**Meta Platforms Intensifies AI Efforts Amid Stock Surge**
Meta Platforms’ stock has seen significant growth this year, largely due to the company’s advancements in artificial intelligence (AI). Recently, reports have emerged that Meta is intensifying its AI initiatives, including attempts to acquire a prominent AI start-up and hiring new AI executives. Earlier this year, Meta sought to purchase Safe Superintelligence but was turned down by its founder, Ilya Sutskever, as reported by CNBC. Safe Superintelligence was valued at $30 billion during a funding round in March. Neither Meta nor Safe Superintelligence has commented on the situation.
At first glance, acquiring Safe Superintelligence might seem unusual, as the company has yet to release any products, focusing instead on developing advanced AI technologies. Nevertheless, Meta has performed well independently, with its stock rising 19% this year. The primary motivation for such an acquisition would likely be to bring Sutskever and his key team members into Meta. Sutskever, a former chief scientist at OpenAI, played a crucial role in developing the technology behind ChatGPT before leaving OpenAI last year after a disagreement with CEO Sam Altman.
In light of the unsuccessful acquisition attempt, Meta CEO Mark Zuckerberg is reportedly negotiating to recruit Safe Superintelligence’s CEO, Daniel Gross, along with former GitHub CEO Nat Friedman. Gross and Friedman are partners in the investment fund NFDG, which has supported several AI start-ups. Unlike Microsoft and Amazon, which have invested in AI start-ups like OpenAI and Anthropic, respectively, Meta has primarily relied on its in-house AI models. However, there are indications that Zuckerberg believes Meta’s AI team requires strengthening.
Recently, Meta completed a significant investment in Scale AI, committing $14 billion for a 49% stake in the data-labeling company, with Scale AI founder Alexandr Wang set to join Meta. The broader context reveals that many AI companies have postponed the launch of their next flagship models due to concerns about insufficient improvements. This situation suggests that the industry’s “scaling law,” which posits that larger and more complex models are inherently more intelligent, may be faltering. Meta is among those facing challenges in achieving breakthroughs, with its “Behemoth” model, initially scheduled for an April release, now delayed until fall or later.
In response to these challenges, AI companies are developing reasoning models that tackle problems step-by-step. However, a recent study from Apple researchers identified “fundamental limitations” in these models, noting that they experience “complete accuracy collapse” when faced with tasks beyond a certain complexity. This finding indicates that the industry may need to explore new techniques to advance AI capabilities.
**FAQ**
**Q: Why is Meta Platforms focusing on AI acquisitions?**
A: Meta is enhancing its AI capabilities to remain competitive in the tech industry, aiming to attract top talent and innovative technologies to bolster its in-house AI development.
