In the intense competition between Meta and OpenAI for AI supremacy, the preferred tool has shifted. Initially , it was unlimited compute, subsequently, $100m signing bonuses. Now, the conflict has progressed to a more peculiar and personal stage: the competition over soups.
TL;DR
- Meta and OpenAI's AI talent war now involves CEOs personally delivering soup to recruits.
- Mark Zuckerberg hand-delivers soup; OpenAI's Mark Chen reciprocates by outsourcing Michelin-star soup.
- The competition highlights the scarcity of advanced AI scientists, fewer than 1,000 globally.
- Personal gestures replace financial incentives, signaling a researcher's high value to CEOs.
During a podcast appearance with tech commentator Ashlee Vance, Mark Chen, who serves as OpenAI's Chief Research Officer, stated that the competition for talent has changed. Chen indicated that Meta has been actively recruiting approximately half of his immediate team, supported by a substantial $10 billion budget for acquiring skilled individuals. However, he also noted that CEO Mark Zuckerberg has personally been involved in these recruitment efforts.
Zuckerberg, Chen said, has personally “hand-cooked” and “hand-delivered” soup to researchers he wanted to recruit away from OpenAI. And it wasn’t a joke, the executive insisted.
“It was shocking to me at the time,” Chen admitted. But in Silicon Valley, if the enemy brings broth, you must respond in kind. Chen confessed he has now adopted the tactic, delivering soup to his own recruits as he hopes to poach talent from Meta. However, he draws the line at manual labor.
“No, no, no … it’s better if you get like Michelin-star soup,” Chen laughed, noting he outsources the job to a high-end Korean-soup spot in the Valley called Daeho.
“These things can be effective in their own way,” he said. Chen is even planning a cooking class offsite to lean into the absurdity.
The pleasant showmanship conceals a more difficult truth: the number of individuals capable of creating and educating advanced large language models is extremely small. Those familiar with the industry estimate estimate that fewer than 1,000 scientists worldwide possess the necessary skills to advance the field independently.
This situation bears a resemblance to past conflicts for tech expertise, when Google and Facebook attempted to outbid each other by offering complimentary sushi, onsite coffee makers, and fitness centers within their facilities. However, the magnitude of this current situation is distinct: financial exits enabling researchers to liquidate their stakes sooner, preferential access to computational resources, and the prospect of shaping the development of influential AI technologies that will define the future.
To attract those who are difficult to recruit, chief executives must replace financial incentives or other assets with personal connections. A CEO arriving at your residence with a meal conveys a significant point: your importance is such that I'm willing to invest my personal time in winning you over.
Chen, in his perspective, is employing the narrative to convey a wider observation regarding the internal experience of the competition for talent at OpenAI. According to Chen, media reports have frequently depicted Meta as merely poaching OpenAI's top employees.
He countered, stating Meta “went after a lot of people quite unsuccessfully,” among them, noting that all of his direct reports had declined the offer.
He proposed that OpenAI's approach to keeping staff focuses less on competing with Meta financially and more on fostering belief: scientists remaining due to their confidence in the lab's trajectory and its prospects for achieving artificial general intelligence ahead of others.
“Even among people who have offers from Meta,” Chen said, “I haven’t heard anyone say AGI is going to be developed at Meta first.”
