Home / Technology / Rebel AI Group Raises Unprecedented Cash After Machine Learning Breakdown

Rebel AI Group Raises Unprecedented Cash After Machine Learning Breakdown

A fractured group of artificial intelligence researchers has set an unprecedented first round of financing for a generic AI-related startup, the latest in an enterprise-building effort to ensure the security of the most powerful technology. in that era

The group split from OpenAI, an organization founded backed by Elon Musk in 2015, to ensure smart AI systems won’t one day confuse and harm their manufacturers. The split followed a difference in the group’s direction after it took a $1 billion investment from Microsoft in 2019, according to two people familiar with the split.

The new Anthropic company, led by Dario Amodei, former OpenAI head of AI security, has raised $1

24 million in its first round, according to research firm PitchBook, the highest leverage for the AI ​​group trying to build general-purpose AI technologies. Rather than building to bring this technology to specific industries. According to figures disclosed in the company’s filing, Anthropic’s round value was $845 million.

The investment, led by Jaan Tallinn, the Estonian computer scientist behind Skype, former Google chief executive Eric Schmidt and Facebook co-founder Dustin Moskovitz also supported the investment.

The break from OpenAI started with Amodei’s departure in December, and has since moved on by nearly 14 researchers, according to estimates. They include Amodei’s sister, Daniela Amodei, president of Anthropic, as well as a group of researchers working on OpenAI’s groundbreaking language automation system GPT-3, including Jared Kaplan, Amanda Askell, Tom Henighan, Jack Clark and others. Sam McCandlish

OpenAI changed course two years ago when it needed support from Microsoft to support the growing hunger for computational resources to drive deep learning. commercial first

Oren Etzioni, head of the AI ​​institute founded by late Microsoft co-founder Paul Allen, said: “They started out as a non-profit. It aims to democratize AI.” “Of course, when you earn $1 billion. You have to generate returns I think their trajectory has become more organizational.”

OpenAI attempts to prevent AI security research from being commercialized later by limiting Microsoft’s presence on board. However, that has also led to internal tensions over organizational directions and priorities. According to a report from a person familiar with the sectarian group.

OpenAI wouldn’t comment on whether disagreeing with the research direction has caused divisiveness. But it said it had made internal changes to more closely integrate research and safety work when Amodei left.

Microsoft has been privileged to tap OpenAI research findings after giving $1 billion to support the group, mostly in the form of technology to support compute-intensive deep learning systems, including GPT-3. Earlier this week, Microsoft said it was embedding a language system in some software build tools so people without programming skills could build their own applications.

The rapid commercialization of GPT-3 is in stark contrast to OpenAI’s earlier handling of technology developed in 2019. The group initially said it would not divulge technical details about its development out of concerns about the “premium e-commerce”. Misuse of a powerful language system even if it was reversed later.

To protect itself from commercial interference, the Anthropic is registered as a public benefit organization with special arrangements to protect its mission of “Develop and maintain advanced AI responsibly for the benefit of humanity,” which includes the creation of a long-term benefit committee made up of individuals who are not affiliated with the company or its sponsors. And who will be the final speaker on the matter? including the composition of the board

Anthropic said its work will focus on “large scale AI models”, including making the system easier to interpret and “Create a method for more tightly integrated human opinions in the development and use of these systems.”

This article has been edited since it was first published to correct the number of researchers leaving OpenAI for Anthropic and Dario Amodei’s previous roles at OpenAI.

daily newsletter

© Financial Times

#techFT brings you news, opinions and analysis on big companies, technology and issues. that creates the sector’s fastest movement from experts from around the world. Click here to get #techFT in your inbox.

Source link