Definitely sympathetic to their policy, but AI tooling and quality are changing quite fast. In a year I'd expect a modification of this as AI agents get better in virtually every possible way.

Is it though? Gains have been marginal and one of the major weak points is working on complex software