Sapience sentience AI and human development

[Denny Crane!]

Bear with me while I work this one through as I write it.

At what point did we differentiate ourselves from apes or dolphins with high intelligence levels in terms of sapience. At what moment did we become self aware and what triggered that change? At what stage in our development did it happen?

What is it that drives our thoughts, that shapes us and drives us to learn more?

At our most basic levels are thoughts and actions are determined by our instincts. This shapes our thought patterns and how we prioritise our thoughts, so where does the desire to learn come from. It is not an instinct driven function on the higher levels of pursuit.

And how in all this would this transfer over to a computer, how could it ever achieve artificial intellignce as it learns only by being programmed to measure input. Is there a difference between a computers programming and a brain in terms of input structure and ability to learn. Could the desire be replaced by a preprogrammed set of instructions?

I've just thrown together a dozen or so jumbled questions from saturday night to see if I can make any sense of them. I'm guessing it made more sense then. If you see a part you can comment on please do, particularly the development of the human race into sentient sapient beings.

Peter

[Gaidin]

And how in all this would this transfer over to a computer, how could it ever achieve artificial intellignce as it learns only by being programmed to measure input. Is there a difference between a computers programming and a brain in terms of input structure and ability to learn. Could the desire be replaced by a preprogrammed set of instructions?

If the input sensors are programmed correctly there is no real difference in the input structure. We just have to tell it what it is seeing before it will understand that. Pretty much the same as any human.

Desire and emotion in general is a different matter for programming as you can't really quantify them in a basic set of ones and zeros. The human brain deals with a broader range of 'values'.

[MoROmeTe]

There's a view that sapience exists because of the sheer number of connections between neurons in our brains...

Conscience, that is quality, is a derivative of numbers, that is quality. So practically an AI needs to have a lot more connections between different operation oriented entities (aka processors). I believe I read a while back that progress was being made with large "neural nets"...

I am now off to search for more information...

[Denny Crane!]

There's a view that sapience exists because of the sheer number of connections between neurons in our brains...

Conscience, that is quality, is a derivative of numbers, that is quality. So practically an AI needs to have a lot more connections between different operation oriented entities (aka processors). I believe I read a while back that progress was being made with large "neural nets"...

I am now off to search for more information...

So the development of our brains is the reason why we acheived self awareness? Does that mean that computers once they reach a certain level will develop the same way without our input or dependant upon programming input.

What kick started the change, what "programmed" humans to suddenly become more than animals? What was the stimulants that changed the homo sapien from that of the primate or neanderthal.

Peter

[ZaPPPa]

Programming AI should follow the same way life and the brain developed.. By evolving. AI needs to make random decisions and measure the succes of its decision. Based on that it can make decision tables for every known situation. If an unknown situation arises, the AI can check and combine the closest resembling known situation and make a decision based on that. Then the new situation is added to the known list. Over a number of years (probably decades) the AI can program itself.

Humans do come with pre-programmed instructions like how to recognize emotions in other humans. All these instructions need to be programmed in the AI. These programmed instructions will be the footprint that will help to determine success of its actions. Any other instructions like morality should not be pre-programmed in my opinion. An instruction like: "Thou shallt not kill" can lead to impossible decision-deadlocks in which the AI cannot continue to function because it is put in an impossible situation. Humans are not pre-programmed not to kill, so AI shouldn't either. AI should learn that killing someone is a bad idea in most cases, but not in all.

As for difference in AI and a human brain: AI has an infallible memory, but will act like an infant when it comes to making decisions. The human brain has poor memory, but is highly adaptable. It is going to take many decades before AI will be anywhere near human level.

[MoROmeTe]

Lat time I heard there were some experiements involving interconnected computers in which several of them were fed false data and the entity was asked to self regulate... AI cannot self regulate as we do, going through conflicting sets of data and choosing one based on criteria. But the AI seems to be able to understand how someone else aka human solves such problems and replicate the behaviour.

So I believe it will take time till some AI is so learned in mechanisms of problem solving that the humans use and becomes able to create new problem solving mechanisms out of the ones it knows... When will this happen? Some say 2012...