[ExI] More thoughts on sentient computers

[Ben Zaiboc]

On 23/02/2023 23:50, bill w wrote:

 > another question:  why do we, or they, or somebody, think that an AI 
has to be conscious to solve the problems we have?  Our unconscious mind 
solves most of our problems now, doesn't it?  I think it does.  bill w


That's a good question.

(If our unconscious solves most of our problems now, it's not doing a 
very good job, judging by the state of the world!)

Short answer: We don't yet know if consciousness is necessary for 
solving certain problems. Or even any problems.

Longer answer: I suspect it is necessary for some things, but have no 
proof, other than the circumstantial evidence of evolution.

Consciousness evolved, and we know that evolution rapidly eliminates 
features that don't contribute to reproductive fitness, especially if 
they have a cost. Consciousness almost certainly has quite a big cost. 
This suggests that it's necessary for solving at least some of the 
problems that we've met over the last 300 000 years (or at least for 
/something/ that's useful), or we wouldn't have developed it in the 
first place. Or if it happened by accident, and wasn't good for 
survival, we'd have lost it. So we can conclude at the very least that 
consciousness has been good for our survival, even if we don't know how.

It strikes me as noteworthy that the kinds of things that our computers 
can do well, we do poorly (playing chess, mathematics, statistical 
reasoning, etc.), and some things that we have evolved to do well, our 
computers do poorly, or can't do at all (hunting and gathering, making 
canoes, avoiding hungry lions, making sharp sticks, etc.). Perhaps 
consciousness is the (or a) missing ingredient for being able to do 
those things. Yes, arms and legs are an obvious advantage, but many 
other animals with arms and legs never developed like we did.
As the former things tend to be abstract mental things, and the latter 
tend to be highly-co-ordinated, complex physical things, maybe 
consciousness has a lot to do with embodiment, and manipulating the 
external world in complex ways successfully. Maybe Big Dog is closer to 
consciousness than ChatGPT (or, more likely, needs it more).

If Big Dog (or whatever the latest iteration of it is called) had 
ChatGPT in its head, as well as all the other stuff it already has, 
would it be able to build a canoe and use it to escape from a forest 
fire, decide where it was safe to stop, and built a hut? That would be 
an interesting experiment.

Ben
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230224/f6527e64/attachment.htm>