Connectionists: Can LLMs think?

[Ali Minai]

Of course, LLMs can’t think. The more interesting question is: What are the
implications of the apparent thinking of LLMs for our understanding of
human thinking? I suspect that, as more powerful LLMs emerge, this question
will become more salient. The space of what cannot be done by AI will keep
shrinking, and all we will have to fall back on is the feeling that humans
are different - which I agree with. We will then be forced to explain much
better how, why, and in what ways is human thinking different than
something very large neural networks can ever achieve. We should be
prepared for the possibility that the answer is “none”.  After all, we have
an existence proof of a very large neural network - the human brain - that
can produce human thinking!

I suspect that, in the end we’ll find out that the brain needs a behaving
body embedded in the real world and the capacity for self-motivated
behavior to be intelligent like a human. Computers in server farms will
just keep simulating some parts of it better and better.

Best
Ali

On Fri, Mar 17, 2023 at 6:51 AM Gary Marcus <gary.marcus at nyu.edu> wrote:

> Frederick, Geoff and others,
>
> by the way I don’t know about “average people” but here is a great example
> of a bright journalist (not an AI expert or ML researcher) with perfectly
> clear recognition of how GPT-4 cannot be trusted,—and a reminder that for
> all its alleged “understanding” GPT-4 is utterly unconstrained by any
> internal process of fact-checking, which is to say that it cannot ground
> its text-pastiching process in reality, another diagnostic of
> discomprehension:
>
> [image: IMG_3988]
>
> for good measure, some subtle lies about SVB, also generated by GPT-4
> (reported by Dileep George).
>
> [image: IMG_5655]
>
> On Mar 17, 2023, at 08:48, Gary Marcus <gary.marcus at nyu.edu> wrote:
>
> 
>
> Average people were fooled by a chatbot called Eugene Goostman that
> ultimately had exactly zero long-term impact on AI. I wrote about it and
> the trouble with the Turing Test here in 2014:
> https://www.newyorker.com/tech/annals-of-technology/what-comes-after-the-turing-test
>
>
> On Mar 17, 2023, at 8:42 AM, Rothganger, Fredrick <frothga at sandia.gov>
> wrote:
>
> 
> Noting the examples that have come up on this list over the last week,
> it's interesting that it takes some of the most brilliant AI researchers in
> the world to devise questions that break LLMs. Chatbots have always been
> able to fool some people some of the time, ever since ELIZA. But we now
> have systems that can fool a lot of people a lot of the time, and even the
> occasional expert who loses their perspective and comes to believe the
> system is sentient. LLMs have either already passed the classic Turning
> test, or are about to in the next generation.
>
> What does that mean exactly? Turing's expectation was that "the use of
> words and general educated opinion will have altered so much that one will
> be able to speak of machines thinking without expecting to be
> contradicted". The ongoing discussion here is an indication that we are
> approaching that threshold. For the average person, we've probably already
> passed it.
>
> --
*Ali A. Minai, Ph.D.*
Professor and Graduate Program Director
Complex Adaptive Systems Lab
Department of Electrical & Computer Engineering
828 Rhodes Hall
University of Cincinnati
Cincinnati, OH 45221-0030

Phone: (513) 556-4783
Fax: (513) 556-7326
Email: Ali.Minai at uc.edu
          minaiaa at gmail.com

WWW: https://eecs.ceas.uc.edu/~aminai/ <http://www.ece.uc.edu/%7Eaminai/>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20230317/b74ba005/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: IMG_3988.jpeg
Type: image/jpeg
Size: 393912 bytes
Desc: not available
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20230317/b74ba005/attachment-0002.jpeg>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: IMG_5655.jpeg
Type: image/jpeg
Size: 355436 bytes
Desc: not available
URL: <http://mailman.srv.cs.cmu.edu/pipermail/connectionists/attachments/20230317/b74ba005/attachment-0003.jpeg>