Masad: Next-gen coding AI will train via AlphaZero-style RL, not human code
Amjad Masad · The Future of Software Creation with Replit CEO Amjad Masad · Y Combinator · Sep 12, 2025 · at 40:14
Amjad Masad (CEO of Replit) responds to an audience question at Y Combinator AI Startup School about model collapse when AI trains on AI-generated code.
“My bet is that Pretty soon we're going to move into more of the alpha zero style of training where you have a more traditional LM that's trained on all of the internet. But then the way to train the next generation of it would be to give it a reinforcement learning environment where, ah, it's generating a lot of problems and doing, like, self-play, where solving these problems, getting feedback on them, and doing it in this, like, massively parallel way. I think this is how we're going to get the next generation software agents. It's not going to be trained on human code, because, like you said, there's not going to be human code, and so we have to solve this.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →