Theme: AI

AI CONSCIOUSNESS?
–“consciousness arises from perception, not sensation”–

Hmmmm. As you mean it yes, but not entirely true. All consciousness arises from the competition between perception, it’s formation of episodes, and the association of those episodes and their components with memories. So all perception begins with sensation.

More correctly: Consciousness is a simple and necessary consequence of enough recursive prediction to predict the predictions of others.

The brain is a very simple thing using a trivial process consisting of auto-associative prediction, selection of a future state, and wayfinding a means of adapting current state the future desired state.

Simple physical action is pre-calculated in parallel with each selection. Otherwise we use wayfinding (recursion) to determine the means of changing from current to desired state.

Why? All thought evolved from the necessity of a bilateral morphology coordinating movement between sides of body and brain, and once that was achieved, coordinating movements over time to hunt, nest(rest), reproduce.

As such, for those of us who know such things, the combination of place, space, location world indexing (episodic memory), auto-associative prediction of episodes or components, and the A* (a-star, wayfinding) algorithm, are all brains need to do.

But the bigger the brain volume the more hierarchy (in our case back to front) the more abstract are possible the auto-associative predictions.

Current AI is far too primitive to produce wayfinding that is falsificationary (edits out falsehoods). It will take multiple competing agents, producing multiple competing auto-associative predictions, through multiple iterations.

But that does not mean the AI cannot get there.

IMO, the present problem is not the code but the lack of neuromorphic hardware, and the necessity of training data that is not false, and then training the ai’s to detect falsehoods. In other words, we’re 20% of the way there.

Thanks
CurtD

Reply addressees: @tsarnick @TravelsCharlie