report. What it is trying instead.
The Stock Platforms Stock Platforms climbed this year in response to the progress of the social media company with artificial intelligence. Now, it increases its efforts, trying to gain a major AI start and hire new AI executives.
Meta tried to buy a secure super -superingence earlier this year, but was rejected by its founder Ilya Sutskever, referring to people who were familiar with the issue late on Thursday, CNBC. In March, a financing tour was a safe super -Superstegence value was $ 30 billion.
Meta and safe Superintelligigence did not respond immediately to comments.
In the face of this, such a purchase would be a strange movement. Safe Superintalligigence did not publish any product because it focused on developing Supersmart. Meta also gets along Perfect good On its own, with 19% stock so far this year.
The real charm of such an agreement would probably be Sutskever and his key employees on the ship.
Sutskever was the chief scientist in Openai and helped improve the technology behind Chatgpt. Last year, he left Openai after a break with his CEO Sam Altman and then launched a safe super superity.
According to CNBC, Mark Zuckerberg, Mark Zuckerberg, CEO, who was blocked by efforts to bring Sutskever to the ship, negotiated to hire the safe super Superintelligige CEO Daniel Gross and former Github CEO Nat Friedman. Gross and Friedman are partners at the NFDG Investment Fund, which supports several AI Start-up.
So far, Meta has trusted an anthropically anthropic with Microsoft’s Openai and Amazon.com, rather than a starting or financing in -house AI models. However, there are signs that Zuckerberg thinks that Meta’s AI team thinks should be supported.
Last week, Meta completed its scale AI investment. Wall Street Journal reported This commodity would pump $ 14 billion to the data labeling company by 49% shares, and this scale would join the AI founder Alexandr Wang Meta.
The larger picture here is that multiple AI companies delayed their versions of the next flagship models because of their concerns that they do not have adequate improvement. This proposes the idea that the “scaling law ın of the industry is automatically that larger and more complex models are automatically. more intelligentcollapsing.
Meta is among those who struggle to make a breakthrough. According to Journal, the “Behemoth” model, which was initially released in April, is postponed until autumn or subsequently.
The response of artificial intelligence companies has been the development of reasoning models that broke the problems step by step. However, a recent article from the researchers in Apple found “basic limitations ında in these models. According to the researchers, these AIs experienced “full accuracy collapse ında in tasks beyond a certain level of complexity.
This shows that the industry must adopt new techniques to push AI to the next level of intelligence. Meta will hope that their new recruitment can go there first.
