So cool to see such a thoughtful and clear writeup of your investigation! Also nice for me since I was involved in creating them to see that 80k’s post and podcast seemed to be helpful.
I think [advising on hardware] would involve working at one of the industries like those listed above and maintaining involvement in the EA community.
What I know about this topic is mostly exhausted by the resources you’ve seen, but for what it’s worth I think this could also be directed at making sure that AI companies that are really heavily prioritising safety are able to meet their hardware needs. In other words, depending on the companies it could make sense to advise industry in addition to the EA community.
University professor doing research at the cutting edge of AI hardware. I think some possible research topics could be: anything in section 3, computer architecture focusing on AI hardware, or research in any of the alternative technologies listed in section 4.
Industry: See section 4 for a list of possible companies to work at.
For these two career ideas I’d just add—what is implicit here I think but maybe worth making explicit—that it’d be important to be highly selective and pretty canny about what research topics/companies you work with in order to specifically help AI be safer and more beneficial.
These experiences will probably update my thoughts on my career significantly.
Seems right—and if you were to write an update at that point I’d be interested to read it!
So cool to see such a thoughtful and clear writeup of your investigation! Also nice for me since I was involved in creating them to see that 80k’s post and podcast seemed to be helpful.
What I know about this topic is mostly exhausted by the resources you’ve seen, but for what it’s worth I think this could also be directed at making sure that AI companies that are really heavily prioritising safety are able to meet their hardware needs. In other words, depending on the companies it could make sense to advise industry in addition to the EA community.
For these two career ideas I’d just add—what is implicit here I think but maybe worth making explicit—that it’d be important to be highly selective and pretty canny about what research topics/companies you work with in order to specifically help AI be safer and more beneficial.
Seems right—and if you were to write an update at that point I’d be interested to read it!