Why Anthropic’s AI talent rejected Mark Zuckerberg’s Rs 862 crore offers
As this race intensifies, the real question may not be how much AI talent can be paid, but what they’re ultimately working toward/
ADVERTISEMENT
As Meta ramps up its AI ambitions, CEO Mark Zuckerberg has launched an aggressive hiring spree to bolster the company’s Superintelligence Labs. Engineers and AI experts are being courted with jaw-dropping salary packages, reportedly reaching $100 million (Rs 862 crore) over four years, in a bid to push Meta to the front of the generative AI race. But not everyone is biting.
At Anthropic, one of the most influential AI labs today, top engineers and researchers are reportedly turning down Meta’s megadeals. Dario Amodei, co-founder and CEO of Anthropic, recently revealed that despite Meta’s persistent poaching attempts, his team has consistently declined the offers.
In an episode of Lenny’s Podcast, Jared Kaplan, an executive at Anthropic, unpacked the thinking behind the resistance. The answer lies in the company’s mission-first culture.
“The best-case scenario at Anthropic is that you influence the future of humanity, whereas the best-case scenario at Meta is that you make money,” said Kaplan, highlighting the stark contrast in motivation and value alignment.
He added that while he doesn't hold it against anyone who accepts such lucrative offers, financial gain isn’t the ultimate driver for the team at Anthropic.
Interestingly, Kaplan also acknowledged that from a business standpoint, $100 million compensation packages may not be as outrageous as they seem. Given the scale and impact of AI models, such figures can still represent a bargain in terms of the value created.
“Paying individuals $100 million over four years is actually relatively cheap compared to the value that is created for the business,” he noted.
Supporting this view, Perplexity AI CEO Aravind Srinivas emphasised that top-tier AI talent is drawn to purpose, impact, and intellectual challenge—beyond just salary.
“You’re encountering new kinds of challenges. You feel a lot of growth, you’re learning new things. And you’re getting richer, too, along the way. Why would you want to go just because you have some guaranteed payments?” he said.
Zuckerberg’s Superintelligence Labs is Meta’s big bet to catch up with the likes of OpenAI and Google DeepMind. While Meta’s Llama models are improving, they currently lag behind cutting-edge AI systems like GPT-4o and Gemini 1.5 in terms of performance and capabilities.
Despite this, Meta AI is now embedded across the tech giant’s ecosystem—from Facebook and Instagram to WhatsApp—indicating the company’s broader push to operationalise AI in consumer-facing products.
But as this race intensifies, the real question may not be how much AI talent can be paid, but what they’re ultimately working toward—and which vision of the future they believe in.
