scasper comments on Advice for Entering AI Safety Research

[scasper]

I think not. Maybe circuits-style mechanistic interpretability is though. I generally wouldn’t try dissuading people from getting involved in research on most AIS things.