Ilya Sutskever just became CEO of Safe Superintelligence after losing their previous CEO to Meta. The AI talent war has gone nuclear: $100M signing bonuses, 4.3% monthly attrition, “pro athlete” pay. This isn’t talent acquisition, it’s talent hysteria🧵

I hope you've found this thread helpful. Follow me @shawnchauhan1 for more. Like/Repost the quote below if you can:.

If you found these insights valuable: Sign up for my free AI newsletter!.

I’m Shawn, a Generative AI Consultant passionate about building AI-driven solutions. I write about AI, startups, and the future of work.

Bottom line: When your CEO quits for a bigger paycheck, you're not competing on vision, you're competing on compensation. The companies that survive this talent war won't be the ones who pay the most. They'll be the ones who build teams that want to stay.

The real opportunity isn't in outbidding Meta for talent. It's in building sustainable teams that can actually execute. Think Amazon's "Day 1" culture, not s talent spending sprees.

What happens next? Three scenarios:. 1. Talent bubble pops, salaries normalize, real building begins.2. Talent forms their own companies, cuts out the middleman.3. AI progress stalls because teams can't stay together long enough to ship.

The smart money is already hedging:. - Anthropic focusing on smaller, stable teams.- OpenAI "recalibrating" compensation (aka cutting costs).- Startups pivoting to AI-adjacent plays with lower talent costs. When everyone's chasing the same 1,000 people, build something different.

Here's what's really happening:. VCs are funding talent wars they can't win. Every $100M bonus is coming from someone's investment round. When talent costs more than revenue, you're not building companies, you're building talent casinos.

The dot-com parallel is eerie:. 2000: Companies paid absurd salaries for "internet talent".2024: Companies pay absurd salaries for "AI talent". Both times, the talent was scarce, the money was infinite, and the music eventually stopped.

Meta's strategy is particularly aggressive:. - $2M+ base salaries for AI researchers.- $100M signing bonuses for top talent.- Still losing people to OpenAI and Anthropic. When you're outbidding everyone and still losing, the problem isn't your wallet.

The irony is rich: Sutskever left OpenAI in May 2024 over safety concerns. Now he's running a "safety-first" startup while the industry burns money on talent that keeps job-hopping. You can't build safe AI if you can't keep your team together.

Safe Superintelligence raised $1B in September 2024 with backing from Google, Nvidia, and A16z. 9 months later, they're scrambling to replace their CEO. If billion-dollar war chests can't retain leadership, what chance do smaller players have?

The real problem? There are maybe 1,000 people on Earth who can actually build AGI. Every company needs them. None want to share. Basic supply and demand says this ends badly for everyone except the talent.

But here's the kicker: most of these mega-offers are failing. Meta couldn't even poach Noam Brown from OpenAI despite $100M+ offers. When $100M isn't enough, you're not in a talent market, you're in a talent bubble.

The talent math is brutal:. - Meta: 64% retention rate in AI division.- Industry average: 4.3% monthly attrition.- OpenAI: "Recalibrating comp" to stop the exodus. Companies are hemorrhaging talent faster than they can hire it.

Here's what happened: Daniel Gross, SSI's co-founder and CEO, jumped ship to Meta on June 29th. Sutskever had to step up as CEO just to keep the lights on. When $1B-funded AI safety startups can't retain their own CEOs, nobody's safe.

The numbers are insane: Meta is offering $100M+ signing bonuses to poach OpenAI researchers. That's not a salary, that's a lottery ticket. When companies pay more for talent than most startups are worth, you know something's broken.

I hope you've found this thread helpful. Follow me @shawnchauhan1 for more. Like/Repost the quote below if you can:.

If you found these insights valuable: Sign up for my free AI newsletter!.