Meta Platforms, Inc. has restricted its engineers from using Anthropic’s Claude and OpenAI’s Codex in light of concerns that AI outputs could be inadvertently incorporated into the company’s training data through a process known as distillation. According to reports by The Information, Meta has directed teams to pause certain tasks involving third party AI models.
The restriction stems from the potential risk of rival AI outputs contaminating Meta’s own AI training data, which could lead to escalations with partner companies. This move is not solely motivated by competitive concerns but also aims to protect strategic partnerships and ensure the integrity of Meta’s proprietary technologies.
Additionally, Google has reportedly placed limits on Meta’s use of Gemini AI models, despite granting access to more computing capacity. In March, Google warned that it could not provide all the Gemini capacity required by Meta, signaling a cautious approach towards sharing cutting edge technology with its rivals.
Meta Platforms invests heavily in AI infrastructure and operates a multimodal assistant integrated across platforms like WhatsApp, Instagram, Messenger, and Facebook. It also develops and open-sources the Llama family of large language models, fostering collaboration among developers and organizations.
While Meta remains an important player in the AI landscape, some investors have identified other undervalued AI stocks that offer greater upside potential and lower risk. For instance, certain AI companies are benefiting from Trump-era tariffs and the onshoring trend, making them attractive investment opportunities for those looking to capitalize on current market dynamics.
Source: https://finance.yahoo.com/technology/ai/articles/why-meta-platforms-inc-meta-224947515.html
Thinking about building an AI product?
Get in Touch