Thanks for all this! I'm not familiar with AI safety, and even if some concepts are used both in AI and economics, I suspect there would still be a lot of retraining involved, but I could be wrong. 

I'll take a look at the blog posts you mentioned! 

Thanks for this! 

I'm in my first-year so I haven't specialized in a subfield yet. I've checked out GPI's research agenda but most economics-y questions look very abstract. 

Thanks a lot for your comment. What you describe is a different route to impact than what I had in mind, but I suppose I could see myself do this, even though it sounds less exciting than making a difference by contributing directly to making AI safer. 

Thanks for sharing your experience and thanks your comment! This is very useful! 

For most people in a PhD in any long-term relevant subject (econ, biology, AI, stats), with a chance of a tenure-track position at a top-20 (worldwide) school, I expect it will make sense to push for that for at least ~3 years, and to postpone worries about pivoting until after that. Because switching subjects reduces those odds a lot.

I had two follow-up questions: first, do you think there is a big difference in impact between getting a tenure-track position at a top-20 school vs medium-ranked school? If so, where does the difference come from? 

Secondly, why do you think switching subjects reduces those odds a lot? Do you think it's because it's unlikely that I would get accepted into a PhD program in AI or because, even if I'm accepted, I'm less likely to get a tenure-track position in this field?