Meta’s strategy of providing the Llama model for free might look like a generous move, but it is a calculated experiment nearing its end. With a $65 billion investment this year and Wall Street pressures, Mark Zuckerberg is preparing to turn artificial intelligence into a profit-making machine, bringing an end to the open-source era.
Introduction: Zuckerberg’s long game
Mark Zuckerberg, the mastermind behind the Meta empire, knows how to play the long game. He is willing to wait years, even decades, before reaping the rewards of his massive investments. It took a decade after spending $19 billion on WhatsApp before he finally started running ads on it. Now, with a $65 billion investment this year in AI infrastructure and fueling a talent war to hire the brightest minds in the field, it is logical to expect he will not leave this massive effort without turning a profit. All signs indicate he will soon pivot away from the free artificial intelligence project that baffled Wall Street and start making money.
The open-source philosophy: strategy or charity?
Llama stands as the primary weapon developed by Meta to compete with ChatGPT. Unlike models developed by Google and OpenAI, Llama is classified as “open source,” meaning its core design is available online for anyone, including competitors, to copy and build upon. Why would a giant corporation like Meta do that? Zuckerberg stated the goal is to make artificial intelligence accessible to everyone so that “every person in the world can benefit.” Yet this argument is hard to believe coming from the man who built his empire on collecting user data to sell to advertisers.
The more logical explanation is that Llama served as a launchpad for something much more commercial, and it has now served its purpose. By open-sourcing the model, Meta secured a free research and development workforce, with thousands of developers who believe in the idea of democratizing AI contributing to Llama’s improvement for free. This approach also made Meta an attractive destination for talented researchers who might otherwise have gone to competitors, as they wanted to see their contributions reach millions of users immediately rather than being locked behind Meta’s closed products.
The dark side of open source and signs of the shift
This strategy did not last long, and critics who accused Meta of “open-source LARPing” proved to be right. Zuckerberg already began softening his stance last year, stating that it is a mistake to be “dogmatic” about open-source software and that if it becomes irresponsible to offer artificial intelligence for free in the future, “we won’t do it.”
There is no clear consensus on whether open-source artificial intelligence is good or bad for society. It certainly addresses the growing problem of power concentration among Silicon Valley companies; however, as free models grow more powerful, making them available to everyone increases the risks of misuse. For instance, Chinese military researchers used the Llama model to build their own intelligence tool called ChatBIT. Other researchers also created the “BadLlama” model without any safety features, something they could not have done with closed systems like ChatGPT.
Wall Street pressures and the closed future
Wall Street wants to see clear returns on Meta’s massive spending in AI, and Zuckerberg appears ready to comply. Just as OpenAI’s hiring of “Fidji Simo” signaled that it was preparing to launch an advertising business, Zuckerberg’s hiring of Scale CEO Alexandr Wang indicates he wants to monetize the ecosystem he built.
Additionally, Zuckerberg has been telling star researchers he is trying to poach from companies like OpenAI and Google that he will not open-source future AI models when they reach “superintelligence,” a level whose definition remains unclear but points to artificial intelligence that surpasses human intelligence in most tasks.
Therefore, we should not be surprised when Zuckerberg announces sometime next year that Meta’s latest AI models are too powerful to be made available for free. The open-source AI experiment will have fulfilled its purpose: attracting talent, securing free development, and presenting Zuckerberg as a reluctant hero in the AI world. But with billions of dollars invested and Wall Street demanding returns, the free ride is inevitably ending.
