OpenAI’s Latest Rivals Are Getting Help From… OpenAI

As artificial intelligence developers find it easier to catch up with OpenAI in reasoning than with traditional large language models—a topic Erin and I wrote about last week—they may be using a tried-and-true strategy.

Researchers at OpenAI believe that some rival AI developers are training their reasoning models by using OpenAI’s o1 reasoning models to generate training data, according to a person who has spoken to the company’s researchers about it. In short, the rivals can ask the o1 models to solve various problems and then use the models’ chain of thought—the “thought process” the models use to solve those problems—as training data, the person said. 

You might be wondering how rival developers can do that. OpenAI has explicitly said it hides its reasoning models’ raw chains of thought due in part to competitive concerns.