Keith Rabois on the Art of Hiring

The most useful piece of advice Keith Rabois says he got in 23 years in tech came from Vinod Khosla: "The team you build is the company you build." Get the people right and every other problem — product, market, design — becomes solvable. Get them wrong and, as Stripe puts it, every problem is a people problem.

And yet, Rabois points out, the most important thing about building a company is the one thing nobody has written a usable playbook for. He's a voracious reader; he's tried every book on hiring; he wouldn't recommend one to a founder. That gap is exactly why his "Art of Hiring" talk is worth your time:

Watch the full talk on YouTube →

"Don't hire the obvious people"

Rabois's most counterintuitive rule is to avoid proven, obvious candidates. His mentor Peter Thiel put it bluntly on a jog around Stanford: don't hire anyone over 30. The logic — by 30 there are enough data points on you that the whole market reaches a consensus about your ability, and once there's consensus, Google, Meta, or OpenAI will outbid any startup for you. Startups win by finding undiscovered talent and building a system that leverages it — the way the 1980 US hockey team beat a Soviet all-star squad with college kids and the right system.

At PayPal that meant something concrete: no general managers. Whoever led design had to be the best designer; whoever led engineering, the best engineer. "Managing people" was never itself a hire.

Barrels and ammunition

The idea most people take from the talk: most of your hires are ammunition — reliable executors. They're not hard to find, and you'll mostly get them right. The rare, valuable people are barrels — the ones who take something from idea to shipped. Barrels are weird ("usually a screw loose"), which is exactly why they're hard to assess and easy to screen out. Rabois is blunt: the people who created the most value at PayPal were the weirdest people there.

Two cheat codes: hire interns (the most upside of anyone — the best go on to found companies), and clarify whether you're hiring to create value or protect it. For value creation, experience is often a handicap — PayPal had 2 people out of 254 who knew financial services. For value protection — general counsel, CFO, compliance — experience helps you see around corners.

You can't delegate the decision

The uncomfortable part: assessing talent is the founder's job, and it doesn't transfer. Jack Dorsey and Rabois personally interviewed the first 400 people at Square; Tony Xu interviewed the first 1,500 at DoorDash. Being world-class at your craft, Rabois warns, is not the same skill as sizing up a stranger who just walked in — and most leaders are good at one, not both.

So what predicts success? He borrows an answer from Derek Jeter, who said he could spot an all-star rookie every time — not from their swing, but from their confidence. Rabois looks for risk-taking (the instinct, not a résumé full of risks — he tells the story of hiring a McKinsey consultant he'd only seen play an absurdly aggressive goalie on his soccer team) and, above all, for people who can positively manipulate the world around them — reframe reality, at a micro or macro level. One thing he deliberately does not screen for: passion for your industry. As Patrick Collison told him, it's arrogant to assume a great candidate wakes up caring about your company — it's your job to make it compelling.

An incomplete team beats a mediocre one

"Mediocrity spreads. Mediocrity compounds." Rabois would rather leave a role open than fill it with an A-minus, because once a mediocre team sets in it's almost impossible to fix. The rare exception is a role where not hiring blows up the company — a compliance hire in fintech, say.

Speed is the startup's advantage

Two closing tactics worth stealing:

• Jack Dorsey's question before extending an offer: "Is there anything that would prevent you from accepting this offer?" It surfaces objections while you can still address them.
• Move fast. Candidates love the team that gives them an offer. Keep the loop short — Rabois wouldn't have a candidate meet more than five people — because every extra interviewer adds little and costs you speed.

And build a cult in the Thiel sense: a shared, contrarian view of the world, and only hire people opting into it. Square was design-driven enough that it made accountants clear a design bar — which is how a design-obsessed accountant named Lauren became the company's best PM.

Where PerfectLaunch fits

Notice what Rabois keeps circling back to: a consistent bar, owned by the founder, applied with discipline and speed. He isn't telling you to outsource judgment — he's warning that most teams have no system behind the most important decision they make. No set stages. No fixed questions. Four interviewers, four different bars.

That's the exact gap PerfectLaunch closes. It doesn't decide barrels versus ammunition for you — it gives that judgment a process: a rubric per role, structured stages, live AI interview coaching, and every answer scored, so the bar holds whether you're interviewing your fifth hire or your five-hundredth. The call stays yours. The consistency, the cadence, and the evidence come built in.

Your next hire is being decided this week. Give it the process it deserves →