Let's hear it, though if you don't have a degree in computer science and a deep knowledge of human psychology, it's likely flawed. Tons of people declare such theories, and to date the only one that appears to be working is the many agents model used by novamente.

*ROTFL*
good one!!!

I think in this you are wrong. A "general purpose thinker", no matter if created by random (biological evolution) or by design is capable of doing anything we can, this includes understanding its/his own design and reproducing the concept in some other way.

Don't you at least agree that by simulating 10^21 spiking neurons and linking them to a honda robot (upgraded with more precise and additional sensors like pain, smelling etc.) and subsequently raising him as if he were our own child would create a mental equivalent of a human being? And if that is so, and if there were a valid simplification so that only a tiny fraction of processing power would be needed, wouldn't that counter your argument?

You will notice I didn't try to simplify the "raising him as if he were our own child" part. This even I believe to be near impossible.

That's what i'm wondering

Actually, it wasn't a joke!

To my wee small tiny mind my computer is amazing. My computer is able to link me the internet. I am able to type messages that people can read on the other side almost instantly - my computer gives me access to masses of information through databases.

My computer isn't able to apply reason.... but this doesn't mean that my computer isn't intelligent.

yeah, this is the ability to analyse and apply reason to come up with a solution.

Yes, but as you suggest, science is continual in its development.
My mind boggles!!
We need people who can apply their minds and focus intensely to find answers.
Fear shouldn't be a reason to stop to flow of new knowledge.

good, it's a fascinating topic.

The computer must first be programmed to tell it HOW to analyse the data.

What we need for strong AI is essentially a program-writing program.

The best description i've ever seen is that programmers already use tricks such as macros and meta-programming and high-level languages to simplify their jobs, in an attempt to automate part of the coding. A strong AI would be the ultimate form of this automation, replacing the programmer completely.

Excellent quote!!!! This is the idea which I started with!

Unfortunately it soon turned out to be a little more complicated. But still it proved to be one of the central ideas in the complete solution.

This is also an excellent quote and you are right. But I am not sure if you know why you are right.

Why should we need anything else than ANNs if the human brain doesn't need anything else?

With ongoing development, do you see replacement of the programmer with AI could be possible Gareth?

FWIW ... for what it's worth. With the implied suggestion .....my thoughts don't count for much.

*FEELINGSCHIZOQUOTINGMYSELF*

but what if you can design a software virtual body, which whilest running, can add code to itself? (this actually is easily possibly) and what if this software extends itself with learning methods which are evolutionarily created to solve the problems the virtual body stumbles upon?

This actually IS the solution, you "just" have to fully understand it and to fully understand the human brain to see that it actually IS the solution. unfortunately plastically understanding the human brain (easier by understanding its development and genesis) is the most complicated part of the process.

PS: I realize two important things are missing in this simplified idea:

1. How to generate practical learning methods automatically. This I cannot explain without a huge amount of background information. Also this is the really work intensive part of the solution.

2. It has to be guaranteed by design of the "virtual brain" that the entity will develop a rich self representation to have an ego. Otherwise it will be a general problem solver and nothing more. We humans are more: we have consciousness. An identity. Qualia. A will. Desires. A personality. And so on. For this I have a solution as well. I will NEVER openly reveal it. But anyone fully understanding the neocortex (I guess hardly anyone does and the ones who do are seen as nutcases like I am) will know the answer immediately.

are you referring to the Tech. Singularity?

The birth of strong AI would itself trigger the singularity as soon as it could self-modify, the very definition of the singularity is this.

I believe that love is not an emotion.

The human brain is just an organic machine, and it's not the only means of doing cognition. You don't need love to think either, the only emotion that could play a serious role in a useful AI would be pleasure for the reward mechanism when training.

err, what is it then?

I don't think that it even exists. I think that people just call attraction "love" if it goes beyond physical superficial attraction.

It's still an emotional state, consisting of very strong attachment to another person - not always sexual either. But regardless, you don't need love to think.

Bots in games don't do any of these things because they aren't strong AIs.

It's a rather complex set of neural circuits firing, which is essentially the same thing.

one of the best statements on this page. though I think that at least two kinds of emotions would be needed for cognition, a positive and a negative one. (pleasure and pain) but lack of pleasure could suffice.
in reality there will always be four kinds of emotions:

- a positive emotion as result of a positive event (P+)
- a positive emotion as result of the absence of a negative event (P-)
- a negative emotion as result of a negative event (N+)
- a negative emotion as result of the absence of a positive event (N-)

they all feel distinctively different in humans and should do so in a strong ai which is "human-like".
each one of those can have an intensity and "color", where color is an abstact placeholder for the complex associations related to the actual event / situation causing the emotion.

if that is so, then every now and then a computer program can step out of its own circuitry.

That still happens due to the processing within the brain, there's nothing "magical" about the mind, it's just a very complex organic machine.

No it doesn't, it's still just blindly following orders (code). If the code is buggy and tells the CPU to do something stupid, it'll do something stupid, this sometimes results in crashing.