[ExI] e: GPT-4 on its inability to solve the symbol grounding problem

Ben Zaiboc ben at zaiboc.net
Tue Apr 18 09:59:00 UTC 2023


On 18/04/2023 06:59, Brent Allsop wrote:
> You said: "/There are no pictures of potatoes being sent back and 
> forth in the brain/. Instead, there are coded signals, in spike trains 
> travelling along axons."
> I believe Geiovani said something similar, when he said there are no 
> pixels in the brain.
> I think I understand what you say in that the name "potato" or "apple" 
> is a referent to a general abstract idea, rather than a specific 
> potato, which makes sense.
> But what is our subjective knowledge of the potato we see, if not a 3D 
> model (a picture?), derived through a very complex process, from 2 
> very noisy and distorted set of 2 2D pixels, from the eyes?
> And when people observe colored  "pictures" in the brian (when they 
> look at potatoes), and display what they see in the brain, on a 
> picture screen, as reported in these many papers 
> <https://canonizer.com/topic/603-Current-Observation-Issues/1-Agreement>, 
> what are they observing, if not pictures?


I'm not sure what your question is. I'm not saying that we don't see 
mental pictures, I'm saying that the pictures are not the things being 
transmitted and received in the brain. Much like in a TV, there are no 
pictures being send around the circuitry, there are electrons in wires.

The pictures are higher-level constructs.

Ben
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230418/d2ee5cff/attachment.htm>


More information about the extropy-chat mailing list