Ramez Naam writes:
> Yes, if de Garis' neural connections are just random, there's no way he's
> going to achieve human-level intelligence, even with 1000x the neurons.
GAs are not random. I do not get his manual block assembly part, though.
> That having been said, his plans are encouraging in two respects:
> 1) Independent of getting human-level (or human-like) intelligence, a neural
> net of that size does seem to hold the promise of being a remarkably
> powerful learning machine.
Size alone is a necessary but not sufficient condition.
> 2) If he can really simulate 10^12 neurons in 2011, then that leaves the
> door open for scanning a human brain and replicating its architecture in de
> Garis' neural simulation. In 2011 presumably the only option would be a
> destructive scan, but even this would be phenomenally ahead of schedule.
On what evidence do you base this absurd optimism?
> The major question in my mind is how rich his simulated neurons are. How
> many synaptic connections do they each have? How frequently are their time
> slices? How closely can the mimic the behavior of human neurons?
Once again, his neurons are highly simplified sketches. They are
Potyemkin's villages of neurons. The tasks of creating an AI and
neuronal emulation are only distantly related. Neuronal emulation is
distinctly harder.
This archive was generated by hypermail 2b29 : Thu Jul 27 2000 - 14:02:44 MDT