On the last day of June, a media outlet called TBPN tweeted a graphic of a highly skilled, highly paid fellow with his game face on. “TRADED” read the all-caps text underneath the image. Someone responded with a Photoshop edit of the man’s name (Jiahui Yu) and his former team’s logo (known as “the Blossom,” apparently) on a stock image of a burning LeBron James jersey, as one does. A few days later, TBPN called similar attention to “BREAKING NEWS” about a “FREE AGENT” of note. “Daniel Gross has entered the trade portal,” read that dispatch. “Rumors say he’s taking his talents to Meta.”
For many of the world’s most specialized performers, and also for the organizations angling to snag them, ’tis the season to try something new. In the NBA, for example, Kevin Durant took a quick break from throwing elbows online to move to the Houston Rockets via “the most complex trade in NBA history.” In the NHL, Toronto Maple Leafs star Mitch Marner decamped for the Vegas Golden Knights, though not without threats of litigation. (No need to Photoshop that burning jersey!) In the midst of all this sports news, though, the wide world of nerds—covered by outlets like TBPN, which stands, cheekily, for Technology Business Programming Network—has been where the real action is.
Yu and Gross are just two artificial intelligence experts who’ve been poached by the tech monolith Meta recently as part of a broader competitive escalation across the still-nascent sector. Call it the summer of free AI-gency! Or, if that’s too cringe, you can put it like The Wall Street Journal did at the end of last month: “With this spending bonanza and frenzy of movement, AI geeks are being treated more like NBA free agents.”
Back in late May, OpenAI, the ambitious operation behind ChatGPT that we at The Ringer last checked in on about a year and a half ago, proudly paid $6.5 billion to acquire a startup launched by one of Apple’s most famous former employees, designer Jony Ive. (The project, skeptically described by TechRadar as an AI-driven wearable akin to “an iPod Shuffle you can wear around your neck,” hit legal snags almost immediately but is now reportedly back on track.) In mid-June, Meta spent more than twice that amount—$14.3 billion, its largest outside investment ever—for a 49 percent share in a controversial data labeling company called Scale AI. As part of the deal, Meta founder Mark Zuckerberg installed Scale’s ascendent 28-year-old founder, Alexandr Wang, as the head of Meta’s new “Superintelligence Lab.” A few days later, OpenAI’s ruthless leader, Sam Altman, mentioned on his brother’s podcast that Meta had been going after his roster of workers, too.
According to Altman, his competitor was offering front-loaded, bonus-laden compensation packages worth “$100 million” in certain cases. “So far, none of our best people have decided to take them up on that,” Altman bragged on the pod. Even so, enough OpenAI staffers did make the leap that a company exec compared the experience to being the victim of a home invasion.
Earlier this week, OpenAI began striking back, plucking employees from Meta, Tesla, and xAI to help scale its business. For a price, it’s safe to assume.
Speaking about the AI talent pool last December, Naveen Rao, the vice president of AI at the company Databricks, said: “The top-tier researcher side is the hard part. It’s like looking for LeBron James.” Rao was certainly onto something with this choice of comparison.
“LMAO,” observed the influencer Andrew Pompliano in early July as the battle for niche, generational talent in the emerging tech sector heated up. “AI engineers are getting paid more than LeBron James, Shohei Ohtani, or Juan Soto. Revenge of the nerds.” An entrepreneur named Smit Patel, who describes himself as “Rick Rubin for startup founders” in his online profile, asked, First Take–ishly: “Is Zuck the GOAT of corp dev?” Another onlooker called Meta “basically the yankees or the dodgers,” warning that now “the world expects rings.” And Sarah Guo, founder of a venture capital firm called Conviction, wrote that this labor market land grab was creating secondary opportunities, too. “There are now folks helping researchers negotiate their comp packages and taking a fee,” Guo said. “Like agents for athletes.”
Even Altman sounded a bit like an everyday bro yappin’ about sportsball recently when he called out departing OpenAI employees for being driven by compensation rather than pure love of the game. “Missionaries will beat mercenaries,” Altman declared, and for a moment, I felt myself yearning for FanDuel lines on a mish-merx head-to-head. (None exist, but you can participate in wagers on Polymarket like: “Will Jony Ive join Meta before August?”)
It makes sense that professional sports would serve as a recurring motif when discussing the goings-on in AI at the moment. If you squint, the sector’s “race toward AGI”—short for artificial general intelligence, which refers to the attainment of true human-level-or-smarter capabilities—isn’t structured all too differently from the journeys teams make to jockey for a Super Bowl or a Formula One title.
So many colorful, obstinate personalities and idiosyncratic infrastructures to manage! Everyone always talking about chips! Such temperamental swings between swagger and doom! (James and Ohtani may not have personally threatened to destroy livelihoods/fully upend society just yet, but there’s still plenty of time.) The various companies operating on the frontier of AI research can feel like pro sports franchises sometimes, with their unique strengths and absurd failings and their staffers who have known (and/or despised) one another for the better part of a lifetime, AAU-ball style.
Apple, for example, is in its rebuild era: cheerfully terrible at making Siri, like, work, yet also terribly cheerful about the extent to which its business doesn’t have to revolve around AI. Google is as omniscient and omnipresent as ever, as Walt “Clyde” Frazier might say. Anthropic has a distinct defense-wins-championships devotion to matters like alignment and AI safety. Thinking Machines Lab is the thinking (wo)man’s choice. xAI, the troubled outfit behind Grok, is [drowned out by the sound of sirens and circus music].
Meta has endless resources, creepy users, an almost entirely unfettered executive, a mediocre model named Llama, and a history of making up for missing moments by buying better businesses, like Instagram or WhatsApp, outright. OpenAI has a confusing and constrained ownership structure, a lotta lore, and a slate of mostly well-regarded products. (I say “mostly” because one of them was once programmed to be so over-the-top obsequious that even people in tech found it off-putting.) In the realm of competition, there’s nothing like a burgeoning top-dog rivalry to make things interesting, and the growing public tension between Altman and Zuckerberg already feels like it could have legs.
In comparison to bigwigs in more buttoned-up industries like investment banking or Big Law, tech leaders have no qualms about doing and saying the kinds of things that most athletes might refer to as “bulletin board material.” And during a recent panel discussion alongside Altman, OpenAI COO Brad Lightcap was asked whether he thought Zuckerberg actually believed in Meta’s quest toward “superintelligence” or if he just says so to entice new workers. “I think he believes he’s superintelligent,” Lightcap said.
There are basically two ways to build teams in sports: You can acquire players through free agency or trades, or you can land fresh talent in the draft. This is true in business, too, as venture capitalist Garry Tan pointed out early this month. “VC firms often overpay for aging LeBrons with max deals at the end of their careers,” he wrote, “when a better strategy is to take as many first round draft picks as you can.” Tan compared the demo day at Y Combinator—the famed, flawed Bay Area startup accelerator that helped bring the world the likes of Airbnb and DoorDash and was, for a time, helmed by Altman—to “the NBA Draft Lottery.”
Just like on draft day, sometimes you hit, and sometimes you bust. Wang’s Scale AI, a unicorn tech company whose secret sauce is a global force of human click-workers paid to label data and solve CAPTCHAs, is a Y Combinator success story. But then there’s a place like Dynamo AI: one of several Y Combinator operations that had the ignominious honor of being on the résumé of a grindset grifter named Soham Parekh.
Parekh recently went viral not for making a hundred million dollars—but rather for somehow managing to get hired for roughly a hundred million different jobs. Parekh’s audacious work history, which he admitted to in an exclusive interview on none other than TBPN, included juggling three to four jobs at once; blaming drone strikes for unfinished assignments; and writing cover letters with lines like: “I am not athletic.” (Respect.) All of this misplaced hustle piqued the attention of one Mark Cuban, who cracked online: “Really excited that the Mavs signed Soham Parekh ! He was kind of under the radar, but I think he is going to do big things starting in Summer League !”
Cuban was promptly clowned for being days late to the game on that one. But back in the AI big leagues, where people are being paid astonishing sums to work one (1) job at a time, there’s a sense that a lot of the real funny stuff is just getting started. (Both funny-haha, and funny-yeahwe’reallgoingtoworkforthecomputersoneday.) In 2018, a New York Times article was written about the “borderline crazy,” “eye-popping salaries” commanded by AI specialists—salaries said to be in the range of $300,000 to $1.9 million. Just yesterday? Meta poached a top AI builder at Apple for a package worth a reported $200 million over a several-year period.
“Apple didn’t try to match the offer,” Bloomberg noted, “as it far exceeds pay at the company for leaders other than Chief Executive Officer Tim Cook.”
Last week, Cook was among the select few invited to Allen & Co.’s annual CEO and power broker mogulfest in Sun Valley, Idaho (a conference that was a big part of the inspiration for the “Argestes” episode of Succession). Altman and Zuckerberg were attendees, too. Over the Fourth of July weekend, Zuck posted a video of himself wakeboarding in an inflatable bald eagle suit, unruffled and unbothered. (We all get the Banana Boat we deserve.) But Altman seemed slightly less relaxed last Tuesday, as he was being interviewed by a scrum of reporters in front of his Sun Valley hotel like an NFL player during training camp. Wearing a conspicuous pair of Vuarnet sunglasses, Altman said he hadn’t seen Zuck yet but looked forward to it. And with that, the other summer league commenced.
