Live data from Hacker News

Yann LeCun does not possess an internal monologue

twitter.com

1–10 of 17 posts

Re: Yann LeCun does not possess an internal monologue

#5
this just seems like some Wittgenstein mistake

>"thinking and reasoning require language"

we are just going to argue about what thinking and reasoning are because the sentence is meaningless.

Clearly animals think and reason and dont have language.

Or clearly thinking and reasoning is a process of symbol manipulation contingent on language itself because nothing else provides the means to ground the symbols nor the rules to encode acceptable manipulations.

Re: Yann LeCun does not possess an internal monologue

#6

this just seems like some Wittgenstein mistake >"thinking and reasoning require language" we are just going to argue about what thinking and reasoning are because the sentence is meaningless. Clearly animals think and reason and dont have language. Or clearly thinking and reasoning is a process of symbol manipulation contingent on language itself because nothing else provides the means to ground the symbols nor the r…

If you've lived with animals, of course they have language - it's what we would call (from a human-centric perspective) "non-verbal". My two dogs absolutely understand and can reason from certain words they learn, one I've raised as a puppy from day 0.

Re: Yann LeCun does not possess an internal monologue

#7

I wonder how many "leaders" in AI don't have one, and therefor don't understand reasoning the same way as others do who do have it (FWIW I have Aphantasia, although it doesn't mean I can't imagine - I just can't see it).

The converse is also true. I’m sure people with internal monologues (I’m not one) are much more optimistic about LLMs.

Re: Yann LeCun does not possess an internal monologue

#8

I wonder how many "leaders" in AI don't have one, and therefor don't understand reasoning the same way as others do who do have it (FWIW I have Aphantasia, although it doesn't mean I can't imagine - I just can't see it).

The converse is also true. I’m sure people with internal monologues (I’m not one) are much more optimistic about LLMs.

Oh I have usually more than one internal monologues going on, but I also think LLMs are just Markov chains with more infrastructure and way over-hyped (because I can reason this)

Re: Yann LeCun does not possess an internal monologue

#9

I wonder how many "leaders" in AI don't have one, and therefor don't understand reasoning the same way as others do who do have it (FWIW I have Aphantasia, although it doesn't mean I can't imagine - I just can't see it).

It's hard to tell if one does or doesn't have an inner monologue/dialog.

It's the standard "is the color I call 'yellow' the same yellow color that you call 'yellow'?" question.

We have no frame of reference for what an inner monologue is really like for other people.

How many gradations are there? 0 = none, 1000 = max (I can hear my voice in my head as clearly and cogently as you can hear my voice when I'm explaining something to you?)

0 = none, 3 = max, and almost all humans are either a 0, 1, 2 or 3?

Until we get neuralink's descendent implanted and can share each other's consciousness, we won't know for sure. And probably not even then.

Re: Yann LeCun does not possess an internal monologue

#10

I wonder how many "leaders" in AI don't have one, and therefor don't understand reasoning the same way as others do who do have it (FWIW I have Aphantasia, although it doesn't mean I can't imagine - I just can't see it).

The converse is also true. I’m sure people with internal monologues (I’m not one) are much more optimistic about LLMs.

> I’m sure people with internal monologues (I’m not one) are much more optimistic about LLMs.

Can I ask why you're "sure" about this? I have an internal monologue and I'm quite skeptical of the LLM hype.

Post reply on HN