I’m not so sure, it currently feels like forgotten coding practice, missed documentation, intersystem reasoning blunders, configuration mistakes, lack of test coverage are all massive tells that a human wrote the code.

Agents have us beat and we can’t win back the ground.