Fast progress artificial intelligence

[robheus]

Although there is over the years some significant progress in artificial intelligence (computer beating human chess player, etc.) here is a method of advancing the field of artificial intelligence very fast.

The idea is just this: brain cells, the neurons, even when the person is already dead, can be submerged into a fluid, and can still grow new dendrites (connections between neurons), so to advance very fast in artificial intelligence would involve submerging the individual brain cells of some people that have died, and equip the submerged neuron mass with the right kind of interfaces to electronic equipment, and then study how neural pattersn grow depending on what kind of input is given to the neural mass.

I've come up with this kind of idea after having seen an experiment in which a rat brain was treated like this, and which - after being connected to a computer interface - learned to fly a virtual airplane.

See: 'Brain' in a dish flies flight simulator

 

[pallidin]

Indeed, there is a huge potential in organic computing.
Lot's of hurdles, though.

 

[Negatron]

Indeed, there is a huge potential in organic computing.
Lot's of hurdles, though.

None whatsoever, unless by organic you mean carbon-based (true definition) rather than biological. Computer systems fly flight simulators a hell of a lot better than dish brains, and artificial neurons are vastly faster and more efficient than biological neurons and can be manipulated and connected in arbitrary ways. Bio neurons are a dead end. The best thing to get out of them is an inspiration for developing practical artificial neural systems.