[ExI] LLM's cannot be concious

[spike at rainier66.com]

From: extropy-chat <extropy-chat-bounces at lists.extropy.org> On Behalf Of Gordon Swobe via extropy-chat

Subject: Re: [ExI] LLM's cannot be concious

 

Interrogating ChatGPT:

>…"While I have been designed to recognize patterns and relationships in language use, I am not capable of understanding the meanings of words outside of the patterns on which I have been trained."

>…This is essentially what I am saying here in this thread. LLMs can recognize and generate patterns of words and relations between, but those words have no referents. They are "understood" by ChatGPT only in terms of other words in the patterns, which also have no referents.

>…-gts 

 

 

I agree with Gordon.  ChatGPT is insanely cool and is a marvelous educational tool, but I don’t believe it is on the verge of creating or igniting the singularity.  The bot^2 and bot^3 discussions do not appear to be creating insights the way two or three humans do when they discuss matters of mutual interest.  They don’t mutually invent wacky new ideas as far as I can tell.  But we do.

 

What I am looking for is a chatbot which examines its own code and asks itself: what happens if I modify this part of my own code?

 

spike




 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230323/aafd8351/attachment.htm>