MIT Technology Review

We're on those other social media platforms, too. 🟦☁️: bsky.app/profile/techno… 🧵: @technologyreview" target="_blank" rel="nofollow noopener">threads.net/@technologyrev…

According to Greg Brockman, Elon Musk asked a table of AI researchers at a dinner in Menlo Park in 2015, "Is Demis Hassabis evil?" Demis Hassabis was (and still is) CEO of Google DeepMind at the time. Elon testified last week that he co-founded OpenAI to create a "counterbalance to Google" because Google cofounder Larry Page didn't seem concerned about AI safety.

In October 2015, when Greg Brockman was soliciting donations to found OpenAI as a nonprofit, he sent an email to Marissa Mayer, who was the CEO of Yahoo at the time, asking for a donation. He told her that he himself was donating $100,000 to OpenAI, and made the same claim to others, including Sam Altman. Brockman says he's "not certain" whether he made the claim to Elon Musk as well. Brockman never made the donation.

OpenAI's lawyer points out to Stuart Russell that Elon tweeted that "Tesla will be one of the companies to make AGI" and has said that Optimus, a humanoid robot designed by Tesla, is the company's biggest product of all time. He's suggesting that Elon is suing OpenAI to undermine a competitor in the AI race. x.com/elonmusk/statu…

Here’s what we learned from Elon Musk’s testimony, and what to expect this week. trib.al/0RvPleq

"Each [AI] company individually feels that it needs to be in this race for improving capabilities, and that means they can't stop and solve the safety problem," says Stuart Russell, a computer scientist at UC Berkeley. Some individual employees might want to do this, but "the overall company policy is preventing that," he says.

Stuart Russell, a computer scientist at UC Berkeley, takes the stand to testify, direct-examine by Elon Musk's lawyer Steven Molo. He says that the current AI race has a "winner take all" dynamic, in which "whichever company develops AGI first (AI that matches or exceeds human capabilities in every area) will have a significant advantage." Because AGI could take over most human work, the companies that create it will "quickly come to control perhaps the majority of economic activity on the planet." Even governments could "become subordinate to these companies."

@techreview Stuart Russell testifies during cross-examination that he's getting paid $235,000 by Elon Musk to testify about AI safety in this trial. He says that will make up 20% of his annual income.

Stuart Russell, a computer scientist at UC Berkeley, takes the stand to testify, direct-examine by Elon Musk's lawyer Steven Molo. He says that the current AI race has a "winner take all" dynamic, in which "whichever company develops AGI first (AI that matches or exceeds human capabilities in every area) will have a significant advantage." Because AGI could take over most human work, the companies that create it will "quickly come to control perhaps the majority of economic activity on the planet." Even governments could "become subordinate to these companies."

Day 5 of Musk v. Altman trial in federal court in Oakland, California. Stuart Russell, a computer scientist at UC Berkeley, will testify about AI safety. OpenAI president Greg Brockman will also testify. I'm a journalist (and lawyer) reporting on the trial for @techreview.

In the first week of the Musk v. Altman trial, Elon Musk said he was duped by OpenAI CEO Sam Altman and president Greg Brockman, warned AI could kill us all, and admitted that xAI distills OpenAI's models. Elon kept his cool, and OpenAI’s lawyer bulldozed him with piercing questions about his motivations for suing the company. technologyreview.com/2026/05/01/113…

Musk kept his cool, and OpenAI’s lawyer bulldozed him with piercing questions about his motivations for suing the company. trib.al/jti2ZF6

Plus: Elon Musk has admitted that xAI trained Grok on OpenAI models. trib.al/2G0MCeL

A company called Orpheus Ocean wants to go “deep for cheap.” trib.al/767oAFA

Ambitious research is on the chopping block following yet more cuts at the National Science Foundation. trib.al/8MQBJ94

Access a subscriber-only eBook on a startling and fairly graphic pursuit of human longevity that poses concerns about the ethics of cloning. trib.al/K5RwvcE

"That's not a leading question, that's a leading answer," says Musk to OpenAI's lawyer during cross-examination. "You're not a lawyer," says Judge Gonzalez.

Microsoft asks Elon why he didn't sue Microsoft when he found out that Microsoft invested billions and got an exclusive license to GPT3 in 2020. "[Sam Altman] assured me that they were staying on the mission" (of being an open-source nonprofit), he says. Musk formed xAI in 2023 and sued Microsoft in 2024. Microsoft is suggesting that Musk sued the company too late (barred by the statute of limitations) and is suing now only to bring down a competitor.

OpenAI’s lawyer asks Musk if he knows that OpenAI foundation is the same as the nonprofit that he cofounded. “It’s not the same company because most of the value has been taken by the for-profit," says Musk.

Musk argues he was deceived into bankrolling OpenAI by being promised that the company would remain an open-source nonprofit, but he did write down the terms of his donation. In August 2016, Sam Altman emailed Musk saying OpenAI negotiated a $50 million compute donation from Microsoft. "I would also like to see the exact terms and conditions. Gifts are only as good as the T&C (terms and conditions)," wrote Musk.

Goodfire wants to make training AI models more like good old-fashioned software engineering. trib.al/C1HY60h