- Caffeine & Chaos By Zubin
- Posts
- Retain A+ Talent: Why $100 Million Isn’t the Magic Answer
Retain A+ Talent: Why $100 Million Isn’t the Magic Answer
Meta waved $100m at OpenAI stars and most said “nah.” Money gets attention, mission keeps loyalty. Here’s the playbook for hanging onto your best people.
Zubin Teherani
August 01, 2025
$100 MILLION. WHAT?
Sam Altman told Reuters that Meta dangled $100 million packages at key OpenAI team members. Some offers were rumored to top $300 million over four years.
Meta pulled out all the stops. Recruiting dinners, direct WhatsApp messages, and the equivalent of Tinder Super Likes from Mark Zuckerberg himself.
The result: nearly everyone stayed put. Three researchers left; the rest replied, “We’re good.”
Somewhere a Meta recruiter is screaming in a glass cage of emotion.
Why Giant Checks Didn’t Seal the Deal
Altman’s explanation: mission. People want to build superintelligence responsibly, not just refinance a beach house.
DeepMind’s Demis Hassabis called Meta’s move “rational”, then reminded everyone purpose > payroll .
Turns out “changing the world” still edges out “changing the trim on my new yacht.”
An Inside Look at OpenAI’s Culture
A viral Hacker News article from someone working at OpenAI provided an in-depth look at the culture of OpenAI and why it’s so unique. It’s worth a read, but it essentially boils down to the following:
Pillar | What it looks like on the ground |
|---|---|
1. Clear, lived-in mission | Every project ladders to “benefit humanity”. The stakes feel high when hundreds of millions users use your product for everything from medical advice to therapy. |
2. Insane talent density | With 17 people, Codex (OpenAI’s agent) was shipped in 7 weeks. Only an absurdly gifted squad pulls that off. |
3. High trust, low politics | Debates are public, searchable threads. Interns challenge VPs, and the best argument wins. No memo wars, no calendar theater. |
4. Autonomy and bias towards action | OpenAI has no product roadmap. Have an idea? Spin up a prototype (four Codex variants existed before launch). Something not working? Stop and pivot. |
5. In the trenches leadership | Execs are highly visible and manage the work, not the people. Managers of managers are rare and credibility comes from contribution, not PowerPoint. |
6. Fanatical customer feedback loop | OpenAI closely monitors customer chatter. If a tweet goes viral, someone will probably look at it. |
7. Ethics matter | OpenAI spends a lot of time trying to do the right thing. Massive resources go to hate-speech filters, self-harm triage, political-bias audits. Nothing ships until red-team sign-off. |
How to Build “I’m not leaving” Energy
Practice | What It Looks Like | Quick Sanity Check |
|---|---|---|
Nail the North Star | Everyone from new hire to CEO can explain why the org exists in one sentence. | Ask a day-2 hire; if they mumble, fix the docs. |
Give Real Ownership | People own problems, not chores. “Here’s the outcome. Now go.” | If you hear “Am I allowed?” more than “How might I?”, tighten trust screws. |
Make Debate Public, Not Political | Design reviews and decisions live in open threads; best idea wins. | Last time a junior’s comment changed a decision? If you can’t recall, open the floor. |
Bias for Action | Idea → prototype in days, not quarters. Small bets, fast kills. | Count new features this month. Fewer than four? Unblock. |
In-the-Trenches Leadership | Execs roll up their sleeves. Credibility = contribution. | Can ICs list three things their VP did this quarter to move the business forward? |
Fair ≠ Flashy Comp | Pay removes anxiety. Equity or profit-share links everyone to upside. | If comp is your only retention lever, you’re in a bidding war you’ll lose. |
Perks are like seasoning. Use enough to enhance the meal, not enough to mask expired chicken.
TL;DR for founders, HR, and anyone sweating poach attempts
Money lands the first date. Mission seals the marriage.
Ownership + trust cost far less than $100m and compound for life.
Culture is the only moat that never shows up as depreciation on the balance sheet.
Ask yourself: Would your top performer shrug at a giant offer? If the answer is “probably not,” start doubling down on purpose and culture before the next recruiter slides into their DMs.