Announcement

Collapse
No announcement yet.
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    My theory has always been that if you understand something well, AI can make you better. But if you just use AI as a substitute for learning you’ll never learn enough to use AI well. Too many students just use AI to do the work for them. This is a particularly horrifying example.

    https://www.theatlantic.com/technolo...cation/684088/. I’m a High Schooler. AI Is Demolishing My Education

    -------------------------------------------
    Richard Williams
    Professor Emeritus of Sociology
    University of Notre Dame
    StataNow Version: 19.5 MP (2 processor)

    EMAIL: [email protected]
    WWW: https://academicweb.nd.edu/~rwilliam/

    Comment


    • #17
      I used Claude to recode psacalc for reghdfe with some other modifications. It took some back and forth, and could be cleaned up a bit, but it works in the simplest cases. I fed it all the code for psacalc, and then directed Claude to rewrite for reghdfe where the contrained and unconstrined models are specified and stored. Of course, one could extract the covariates for the unconstrained model, but I just gave it the two models (more a hausman-test style approach). https://www.statalist.org/forums/for...acalc-from-ssc

      I suspect it would have taken many hours to recode on my own. Got this done in less than an hour.
      Last edited by George Ford; 27 Sep 2025, 08:25.

      Comment

      Working...
      X