On Recruiting

Recruiting is the founder’s most critical and non-delegable function — the early team is the company’s DNA, and the moment a founder stops directly recruiting and managing everyone, the company loses its ability to go from 0 to 1.

• Founders can delegate almost everything except recruiting, fundraising, strategy, and product vision.
• Outsourcing recruiting introduces a “fly-by-wire” layer: other people won’t have the same selectivity as the founder, and the quality of every hire reflects the founder’s own caliber.
• You will never be able to hire anyone better than yourself early on — all you bring is you, so you must be at the level of the people you want to attract.
• The best people only want to work with the best people. Being surrounded by anyone below their level is cognitive load and signals they should be elsewhere.
  • A test: if you wouldn’t be comfortable letting a candidate randomly interview any member of your current team for 30 minutes, that person you’re worried about is the one you need to let go.

The ideal early hire is a genius: self-motivated, low ego, highly competent, and multidisciplinary.

• Warren Buffett’s triad — intelligence, energy, integrity — plus low ego. Low-ego people scale; high-ego people require constant management.
• Great people are “first among equals” (primus inter pares): each has deep expertise and taste in a specific domain, and peers defer to them in that area.
• They are technical, artistic, constantly generating new knowledge (creative), and automating the tedious parts of their job through code, product, or AI.
  • Automating through process or adding people is the worst form — it introduces cogs who will eventually be replaced and drags down the culture.
• Every great engineer is also an artist — art is anything done for its own sake, done well, and creating beauty or strong emotion. Great engineers express themselves through their craft.
  • Example: AirPods as a work of industrial art — sculpted to fit the human ear, mass-producible, with satisfying magnetic clicks and G3 curves that are hand-sculpted, not computer-generated.

Recruiting requires extreme creativity and rule-breaking — which is exactly why it can’t be outsourced.

• Every hire to a great team requires breaking some rule: compensation structure, location, title, start date, who they report to, equity, hours, or role boundaries.
• The best people don’t fit into neat boxes. Trying to hire for rigid roles is a large-company practice that kills small-company agility.
• Outsourced recruiters or HR can’t break rules they don’t own, and they can’t bring you weird, high-risk, high-reward people because they lack the taste.

Find undiscovered talent before anyone else does.

• If you can easily identify someone, so can everyone else — by the time someone is famous on Twitter or pedigreed with awards, it’s too late.
• Elon Musk’s playbook: pick an audacious mission (Mars, AGI, robots), frame it as large as possible, and be early — this attracts the best engineers before the space gets crowded.
• Source from the edges: look for tinkerers on weird side projects (e.g., unusual ML algorithms for micro-weather forecasting), engage genuinely with their work, and build real relationships — not as a recruiting tactic but out of authentic interest.
• Makers have taste in other makers. Only builders can spot other builders. This judgment cannot be outsourced.
• Example: hiring an assistant spotted at a restaurant who had never worked in tech but demonstrated quality and care in everything they touched.

Early teams are monocultures — and that’s a feature, not a bug.

• Early teams look like cults: monomaniacal, weird in similar ways, and mission-oriented. Mixing too many different kinds of people leads to bland averaging and endless arguments.
• The founder’s personality is the company — their taste, non-negotiables, and values dictate who gets hired. Good founders are extremely judgy and opinionated.
• Products must be opinionated to be simple. Simplicity requires removing every unnecessary click, button, setting, and choice. Settings menus are an abdication of responsibility to the user.
  • ChatGPT succeeded because it was simpler than Google — just talk to it like a human and get a straight answer.
• “Limit the number of details and make every detail perfect.” — Jack Dorsey

Unscale the company to protect creativity.

• Scale is the enemy of 0-to-1 creation. It only takes a small group of brilliant people to build something great.
• Steve Jobs put the Macintosh team in a separate building to prevent cross-pollination. Elon Musk uses standing meetings and tells people to walk out. Bezos uses two-pizza teams.
• At the speaker’s current company, they don’t use Slack. Slack degenerates into low-signal group chat — people asking random questions, politicking, and asymmetrically wasting each other’s time.
  • Without Slack, people must think through problems themselves, then track down the right person — which is interruptive by design, and that’s the point. It forces thoughtfulness and keeps the team small.
• Give people large blocks of uninterrupted “maker time.” Creativity requires boredom, not busyness. Let people be bored rather than filling their days with make-work.
• Never hire an assistant (per Nassim Taleb) — it paradoxically makes you busier by expanding your scale.

Learning happens through iteration, not hours — and most work will be thrown away.

• Mastery isn’t 100,000 hours — it’s the number of learning loops: do something, test the result, distill the insight, make a new creative guess, repeat.
• Great people distill insights from every iteration. A company’s secret isn’t one big bet — it’s thousands of daily insights built on each other.
• “Wandering through the idea maze” (Balaji Srinivasan): take left turns, right turns, and backtracks. The biggest impediment is pride — people lock into their original vision instead of navigating.
  • This is why big companies can’t catch startups: by the time they copy where the startup was, the startup has moved deep into a different part of the maze.
• Good teams throw away far more product than they keep. Most ideas are half-baked, and that’s fine — it’s a search process.
  • The common objection at scale: “We don’t want to try this because it probably won’t work.” The founder must power through this and get the team comfortable with repeated small failure.

Assess candidates by whether they generate new knowledge.

• Peter Thiel’s question: “What important truth do very few people agree with you on?” — tests for independent thinking.
• Naval’s version: “What do you care about that isn’t popular?”
• Even in hobbies: someone generating new knowledge will come up with novel theories about squash within an hour of learning it.
• Self-motivation is non-negotiable. You shouldn’t have to push people or ask “what did you get done this week?” They figure out how to contribute once they know the direction.
• High ego people destroy teams. Low ego + self-motivated + genius-level capability is extremely rare.

Curate people ruthlessly — including firing.

• The prime directive: never compromise on talent. Take a short-term hit on customer experience before taking any hit on team quality.
• “Geniuses only” is the motto — if someone isn’t operating at that level, either you’ve prematurely entered a scaling phase or you need to show them the door.
• You will always make mistakes. If you’re not firing, you’re diluting the team and will only recruit people weaker than the weakest person you kept.
• Don’t hire to fill roles — collect and warehouse geniuses. Great people find ways to contribute even without a defined slot.
• Burnout usually means “I want to quit” — it’s a sign the work isn’t working or isn’t enjoyable, not that someone needs a vacation.
• Sometimes you meet the right person at the wrong time — life problems, health issues, or internal struggles make them unable to function at the level you need. That’s sad but common.
• Fun means you’re learning at the edge of your capability (David Deutsch). If it’s anxiety-inducing, it’s beyond your capability. If it’s boring, you’re not learning.