This is the one thing more valuable than data in AI, says Meta CEO Mark Zuckerberg

Meta flexed its AI muscle with the launch of Llama-3 models and integrated it into WhatsApp, Messenger and Instagram. Mark Zuckerberg, CEO, Meta, is buoyed by what the company can do in the AI race. In an interview with The Verge, Zuckerberg said that data is not the most valuable thing when it comes to training AI models. “The thing that I think is going to be more valuable is the feedback loops rather than any kind of upfront corpus,” Zuckerberg said in the interview.

To train, retrain and improve AI models, feedback loops are used based on their previous outputs. These loops are algorithms that inform and educate AI models when they are making mistakes. Furthermore, they give the models the data which is required to boost their performance. “Having a lot of people use it and then seeing how people use it and being able to improve from there is actually going to be a more differentiating thing over time,” Zuckerberg told the publication.

Synthetic data is something that Meta and other players in the AI race would be looking at. As the name suggests, this data is artificially generated and replicates data based on real-world events.

“I think there's going to be a lot in synthetic data, where you are having the models trying to churn on different problems and see which paths end up working, and then use that to reinforce,” said Zuckerberg.