Mark Zuckerberg on Business Philosophy

The strategic self-interest was unusually explicit. Meta's business is building the best experiences, Zuckerberg wrote, so it must always have access to the best technology and never be locked into a competitor's closed ecosystem. He cited a formative experience: building services constrained by what Apple allows on its platforms, with the way it taxes developers, applies arbitrary rules, and blocks product innovations. Since Meta does not sell access to AI models, open-sourcing Llama does not undercut its revenue the way it would for closed providers, one reason, he noted, that closed providers lobby governments against open source. Precedent also mattered: Meta saved billions open-sourcing its server, network, and data center designs through the Open Compute Project, and it benefited from the ecosystems around its open tools PyTorch and React, whose communities returned improvements Meta could never have engineered alone. Openness, consistently held, had compounded before.

Zuckerberg closed the argument at the level of society. Open source, he wrote, is necessary for a positive artificial intelligence future, because AI has more potential than any modern technology to raise productivity, creativity, and quality of life and to accelerate medical and scientific research. Openness, he argued, widens access to the technology's benefits around the world, keeps power from pooling inside a handful of corporations, and lets the technology spread across society more evenly and more safely. Against the running debate about the dangers of releasing model weights, he took the position that open source AI will be safer than the closed alternative, because transparency invites scrutiny and broad access prevents any single actor from governing the technology's capabilities. The bet, as with Linux, was that the ecosystem around the standard ends up mattering more than any one model.

Zuckerberg framed the cuts inside the company's purpose rather than against it. Meta builds new ways for people to feel closer, he wrote, a fundamental human need that may matter more in a complex world, and it does leading work across advanced technologies that it distills into products: artificial intelligence for creative expression and content discovery, the metaverse for a realistic sense of presence, new media formats for richer experiences, encryption for private communication, and business tools for reaching customers. The pitch to employees was that if someone wants to invent the future or apply the best ideas at the greatest scale, Meta remains the place to do it. The efficiency program, in this telling, existed to keep that long-term technology vision funded through a difficult advertising environment, not to replace it. The savings the program produced, in other words, were earmarked for the long-horizon bets that advertising alone had been carrying.