A $1.6 billion AI startup is winning talent from Big Tech not with bigger paychecks, but by promising something money can't buy: research that still matters in two months.
The Summary
- Axiom Math founder Carina Hong told Business Insider that AI veterans are leaving Big Tech for neolabs to escape "PSC culture" and constant research pivots
- The startup, valued at $1.6 billion, is building AI systems that solve complex math problems with reasoning, tackling AI's accuracy problem
- The talent war isn't about compensation anymore. It's about research direction stability and knowing your work will ship before priorities shift again.
The Signal
Carina Hong is pulling senior researchers from Meta, Google, and the rest of Big Tech by offering something their six-figure compensation packages can't: certainty that what they build today won't be dead code by summer. At Business Insider's The Long Play event, she laid out why industry veterans are trading job security for research security.
The breaking point isn't money. It's Meta's PSC system and its equivalents at other giants. PSC stands for the biannual performance review gauntlet where peers and managers evaluate you every six months. But the real killer is what Hong calls directional whiplash: research priorities that change weekly or monthly, leaving senior engineers building on sand.
"When you're in a big lab and your research direction changes on like a monthly, if not weekly basis, that's actually another sort of security they are pursuing."
Axiom Math is building AI mathematicians, systems that can solve complex math problems through actual reasoning rather than pattern matching. The company hit a $1.6 billion valuation chasing a specific problem: AI is transformative but still often wrong. That focus matters to researchers tired of being re-org'd into whatever the C-suite decided matters this quarter.
The neolab advantage isn't about ping pong tables or unlimited PTO. Key factors driving talent to neolabs:
- Research direction stays consistent long enough to actually ship
- No biannual performance review theater
- Work has clear impact trajectory beyond quarterly earnings calls
Hong's pitch is working. She's staffed up with what she calls "industry veterans" who've seen the Big Tech machine from the inside and want out. These aren't fresh grads chasing equity lottery tickets. They're senior people who've already made money and now want to make something that lasts longer than the next strategy memo.
The Implication
Watch for this pattern to accelerate. As Big Tech centralizes around closed-source frontier models and quarterly earnings pressure, the best agent infrastructure work will increasingly happen at focused neolabs. If you're building in the agent economy, your talent pipeline might not be Stanford new grads anymore. It's 10-year Meta veterans who are done playing politics and want to ship reasoning systems that actually work.
For researchers stuck in the PSC cycle: smaller companies with specific technical bets are hiring. The trade is job security for research security, but if you're senior enough, you're probably not worried about getting hired again anyway.