Two ideas for projects/​exercises, which I think could be very instructive and build solid instincts about AI safety:
Builder-breaker arguments, a la ELK
Writing up a safety case (and doing the work to generate the underlying evidence for it)
Two ideas for projects/​exercises, which I think could be very instructive and build solid instincts about AI safety:
Builder-breaker arguments, a la ELK
Writing up a safety case (and doing the work to generate the underlying evidence for it)