Sam Altman recently revealed that Meta has poached at least nine employees from OpenAI. This recruitment strategy for its new superintelligence division has drawn criticism as "somewhat unpalatable." Reports indicate Mark Zuckerberg is offering $100 million signing bonuses to attract engineers from Altman's organization and other tech firms.
After Zuckerberg disclosed the personnel lineup for the Meta Superintelligence Lab researching both Llama models and cutting-edge AI systems potentially surpassing human capabilities, OpenAI's CEO commented on the matter in internal Slack communications reviewed by Wired.
Altman acknowledged "several exceptional individuals" have joined Meta but noted Facebook's parent company failed to secure top-tier talent, having to "go far down the OpenAI candidate list." This aligns with his previous podcast remarks where he stated "not a single (OpenAI) top talent" accepted Meta's offers.
"They've been trying to recruit people for a long time, I can't even count how many times they've tried to lure someone to be their chief scientist," Altman wrote in Slack.
The person ultimately appointed to lead the superintelligence lab is 28-year-old Scale AI co-founder Alexandr Wang. Meta invested $14.3 billion to acquire 49% of Wang's data annotation startup, securing his position as part of the deal.
"I'm proud of our industry's mission-driven approach; certainly there will always be mercenaries," Altman added. "Missionaries will outlast mercenaries."
While OpenAI maintains its commitment to ensuring AI benefits humanity, recent decisions like shifting to a profit-driven business model and partnering with the US Department of Defense suggest similar corporate ambitions seen at Meta.
When Meta uses money to entice talent, it's mercenary behavior; when OpenAI does it, it's a mission
Though Altman dismissed Meta's financial incentives as creating "deep cultural issues," his memo didn't entirely reject similar strategies. He mentioned OpenAI is reassessing organizational compensation, a point previously raised by lead scientist Mark Chen.
Altman also claimed "OpenAI stock has more upside potential than Meta's stock" due to his company's exclusive AI focus. "While Meta moves to their next big thing or defends their social media moat, we'll be here day after day, year after year, striving to be better than anyone else," he wrote.
However, OpenAI will "fairly" increase compensation across the board rather than targeting specific candidates, according to the CEO.
OpenAI vs. Meta: The Superintelligence Race
Altman's suggestion that Meta lacks his company's AI focus has merit. After all, Zuckerberg's company began as Facebook and maintains dominant social media control. Its 2021 name change to Meta solidified virtual reality commitments with ongoing investments in that space.
While pursuing AI since around 2013, Meta struggled post-ChatGPT to match OpenAI and Google's progress. Strategic missteps included open-sourcing models that allowed DeepSeek to create more advanced, cost-effective derivatives.
Their latest Llama models performed poorly, their AI chatbots gained negative reputations, and top engineers have been leaving the company.
In the memo outlining who joined the Meta Superintelligence Lab, Zuckerberg asserted his company has "unique capabilities" to deliver superintelligence through massive computational resources, proven product delivery to billions, AI hardware expertise like smart glasses, and an organizational structure enabling bold initiatives.
Meanwhile, OpenAI has focused exclusively on advancing general AI since 2015. In January, they announced superintelligence as their primary 2024 goal. Last month, Altman revealed his company "recently built systems surpassing human capabilities in many aspects." As the superintelligence race intensifies, OpenAI is strengthening ties with major tech companies to maintain its lead.
