Meta is the only prominent US lab that’s been releasing open-source AI models for the last two years, and its AI Chief has given some insight into its unusual business choice.
Yann LeCun, Meta’s Chief AI Scientist, recently explained the company’s open-source approach to AI development. His argument centers around the collaborative nature of scientific progress, the acceleration of AI research through shared knowledge, and how this strategy ultimately benefits Meta’s business model, even if indirectly.
“Good ideas can come from anywhere,” LeCun said on a podcast. “Nobody has a monopoly on good ideas. So, you know, I think (AI) is still a major scientific challenge and we need everybody to contribute.”
LeCun then outlined Meta’s strategy: “So the best way we know how to do this in the context of academic research is you probably should go in open source as much as you can and you get people to contribute.” He pointed to recent history as evidence of this approach’s efficacy. “And I think the history of AI over the last 10 years really shows that. I mean, the progress has been fast because people were sharing code and scientific information.”
However, he acknowledges a shift in the field. “Some, you know, a few players in the space started closing up over the last three years because they need to generate revenue from the technology.” LeCun then contrasted this with Meta’s strategy. “Now I admit that we don’t generate revenue from the technology itself. We generate revenue from ads, and those ads rely on the quality of products that we build on top of the technology. And they rely on the network effect of social networks and the conduit to the to the people and the users. And so the fact that we distribute our technology doesn’t hurt us commercially. In fact it helps us.”
LeCun’s explanation provides valuable insight into Meta’s strategy. By open-sourcing its AI models, Meta fosters a collaborative environment that accelerates the overall development of AI. This rapid progress, in turn, allows Meta to build better products, enhancing user experience and strengthening the network effect that drives its advertising revenue. While some companies see AI as a direct revenue stream, Meta views it as a foundation for improving its core products and services. This approach allows them to benefit from collective advancements in AI while potentially hindering competitors who have adopted closed models and are therefore unable to benefit from the same synergistic approach.
It also likely helps that Meta faces the least amount of disruption from AI compared to other tech giants — Google’s search business could be threatened by AI, as could Microsoft’s many software products. Meta relies mainly on network effects of having billions of human users, and won’t likely be particularly impacted by AI. As such, Meta likely wants AI to progress as rapidly as possible, and doesn’t mind sharing its research with other companies. Also, by open-sourcing its models, Meta hits its rivals where it hurts most — companies like Google and OpenAI spend billions of dollars into producing their AI models, but with Meta giving away its own models for free, it depresses the price-point for closed-source AI models. Meta’s decision to open-source its models will ultimately benefit the field of artificial intelligence, but it also makes plenty of business sense for the company to go the open-source route with AI.