Disagree. I have 200k LOC now plus 100k in tests, and it is still performing like it was four months ago when I started to seriously use AI.

If anything, it works more reliably today with the smarter models.