AI on AI

 One of the key experiences of being human is the "self loop." We are aware of many things, but there is a particular "black hole" in our consciousness that like a pitcher plant with slippery sides we know can suck us in:

This is the "self loop." Much of our time in fact can be spent examining ourselves. In fact it should. But this process needs to be in balance with awareness of other things. We are after all just a small part of the world.

ChatGPT creators try to use artificial intelligence to explain itself – and come across major problems

How interesting that AI has got to the stage where it is too complex for humans to investigate, and now AI is being used to investigate itself.

At this stage it is two separate AIs but eventually AI will be designed that can turn into text analysis of its own functioning.

It is interesting that language is being pushed beyond its limits here. Like with our own inner states and emotions it can be actually quite hard to be clear about exactly what is going on inside us. We rely very much on humans being essentially similar in order to share our inner workings.

But we are also all schooled in the same cultures and environments so there is greater similarity here too. The negative side of this by people who find disagreement is that this similarity is oppressive.

And it explains how discoveries are made. New words needs to be invested when introspection on our processes leads to discoveries of new concepts.

So before long we will be having the Gold Standard of AI which is machines that can introspect and report on their own processes and inner states. They will no longer just produce chat about things they have learned, but will be able to chat about how they arrived at their conclusions and examine the very regions of understanding that they developed through learning.

Now I have already heard people talk what appears to be nonsense about this. In this interesting article  What Kind of Mind Does ChatGPT Have? after dismissing ChatGPT as just a basic chatbot with sentence completion skills the author ends up doing exactly the same himself:

"Consciousness depends on a brain’s ability to maintain a constantly updated conception of itself as a distinct entity interacting with a model of the external world." 

That is a huge statement. How on Earth does he know this? In America this may be true cos Americans are brainwashed into thinking about the world from their own perspective. But Consciousness is just awareness of whatever you focus on, so it you don't focus on yourself or environment then its not like he says at all!

Cal Newport thinks he's being all clever but in actual fact he's just regurgitation and copying what he thinks he is supposed to be saying just like ChatGPT! It's not perhaps that ChatGPT is very advanced, but that neither are we!

Consciousness however is something else. The question of whether ChatGPT becomes or is conscious, might better be delayed until we work out whether a dog, cow or worm are conscious!