> I'm not sure I agree with this. Every neural net is different. Beyond a
> very simple level, they will all learn information slightly ifferently
> and the structures in their brains will be different. It may be
> very hard to export data from one AI to another unless backwards
> compatibility is one of the design requisites of the new AI, in which
> case you won't be able to put any funky new features in it (any more than
> putting monkey neurons in your brain would work - the structures are
> almost certianly far too different,
> and the arrangments caused by learning will also differ vastly)..
Ah, I see. You assume that AI must be based on neural nets. Hang on while I shift gears...
My previous comments apply to algorithmic AI - more advanced versions of today's expert systems, for example - or to systems that can use either approach as needed. For a purely neural net system the parameters are different - I'm not sure if it would transcend or not.
However, it still don't share our learning disabilities. Check out the
links Doug Bailey just posted
(http://www.news.bbc.co.uk/hi/english/sci/tech/newsid_250000/250343.stm if you missed it) for an example of modern neural nets. A brain built on their system can exchange knowledge with any other brain based on the same system via what amounts to a direct data dump - in other words, you can copy skills from one brain to another. It also learns very fast - each module can go from blank to fully optimized in less than a second.
Such systems may or may not be capable of a fast self-enhancement process, but they can go from childhood to adolescence almost instantly. Once you get a human-equivalent version and teach it to read, it will be able to learn technical skills as fast as it can process data (which is likely to be very fast indeed). Social skills will take longer, since you need interaction to learn, but it will still be much faster than a human.
Billy Brown, MCSE+I