and loose with the use of those words and that's probably why there are so many positive results that later prove to be duds.

Now is probably a good time to re-educate computer scientists and AI researchers about planning and reasoning.

Wikipedia has a very short introduction to STRIPS planning:

https://en.wikipedia.org/wiki/Stanford_Research_Institute_Pr...

And here's a more comprhensive, but still short, introduction to automated planning and scheduling in general:

http://www.spacebook-project.eu/pubs/CogSci-13.pdf

This is exactly right.

Planning and reasoning have long standing meaning in AI, based on explicit knowledge representation and inference. It's a subspace of AI that predates LLMs by decades.

But LLMs don't do that kind of planning and reasoning, and CoT is very clearly not a reasoning technique in the AI sense.

Isn't the obvious approach to use thr LLM to write the code or definition for the STRIPS system and then use that system to search for the plan?

Have they tried that?

And there comes the fact LLMs are actually terrible at coding beyond advanced regurgitation, precisely because their single-pass architecture does not give them Turing completeness to model a Turing machine, which may also be the reason they model the full chain of cause and effect to formulate a coherent plan...

Turing machines are an irrelevancy for programming, and LLMs can code ... not well, but at a higher level than regurgitation. ("Advanced" regurgitation can, of course, describe anything.)

There are people working on that! Google for LLM + PDDL.

There are big problems with hallucinations because LLMs are not smart enough to know when they're starting to make mistakes.

But there's lots of work in this area, and generally in different ways to nail neural and symbolic systems together.

Reply on news.ycombinator.com