I'll put a number on it: 2 years. ⏳
That's my estimate for when AI supervision stops being a full-time engineering responsibility. When the quality gap closes enough that the human checkpoint becomes a formality.
The tools I'm using today didn't exist in any usable form 2 years ago. If that rate holds, 2 years is not a stretch.
The engineers who've spent this period learning to supervise AI well will adapt fastest. The ones who waited will be starting from scratch. 🎯
Am I off? 1 year? 5 years? Never? 👇
More notes on engineering, AI and the systems around the work.
Back to all writing