Why Anthropic Cannot Buy the Loyalty It Is Paying For

Anthropic pays the AI industry's biggest salaries, yet CEO Dario Amodei reportedly worries new hires care more about money than the mission.

A single line in an Axios report on AI talent wars set off days of mockery this week. A source said Anthropic CEO Dario Amodei has grown worried that new hires join for the money, not the mission.

The internet did not let that pass quietly. Anthropic reportedly pays more than any other AI lab, including OpenAI. One widely shared reply put it bluntly: "Breaking news, people work for money."

The Joke Everyone Made

Engineer Gergely Orosz, who writes the widely read Pragmatic Engineer newsletter, landed the sharpest version of it. If mission mattered more than pay, he argued, a company could simply offer less and see who stayed.

The numbers back him up. A brand marketing events lead role at Anthropic lists a salary range of $320,000 to $400,000. The national average pay for that job is around $59,400, according to the US Bureau of Labor Statistics. Anthropic's offer runs roughly six times higher.

Distinguished AI researchers at top labs can earn $1 million a year with little trouble. Some are now negotiating sign-on equity ahead of potential IPOs at Anthropic or OpenAI, according to Axios.

People Keep Leaving Anyway

Underneath the punchlines sits a real hiring headache. Frontier AI has collapsed into one small, tightly connected talent pool, and the same few hundred researchers keep circulating through it:

• Lilian Weng, a co-founder of Thinking Machines Lab, left last week and turned up at OpenAI days later. She was the fourth co-founder to exit that company in a year.

• Google lost researcher Noam Shazeer to OpenAI and Nobel winner John Jumper to Anthropic in June.

• Meta spent heavily to assemble a superintelligence team under Alexandr Wang, then watched several of those hires leave again, some straight to OpenAI.

More than 1,300 employees across AI labs recently signed a letter warning that development could be outpacing safety controls. The people building these systems are often the ones most publicly worried about them.

What Money Cannot Buy?

Once every lab can offer millions, salary stops distinguishing anyone. Mission becomes the one lever left that pay can't fake. Researchers told Axios that compensation isn't their only draw. Compute access, creative control, and their agreement with a lab's approach all shape where they land.

Retention pressure like this only works against a company if it also has to keep shipping. Cynoteck tracked how fast Anthropic has been moving on the product side in its coverage of Claude Opus 5, the company's fourth new model in under two months. A pace like that depends on a research team that stays put, not one that's constantly rebuilding itself.

A Problem That Isn't Just Anthropic's

Nothing about this is specific to one lab. Any company competing for AI talent right now faces the same math: a tiny labor pool, sky-high pay ceilings, and researchers who can walk within weeks for a better deal elsewhere.

Businesses trying to build an internal AI team should plan to compete on more than a paycheck. Interesting problems, genuine autonomy, and access to real compute now carry nearly as much weight as salary. A pure pay strategy eventually loses because a bigger lab can always beat it by one more dollar. Teams that can't win that fight outright can still build real capability by working with Cynoteck's machine learning development services rather than trying to staff every AI role in-house from scratch.

The Short Version

• Anthropic CEO Dario Amodei reportedly worries that new hires join for pay rather than mission, despite Anthropic paying the industry's highest salaries.

• One Anthropic events role lists a salary roughly six times the national average for that job.

• Frontier AI talent moves in a small, interconnected pool, with researchers jumping between OpenAI, Anthropic, Google, and Meta within months.

• Researchers point to access to computing, influence, and ideology as motivators alongside pay.

• More than 1,300 AI lab employees have signed a letter warning development may be outpacing safety controls.

Where This Goes From Here

Expect churn to continue as labs move toward potential IPOs and researchers negotiate equity to match. The number that matters here was never the salary figure. It's which labs manage to hold onto their best people once the paycheck stops being the most interesting part of the offer.