What the Mass Talent Exodus Could Mean for the Future of AI

A few years ago, I started noticing something strange in San Diego.

Engineers I'd known for years – people who had spent their careers at Qualcomm, General Atomics, and other legacy defense contractors – were quietly disappearing from their jobs.

Not because of layoffs or retirement. They were walking across town to new jobs at companies most investors had never heard of...

Anduril. Shield AI. Oura.

The people leaving were among the best in their fields – autonomy specialists, machine-learning experts, and the engineers other engineers wanted to work with.

I didn't have access to their offer letters. I couldn't see their stock options. But I could see movement.

These engineers were moving into artificial intelligence ("AI").

Within months, the announcements started rolling in: massive funding rounds, major defense contracts, and valuations that made national news.

The capital followed the talent. As I'll show today, it almost always does...

Talent Is Truth in Motion

I've written before about observable surfaces – places where real behavior reveals what's coming before Wall Street models it.

Consumer behavior is a big one, as I said back then. But there's another surface hiding in plain sight: where the best people choose to work.

Employees at tech firms – especially top engineers and executives – have information that never appears in U.S. Securities and Exchange Commission filings. They know whether the technology actually works. They know if leadership has a real plan. They know when the company culture starts to rot.

Nondisclosure agreements keep them quiet. But they can move. And when they do, they're telling you everything.

How Employee Departures Powered Two AI Startups Worth Billions

What I watched happen locally in San Diego has played out on a massive scale in the AI space.

Notably, it happened with ChatGPT creator OpenAI...

Back in December 2020, Dario Amodei was OpenAI's vice president of research. He had helped build GPT-2 and GPT-3 – the models that made OpenAI famous. His sister, Daniela Amodei, ran safety and policy.

Then they left. And they didn't leave alone.

A cluster of senior researchers – the very people who had built the technology – walked out the door. Together, they founded a new startup...

They called it Anthropic.

At the time, most people didn't notice. OpenAI was raising billions. ChatGPT was about to change everything. Why pay attention to a handful of quiet departures?

But those departures were the signal. The people closest to the technology had seen something about the company's priorities and where the real opportunity was.

Today, Anthropic is valued at $350 billion. Its revenue grew 10-fold in each of the past three years. According to SignalFire's 2025 State of Tech Talent Report, OpenAI employees are now eight times more likely to leave for Anthropic than vice versa.

The best people vote with their careers. And right now, they're voting decisively.

AI's massive market capitalizations also tend to follow the important names. You don't pay "war prices" for talent unless you believe you're fighting for the future.

In May 2024, Ilya Sutskever left OpenAI to start a new venture called Safe Superintelligence, which investors valued at $32 billion by April 2025.

It had no product. No revenue. Just Sutskever and a handful of researchers who had touched the frontier.

That's what talent is worth when you've seen the future firsthand.

The same pattern showed up at upstart defense contractor Anduril. When this company set out to challenge legacy players, it recruited Tom Keane – who had spent 20 years at Microsoft building Office 365 and Azure – along with leaders from Palantir Technologies, SpaceX, and Splunk.

The result? Anduril grew from 700 employees to more than 4,000. Its valuation climbed from $1 billion to $30.5 billion... And now it's winning contracts once reserved for companies a century old.

The talent moved first. The money followed.

Your Edge as an Individual Investor

Here's an uncomfortable truth: The best employees always have options.

When a once-booming company changes trajectory, the most talented people see it first. And because they're the most in-demand, they're the first to find better opportunities.

By the time an exodus is obvious to outsiders, the top talent is long gone.

You don't need insider access to track talent movement. The surfaces are public...

• LinkedIn: This career-networking app shows job changes in real time. If you notice five engineers from the same team moving to the same startup... that's no coincidence.

• Glassdoor: This site's workplace reviews reveal sentiment before it becomes news. Rising complaints about leadership or sudden silence from a once-active company are early warnings.

• Hiring patterns: These trends speak volumes. Freezes mean trouble. Unexpected surges mean urgency.

• Quiet executive departures: This is often what matters most. A CEO leaving makes headlines... But a chief product officer slipping out the back door is often a bigger signal that most folks miss.

Institutions can't act on this information. They won't green-light trades based on Glassdoor reviews. And you can't plug "top five engineers quit" into a financial model.

But as an individual investor, you don't have those constraints.

You can check LinkedIn in minutes. You can notice when the people building the future start clustering in one place.

The truth shows up in movement before it shows up in results.

When engineers leave, they know something. The only question is: Are you paying attention?

Good investing,

Josh Baylin