Meta’s Future AI Model May Come With a Price Tag
▼ Summary
– Meta is developing a new AI model, code-named Avocado, which it might charge for access to, according to a Bloomberg report.
– This potential move could mark a shift from Meta’s previous “open source” AI strategy, which CEO Mark Zuckerberg had called “the path forward.”
– The company’s previous Llama 4 AI model had a disappointing release, involving benchmark manipulation and a delayed “Behemoth” version that was later scrapped.
– In a recent memo, Zuckerberg indicated Meta may need to be more selective about open-sourcing AI to mitigate potential safety risks.
– Zuckerberg is now heavily involved with a new, siloed team called TBD Lab, working closely on these AI efforts.
The tech industry is closely watching a potential shift in Meta’s artificial intelligence strategy, as reports indicate the company’s next major AI model might not be freely available. Meta’s upcoming AI system, internally code-named Avocado, could represent a departure from the company’s previous commitment to open-source development by introducing a paid access model. This move would mark a significant change from CEO Mark Zuckerberg’s past statements championing open-source AI as the definitive path forward for the field.
According to recent reports, the company is actively developing this new model, which may come with a price tag for developers and businesses seeking to use it. This development follows the somewhat troubled release of Meta’s last major open-source model, Llama 4. That launch was marred by controversy, including allegations that the company manipulated standard AI performance benchmarks and the subsequent delay of a more powerful planned version known as “Behemoth.” Zuckerberg ultimately decided to abandon that particular development track in favor of pursuing a new direction.
The strategic pivot appears to be influenced by growing considerations around AI safety and commercial viability. In a late July memo discussing the concept of “personal superintelligence,” Zuckerberg hinted at a more cautious approach. He suggested that to properly address potential safety risks, Meta would need to be more selective about which technologies it releases openly to the public. This indicates a balancing act between fostering innovation through open collaboration and managing the reputational and practical dangers of powerful, freely available AI.
Internally, this shift has prompted a reorganization of resources and focus. Zuckerberg is now reportedly dedicating a substantial amount of his personal time and energy to a specialized team known as the TBD Lab, working directly with new high-profile hires in the AI space. This group operates in a dedicated, isolated workspace near the CEO’s office at Meta’s headquarters, underscoring the project’s priority and the desire for focused, secure development. The creation of this siloed environment suggests a concentrated effort to build advanced AI capabilities under a different framework than before, one that may prioritize controlled access over widespread, free distribution.
(Source: The Verge)
