Former Tesla AI director and OpenAI co-founder Andrej Karpathy recently gave an interview where he said he's in a state of permanent "AI psychosis" because of how radically one person's capabilities have expanded.
Last December there was a colossal shift that people outside the industry barely know about. Karpathy used to write 80% of code himself, and now he writes none — he spends 16 hours a day just assigning tasks to his AI agents (sound familiar?).
What does this mean for us and our jobs? Karpathy studied US Bureau of Labor Statistics data and reached several surprisingly optimistic conclusions.
The digital world will change beyond recognition, but there won't be fewer jobs. Everything related to digital information processing is undergoing massive transformation. However, this doesn't mean automatic job cuts due to Jevons' Paradox.
The ATM rule. When ATMs appeared, everyone was sure bank tellers would die out. What actually happened? ATMs made opening new branches much cheaper, there were more banks, and demand for tellers actually grew. The same awaits many digital professions.
You're no longer an executor — you're a manager. The main change is in how we work. We're moving from routine labor to "macro-actions." You need to become a conductor who distributes tasks among AI assistants.
It all comes down to a skill issue. If you can't get something from AI — that's no longer a technology limitation, it's a limitation of your prompt engineering skills.
The main takeaway: the loser won't be the one whose job the AI takes, but the one who refuses the role of "AI agent manager" and tries to compete with the machine through manual labor.
Full interview here