[extropy-chat] SIAI seeking seed AI programmer candidates

Eliezer Yudkowsky sentience at pobox.com
Wed Jun 2 09:42:27 UTC 2004


Eugen Leitl wrote:
> 
> So you have a Ph.D. in three different disciplines (that be neuroscience,
> computational and/or wet), practical hardware design (at least at FPGA level,
> preferrably ASIC), have practical experience with high-performance numerics 
> codes >kNode country, and can handle complex nonlinear dynamics?
> 
> You can assume that those scientists and engineers can program.
> High-performance parallel numerics, that is, which most programmers have
> absolutely no clue of.

This is SIAI's project, Eugen, not yours, don't embroider our job 
requirements.  I don't currently expect to require anything special in the 
way of hardware.  Lots and lots of cognitive science background is a good 
idea.  The main math I already know will be required is Bayesian 
probability theory and expected utility decision theory, not knowing the 
equations, but feeling them in your bones and using them to see everything 
you work with.  I once read a fragment of an online physics course whose 
introduction noted that to understand that course you needed to do basic 
integrations and differentiations with your spine, not your brain.  What 
calculus is to physics, Bayes is to AI.

I was not planning to brute-force the problem, nor to use architectures 
inspired by specific details of the human brain.  The massively 
interdisciplinary cognitive science background required is essentially the 
equivalent of being literate and having a GED, rather than being a 
primitive tribesperson from the Islets of Langerhans.

-- 
Eliezer S. Yudkowsky                          http://singinst.org/
Research Fellow, Singularity Institute for Artificial Intelligence



More information about the extropy-chat mailing list