Meta Platforms is navigating a pivotal period in its artificial intelligence development. Recently, the company's Chief Technology Officer, Andrew Bosworth, candidly expressed his dissatisfaction with the previously released Llama 4 AI software, deeming it a "disappointment" due to its perceived lack of specific strengths. However, Bosworth quickly shifted focus to a forthcoming internal AI model, which he enthusiastically described as "looking really good," hinting at significant advancements. This strategic pivot in AI development coincides with a restructuring of Meta's AI division, now under the leadership of Alexandr Wang. Adding to the organizational shifts, Meta's Reality Labs division is implementing workforce reductions, adjusting its focus as virtual reality headset sales encounter headwinds, while its Horizon Worlds platform sees encouraging growth on mobile devices.
Meta's AI Journey: From Llama 4's Challenges to Future Prospects
In a recent interview at the World Economic Forum in Davos, Switzerland, Meta Platforms Inc. CTO Andrew Bosworth offered a candid assessment of the company's open-source AI software, Llama 4, which was launched last year. Bosworth openly admitted that Llama 4 had proven to be a "disappointment," citing its lack of a clear specialty or outstanding performance in any particular area. He remarked that the model "wasn't amazing."
However, Bosworth's outlook brightened when discussing Meta's next-generation AI model. This new iteration, currently being tested internally by employees, has received high praise from the CTO, who declared it "really good." This promising new AI endeavor marks the first significant development since Meta undertook a major reorganization of its AI division, placing it under the direction of Alexandr Wang.
Reports suggest that this innovative AI model could be publicly released in the first half of the current year, although internal deliberations are ongoing regarding the extent and nature of its open-sourcing. Concurrently, Bosworth also disclosed that Meta's Reality Labs division has implemented a 10% reduction in its workforce. This decision comes as sales of VR headsets have been slower than anticipated, prompting the company to shift its attention within its Horizon Worlds platform towards mobile applications, where it has observed stronger user engagement.
The leadership of Meta's AI initiatives has not been without its own internal dynamics. Alexandr Wang, the founder of Scale AI, joined Meta in 2025 following the company's substantial investment of over $14 billion for a 49% stake in his startup. Recent reports indicate that Wang has expressed feeling "suffocated" by the intense scrutiny and direct involvement of Meta CEO Mark Zuckerberg in the AI project, suggesting that such oversight may be impeding progress. Conversely, Yann LeCun, a former Meta scientist and pioneering figure in AI, has voiced concerns about Wang's youth and perceived inexperience, implying that the 28-year-old co-founder may not yet fully comprehend the intricate workings of advanced AI research teams. As these internal developments unfold, the financial community awaits Meta's fourth-quarter earnings report, scheduled for Wednesday, January 28, after market close. Industry analyses suggest that Meta currently exhibits a robust short-term price trend but lags in medium and long-term performance, holding a weak momentum ranking according to Benzinga's Edge Stock Rankings.
The candid assessment of Llama 4's performance and the enthusiastic anticipation for Meta's upcoming AI model highlight the fast-paced and competitive nature of the artificial intelligence landscape. It underscores the challenges even tech giants face in consistently delivering groundbreaking innovations. The internal discussions surrounding leadership, strategic direction, and open-sourcing policies reflect the complex decisions at the heart of technological advancement. Ultimately, Meta's journey with AI serves as a powerful reminder that progress often involves acknowledging shortcomings, adapting strategies, and continually pushing the boundaries of what's possible, even amidst internal friction and market pressures.