This is the way. At this point I’d consider it malpractice to not run your dissertation by an AI looking for both obvious errors and obvious improvements. The norm is going to shift very quickly.

I have a plan to revisit my PhD and postdoc code with Claude to see if I can make progress on some of the problems I stalled out on back in the day. Pure hobby curiosity at this point since I lack the ambition to pursue the potential applications or even more publication now, but I do wonder how much unrealized potential is buried beneath many dissertations.

PhD work is weirdly high variance. You have the least experience you’ll ever have, but also the most imagination and cognitive acuity you’ll ever have. I bet there’s a lot more rough diamonds buried in dissertations than in mature work.