upvote

points

by diehunde8 hours ago |
comments
upvoteby hackinthebochs8 hours ago|
next
[-]
>LLM can’t beat an avg chess player.

Why should that matter?

reply
upvoteby janalsncm7 hours ago|
parent|
next
[-]
If something has general intelligence it should be able to read the rules of a game and follow them. Therefore an artificial general intelligence (AGI) should be able to do this.

So we have a situation where very powerful and influential people are saying we will have AGI in 6 months (if we don’t already), yet the facts on the ground are so clearly pointing in the opposite direction.

reply
upvoteby Gregkion47 minutes ago|
parent|
next
[-]
So we humans are not a general intelligence then?

And the stuff i'm using LLMs daily is just fake?

I see i see. I will see myself out of this weird discussion while I let an LLM continue doing a lot of interesting things.

reply
upvoteby dosisking8 minutes ago|
parent|
[-]
> And the stuff i'm using LLMs daily is just fake?

It simply means that LLMs are smarter than you, but not smarter than the average person

reply
upvoteby hackinthebochs7 hours ago|
parent|
prev|
[-]
I would bet a lot of money that Astra can follow the rules of chess (perhaps if repeated within the context window). Also, this is a different argument than what I responded to.
reply
upvoteby minraws6 hours ago|
parent|
next
[-]
I can write you a benchmark to prove it even with a heavy handed system prompt Astra will make an illegal move during the course of the games first few moves are generally ok since it's just throwing out learned moves.
reply
upvoteby hackinthebochs3 hours ago|
parent|
[-]
I'd genuinely like to see the results of that.
reply
upvoteby janalsncm5 hours ago|
parent|
prev|
[-]
I would definitely take you up on that.
reply
upvoteby lelanthran3 hours ago|
parent|
prev|
[-]
> Why should that matter?

Because we want to use this as a replacement for humans, and the average human can learn the rules of chess without needing to see the rules explained hundreds of thousands of times in millions of games.

So, yeah, it matters if a model has millions of examples of something in its training set and still cannot follow the rules.

reply
upvoteby hackinthebochs3 hours ago|
parent|
[-]
We're not talking about learning the rules of chess here, but playing a competent game from just being shown the rules. Why is it so hard for people to keep track of the thread of discussion?
reply
upvoteby ncruces39 minutes ago|
parent|
[-]
But we are. The models can't even follow the rules: they try illegal moves all the time.
reply
upvoteby lostmsu7 hours ago|
prev|
[-]
The fact that LLMs can play chess at any level is a strong indication we are in AGI.
reply
upvoteby recursive7 hours ago|
parent|
next
[-]
Can they if they frequently make illegal moves?
reply
upvoteby bigstrat20036 hours ago|
parent|
prev|
[-]
No it isn't. Computers could play chess long before LLMs, better than LLMs can in fact. That didn't make them AGI.
reply