[Linkpost] How To Get Into Independent Research On Alignment/Agency
By Jackson Wagner @ 2022-02-14T21:40 (+10)
This is a linkpost for “How to get into Independent Research on Alignment Agency”
John Wentworth is an AI alignment researcher who works to refine our concept of what it means for a system to have “agency” and how that relates to future attempts to build aligned AI. (You can read more about his research agenda here.) In his post, he describes:
- What’s fun about the job — “I’ll make about $90k this year, and I set my own research agenda. I deal with basically zero academic bullshit, and I publish the bulk of my work right here on LessWrong/AF. Best of all, I work on some really cool technical problems which I expect are central to the future of humanity.”
- A detailed program for how you too could get into the business of alignment research, including what topics to study in the beginning, how to apply for grants, and how to get community engagement with your ideas.
- Who might be a good fit — having a grad-level education in an adjacent technical field like math or computer science is probably helpful, but more importantly: “if your reaction to [the idea of being an independent AI researcher] is ‘Where can I sign up?’, then this post is for you.”
He also shares insightful tips about:
- The importance of explaining your ideas and overall research agenda as a way to create shared understanding in the “preparadigmatic” field of agency: “For the field, such discussion is a necessary step to spreading ideas and eventually creating a paradigm. For you, it’s a necessary step to get paid, and to get useful engagement with your work from others.”
- How to interact with value-aligned EA grantmakers: “A low-bullshit grantmaking process works both ways. The LTFF wants to do object-level useful things, not just Look Prestigious, so they keep the application simple and the turnaround time relatively fast. The flip side is that I expect them to look very unkindly on attempts to make the applicant/application Sound Prestigious without actually doing object-level useful things.”
- How to design a useful research agenda, starting with the Hamming Question and trying to find your comparative advantage within the field: “apply whatever knowledge, or combination of knowledge, you have which others in the field don’t.”