OpenAI Sounds the Alarm

This week, the world’s most important artificial intelligence company was closed. OpenAI gave its entire staff a week off to “recharge,” a seemingly generous perk for a workforce relentlessly pushing toward building a world-changing technology.

But this was not a wellness initiative. It was a strategic retreat in the middle of a brutal, high-stakes war for talent that is now threatening to shatter the company’s carefully crafted identity.

The enemy is Meta Platforms, the social media empire that includes Facebook, WhatsApp, and Instagram. According to OpenAI’s own CEO, Sam Altman, their tactics are getting ugly. In a recent Slack message to employees reviewed by WIRED, Altman addressed the departure of several key researchers poached by Mark Zuckerberg’s company.

“Meta is acting in a way that feels somewhat distasteful,” Altman wrote, acknowledging the “giant offers to a lot of people on our team.” He framed the current moment as a predictable, if chaotic, new phase. “We have gone from some nerds in the corner to the most interesting people in the tech industry (at least),” he wrote. “AI Twitter is toxic,” he continued, adding: “I assume things will get even crazier in the future. After I got fired and came back I said that was not the craziest thing that would happen in OpenAI history; certainly neither is this.”

The message highlights the rising tension in the war for AI talent. But it also reveals something deeper: OpenAI, the most prominent lab in the generative AI race, may be struggling to keep its own people on board. For years, OpenAI has operated with the fervor of a quasi-religious mission. The goal was not just to build products; it was to birth Artificial General Intelligence (AGI) for the benefit of humanity. The work was hard, the hours long, but the mission itself was presented as the ultimate compensation. Now, Zuckerberg is calling that bluff, making a cynical bet that every missionary has a price, and it seems he’s being proven right.

The conflict has become so intense that it’s now creating collateral damage among OpenAI’s closest allies. In a stunningly ironic twist, Ilya Sutskever, the OpenAI co-founder who left to start his own safety-focused AI lab, is a direct victim. This week, he announced that Daniel Gross, the CEO of his company, Safe Superintelligence (SSI), has left. Gross is joining Meta.

“As you know, Daniel Gross’s time with us has been winding down,” Sutskever posted on X (formerly Twitter) on July 3. ”And as of June 29 he is officially no longer a part of SSI. We are grateful for his early contributions to the company and wish him well in his next endeavor. I am now formally CEO of SSI, and Daniel Levy is President. The technical team continues to report to me.”

Sutskever also confirmed reports that Meta had approached Safe Superintelligence for a potential acquisition. “You might have heard rumors of companies looking to acquire us. We are flattered by their attention but are focused on seeing our work through,” he wrote, adding, “We have the compute, we have the team, and we know what to do. Together we will keep building safe superintelligence.”

I sent the following message to our team and investors:
—

As you know, Daniel Gross’s time with us has been winding down, and as of June 29 he is officially no longer a part of SSI. We are grateful for his early contributions to the company and wish him well in his next…

— Ilya Sutskever (@ilyasut) July 3, 2025

This is the backdrop for Altman’s memo, in which he attempts to rally his own troops by taking the moral high ground. He dismissed Meta’s recruiting success, claiming they “didn’t get their top people and had to go quite far down their list,” and that he had “lost track of how many people from here they’ve tried to get to be their Chief Scientist.”

He framed the conflict as a battle of ideals. “I am proud of how mission-oriented our industry is as a whole; of course there will always be some mercenaries,” he wrote, before declaring, “Missionaries will beat mercenaries.”

But in the same message, he quietly conceded that the mission may no longer be enough. He noted that OpenAI is assessing compensation for the entire research organization, promising to do it “fairly and not just for people who Meta happened to target.” It’s a stunning admission. To stop the bleeding, Altman is being forced to play Meta’s game.

He then made his final pitch, arguing OpenAI is the only place truly dedicated to the cause. “We actually care about building AGI in a good way,” he added. “Other companies care more about this as an instrumental goal to some other mission. But this is our top thing, and always will be. Long after Meta has moved on to their next flavor of the week…we will be here, day after day, year after year.”

Viewed through this lens, the mandatory vacation looks less like a perk and more like a desperate defensive maneuver. It’s an attempt to stanch the bleeding, to get employees away from their workstations and the constant ping of recruiters, and to prevent a full-blown crisis of confidence.

Our Take

OpenAI is still the face of generative AI. It has the most famous chatbot, the biggest media profile, and the deepest partnership with Microsoft. But its grip on elite talent is slipping.

Meta, meanwhile, has money, momentum, and a ruthlessness it no longer feels the need to hide. Zuckerberg is not just building an AI lab. He’s building a recruiting machine designed to buy the best army.

As for Safe Superintelligence, it now becomes the third node in an increasingly fractured landscape, an independent alternative to the titans, run by one of OpenAI’s original architects.

Altman may still believe that “missionaries will beat mercenaries.” But missions don’t retain people when nine-figure offers are on the table. Culture does. And this week, the cracks in that culture are starting to show.