Throughout a company-wide all-hands assembly on Thursday, a few of Meta’s high executives had been requested concerning the “$100 million signing bonuses” that OpenAI CEO Sam Altman claimed that they had been providing to poach his staff.
“Sam is simply being dishonest right here,” Andrew Bosworth, Meta’s CTO, stated on the assembly when requested about Altman’s remarks. “He’s suggesting that we’re doing this for each single individual… Look, you guys, the market’s sizzling. It’s not that sizzling.”
The “$100 million bonus” headline has rightfully develop into a meme on social media since Altman stated the quantity on his brother’s podcast. “What Sam neglects to say is that he’s countering all these gives, making a small marketplace for a really, very small variety of people who find themselves for senior, senior management roles” within the new superintelligence AI team Meta is building, Bosworth informed Meta staff right now. “That isn’t the overall factor that’s taking place within the AI house. And naturally, he’s not mentioning what the precise phrases of the supply are. It’s not [a] sign-on bonus. It’s all these various things.”
Bosworth then referenced latest tales a few handful of OpenAI researchers who’re becoming a member of Meta and stated there are “fairly a number of extra within the pipeline that I can’t announce or share proper now.”
“Sam is thought to magnify, and on this case, I do know precisely why he’s doing it, which is as a result of we’re succeeding at getting expertise from OpenAI,” he stated. “He’s not very completely happy about that.”
On the Thursday assembly, there have been many staff current from the corporate’s engineering “bootcamp,” a multi-week onboarding program that assigns new hires to numerous groups. “For all the brand new bootcampers right here, you didn’t screw up,” Bosworth stated to laughs and claps from the viewers. “You made an awesome determination. Comp is correct the place it must be.”
Bosworth wasn’t the one Meta exec to say OpenAI through the inner assembly. CPO Chris Cox additionally acknowledged that, whereas Meta AI has one billion month-to-month customers, engagement “will not be almost as deep as the way in which that individuals are utilizing ChatGPT.” The standalone Meta AI app has solely 450,000 day by day customers, he informed staff, and “lots of these people” are utilizing it to handle their Ray-Ban Meta glasses.
“We’re not going to go proper after ChatGPT and attempt to do a greater job with serving to you write your emails at work,” Cox stated. “We have to differentiate right here by not focusing obsessively on productiveness, which is what you see Anthropic and OpenAI and Google doing. We’re going to go deal with leisure, on reference to pals, on how individuals reside their lives, on all the issues that we uniquely do nicely, which is an enormous a part of the technique going ahead.”
Meta declined to touch upon the interior assembly.
After I spoke with Jason Rugolo on Thursday, I needed to know why he’s suing probably the most influential firm in tech.
Rugolo’s AI system startup, Iyo, not too long ago won a temporary restraining order that bars OpenAI from utilizing the “io” model for Sam Altman’s new {hardware} division with Jony Ive. In response, Altman took to his X account to recommend that Rugolo filed his trademark lawsuit as a result of OpenAI refused to put money into or purchase Iyo, which is gearing as much as launch its first AI-powered, in-ear headphones later this yr.
Rugolo acknowledges (and paperwork submitted to the court docket affirm) that he pitched Altman on investing a number of instances. He additionally mentioned an acquisition with io group members this yr. Nonetheless, he says his lawsuit isn’t a part of some revenge campaign, however quite meant to remove any confusion between his forthcoming Iyo One headphones and Altman’s io.
Trademark lawsuits are a dime a dozen, however this one has damaged by means of for good purpose. There’s intense curiosity in what Altman and Ive are constructing (the primary system apparently won’t be an “in-ear” product or a “wearable”), and the case is a Rorschach check for the way you’re feeling about Altman, who’s undoubtedly polarizing.
“I had a large change in opinion on the man,“ Rugolo tells me of Altman. “Whereas I used to be assembly with them, I used to be below the spell of Sam Altman being an awesome entrepreneur and a very attention-grabbing individual. That broke fairly immediately after their public announcement [of io].”
“Am I getting screwed right here?” Rugolo remembers considering. “After I talked to him on the cellphone and he made a Sopranos risk to sue me, I used to be similar to, ‘Alright, this man is a nasty dude.’” Now, he says that Altman is making an attempt to “manipulate the arguments within the public sphere” and “make me appear like a cash grubber or a sore loser, and I simply don’t assume it’s gonna work.”
“It is a baseless trademark dispute and never a case about stolen concepts or expertise,” OpenAI spokesperson Kayla Wooden says in an announcement shared with me. “Iyo demoed a product in Could 2025 that didn’t operate correctly or meet our requirements in hopes that we’d purchase Iyo. We handed. Jason Rugolo was additionally nicely conscious of the io identify and by no means raised issues earlier than our announcement.”
Due to the hundreds of thousands of {dollars} he not too long ago raised from his producer, Pegatron, and a billionaire whom he refuses to call, Rugolo says Iyo has sufficient runway to final it by means of the top of 2026. After I ask if the system he teased in his viral TED talk final yr is certainly transport later this yr, he says he’s about to fly to China to “mainly be residing on the manufacturing unit.”
Whereas he’s able to undergo the authorized discovery course of and take his case to trial, he hopes that OpenAI will “put their weapons away” and “full like grown-ups on product.”
“I’ll meet them out there,” he tells me. “We are going to each attempt to launch stuff that’s actually cool and see if we will serve our clients. They’ll simply compete pretty and cease utilizing the identify. They’ve among the greatest designers on the earth, apparently. Consider a brand new identify. You simply can’t use the one which I informed you about already, and that I’ve been utilizing since 2019.”
To this point, Runway is thought for bringing generative AI to Hollywood. Now, the $3 billion startup is setting its sights on the gaming trade.
This week, I used to be granted entry to a brand new interactive gaming expertise that Runway plans to make out there to everybody as quickly as subsequent week, in response to CEO Cristóbal Valenzuela. The patron-facing product is presently fairly barebones, with a chat interface that helps solely textual content and picture technology, however Valenzuela says that generated video video games are coming later this yr. He says that Runway can be in talks with gaming firms about each utilizing its expertise and accessing their datasets for coaching.
Primarily based on his latest conversations, Valenzuela believes the gaming trade is in an analogous place to Hollywood when it was first launched to generative AI. There was appreciable resistance, however over time, AI has been progressively adopted in additional areas of the manufacturing course of. Valenzuela says Amazon’s latest present, Home of David, was made partly with Runway’s expertise, and that his firm is working with “just about each main studio” and “a lot of the Fortune 100 firms.”
“If we might help a studio make a film 40 % quicker, then we’re in all probability gonna be capable of assist builders of video games make video games quicker,” he says. “They’re waking up, they usually’re transferring quicker than I’d say the studios had been transferring two years in the past.”
Naturally, I couldn’t let Valenzuela get off our Zoom name with out asking him about his recent acquisition talks with Zuckerberg: “I believe we’ve got extra attention-grabbing mental challenges being impartial, and remaining impartial for now.”
Nobody is aware of what AGI really means. That a lot is evident from this glorious deep dive from The Information into Microsoft’s take care of OpenAI. There was lots of good reporting on the negotiations between the 2 firms, however this piece is probably the most complete and detailed I’ve seen but. It states that Microsoft will now not obtain unique entry to OpenAI’s IP if it achieves “adequate AGI,” which is contractually outlined as when OpenAI’s board determines that the AI “has the potential to generate” the utmost earnings its buyers are entitled to. Amazingly, OpenAI doesn’t have to truly generate these earnings.
Two under-the-radar offers: Though they haven’t garnered many headlines, OpenAI introduced an attention-grabbing partnership and a small acquisition this week. The primary is a deal with Applied Intuition to “advance next-generation, AI-powered experiences in autos.” The second is the acquisition of the small group at Crossing Minds, an AI startup that helped e-commerce firms supply extra personalised product suggestions. “Personally, becoming a member of OpenAI’s analysis group to deal with brokers and data retrieval is a singular honor,” Crossing Minds founder Alexandre Eobicque writes. “These are exactly the issues I’ve all the time been obsessed with: how programs study, purpose, and retrieve information at scale, in real-time.”
Some attention-grabbing profession strikes in tech:
- The three founders of OpenAI’s analysis workplace in Zurich, Lucas Beyer, Alexander Kolesnikov, and Xiaohua Zhai, confirmed they’re becoming a member of Meta. (“No, we didn’t get 100M sign-on, that’s faux information,” writes Beyer.) Trapit Bansal, a former OpenAI researcher who “began the RL for reasoning effort with Ilya Sutskever” and co-created the o1 mannequin, is also joining Meta’s new lab.
- Elon Musk is cleansing home at Tesla. He reportedly fired Omead Afshar, his longtime fixer and head of producing. HR chief Jenna Ferrua is also out.
- Nate Mitchell has joined his fellow Oculus co-founder, Brendan Iribe, on the AI glasses startup Seasame, the place he’ll be chief product officer.
- Databricks hired Alan Davidson, the previous Assistant Secretary of Commerce for the NTIA, to be its head of presidency affairs.
- Nameless tech employees describe how AI has killed their jobs.
- Anthropic revealed research on how people are using Claude for emotional support. I’d like to see this sort of analysis from OpenAI.
- OpenAI’s Sam Altman and Brad Lightcap went on the live Hard Fork podcast. The start of this convo was fairly enjoyable — and a bit awkward — to observe from the viewers.
- The state of consumer AI from Menlo Ventures.
- “Inside Silicon Valley’s anti-college movement.”
- How Venice is bracing for Jeff Bezos and Lauren Sanchez’s marriage ceremony this weekend.
If you happen to haven’t already, don’t forget to subscribe to The Verge, which incorporates unlimited access to Command Line and all of our reporting.
As all the time, I welcome your suggestions. You’ll be able to reply right here or ping me securely on Signal.