The next big breakthrough will be AIs learning on the job

Knowing whether an AI succeeded is only part of making a task suitable for training. In this essay, Dwarkesh Patel argues that repeated practice also needs to be affordable and reproducible. A coding exercise can run in many separate copies of the same software environment. A real-world business decision may take months to produce feedback, and the world cannot simply be reset for another attempt. That difference motivates his interest in systems that learn from a small amount of experience. It extends his argument about retaining lessons from work: progress must include learning effectively from situations that cannot supply unlimited practice.

“You can’t have a thousand agents go try the same checkout flow on Amazon.com. Because Andy Jassy will find and detect your bots and shut your ass down.”

Dwarkesh Patel

“We’ve got some genius grad student who has never been allowed to take an internship.”

Dwarkesh Patel

“Around 30-50% of a lab’s compute goes to inference, and that compute is currently not really doing anything productive in helping improve the model. What a waste!”

Dwarkesh Patel

“the way you get better at your job is not by recalling the transcript of what happened through every single day with perfect fidelity”

Dwarkesh Patel

“I just don’t think you can accumulate new skills by passing yourself notes.”

Dwarkesh Patel