[ExI] Self improvement was Did Hugo de Garis leave the field?

Eugen Leitl eugen at leitl.org
Sun Apr 24 11:58:43 UTC 2011


On Wed, Apr 20, 2011 at 10:12:41PM -0700, Keith Henson wrote:
> Thought experiment.
> 
> Build an AI.  Put a clock rate control in that it can mess with.

Advanced computers won't have global oscillators, being more
asynchronously coupled, free-running assemblies of signalling 
systems, but for the purposes you could implement a throttling
option which would be equivalent to a clock.
 
> If it is measuring it's performance against an outside clock, it is
> going to notice that it does better with the clock rate on fast.
> 
> So it sets it's clock rate to fast, effectively becoming smarter.

Yes, but a dog running for a megayear won't come up with quantum
cronodynamics. Wait, I won't come up with that, either.
 
> Seems to me this is self improvement.

If you have more time you can sample the problem space more
thoroughly, but this  obviously has a ceiling, unless you're trying
to make a smarter system, and are actually making progress.

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE



More information about the extropy-chat mailing list