Brian Chesky on Founder Mode, the CEO Job, and Redesigning Airbnb for AI

Airbnb’s Brian Chesky argues that founders are born but CEOs are made — the CEO job is counterintuitive, has to be learned, and in the AI era demands an even more intense version of founder mode: fewer layers, no pure people-managers, and a company redesigned from scratch.

Key ideas

1. No one is born a good CEO. Founding is innate — you learn by doing, which suits a founder. The CEO job is counterintuitive, and trial-and-error is expensive: a mis-hire builds an empire, leaves, and costs years to unwind. Chesky learned it the hard way when the pandemic took 80% of Airbnb’s business in eight weeks and forced him into wartime founder mode.
2. Founder mode becomes AI founder mode. With information on demand, the leader can go deeper into the details than the old meeting-heavy cadence allowed. The redesign points to asynchronous work, far fewer management layers, and the end of the pure people-manager — every manager must stay technical (‘you manage people through the work’).
3. Make the problem as small as possible. A monopoly of a tiny market beats a small share of a big one; heat a bathtub, not an ocean. Airbnb’s Project Hawaii ran ten-person startup teams inside the company, and new verticals now go one → ten → many rather than launching at global scale.
4. The founder-led endurance paradox. The longer a company stays in founder mode, the more it can be handed off and endure — Disney and Apple ran the founder’s playbook for decades after his death. A ham-sandwich business and a founder-shaped ceiling are both true; the founder builds the moat that lets a ham sandwich run it later.
5. From adulation to craft. Chasing status is a cup with a hole in the bottom; the artist makes for himself. Chesky reframes the AI-era task as redesigning Airbnb’s atomic unit from a home to a person — authenticated identity, a preference library, a real-world social graph — and disrupting the app before agents do.

Summary

No one is born a good CEO

Chesky opens on the thesis that organises the whole conversation: people are born founders — it is innate, learned by doing — but no one is born a good CEO, and the job is counterintuitive enough that most of a founder’s instincts are wrong. Trial-and-error, the founder’s native mode, is ruinous in the CEO seat, because a wrong senior hire builds an organisation that then takes years to unwind. Through the 2010s he over-delegated to professional managers and felt he was ‘in a car without a steering wheel’ — a company of 7,000 running itself. The pandemic, which erased 80% of the business in eight weeks, forced the shift from peacetime to wartime; he took control of every review for two to three years, not to micromanage forever but to know enough to empower people later. His formative training is industrial design (RISD): the designer is the product manager, judged only if the product sells, so marketing, manufacturing and the user’s journey are one job.

From founder mode to AI founder mode

Founder mode — Paul Graham’s coinage for Chesky’s practice — was mechanically a meeting culture: ~35 hours a week of group reviews, a full chain of command in the room, the CEO speaking last and ratifying every decision. AI inverts the mechanism. With everything on demand, the leader goes into more detail asynchronously, with fewer meetings and fewer layers (he invokes the Catholic Church’s four layers against corporate seven-to-nine). Two kinds of people will not survive: pure people-managers who run one-on-ones like a therapist, and the rigid who will not evolve. His rule — ‘you manage people through the work’ — means every manager stays a hybrid IC: engineers code, and ‘whatever the version of that is for your field, you need a[t] code.’ He predicts a consumer-AI renaissance in 12–24 months, against today’s near-total enterprise tilt (159 of the last YC batch’s 175 companies were enterprise).

Make the problem as small as possible

The recurring operating principle is compression of scope. Peter Thiel’s counsel — better a monopoly of a tiny market than a small share of a big one — pairs with Paul Buchheit’s Gmail rule: ship only once 100 people inside love it, because once 100 love something, 100 million will. You cannot heat an ocean; heat a bathtub. Project Hawaii applied this internally: a lean team of ten-to-twelve treated like a startup, crawl → walk → run → fly, that produced ~$200m of internal revenue in year one and hundreds of basis points of conversion. New businesses now follow one → ten → many — pilot one city, prove product-market fit by hand, then industrialise — the discipline Airbnb had lost when it launched Services and Experiences in 100 cities at once.

The founder-led endurance paradox

Chesky reconciles two opposed investing ideas — Buffett’s ‘buy a business a ham sandwich could run’ and the view that a founder’s growth sets the company’s ceiling — through Walt Disney. Disney was a personal expression, founder-led for decades, and 50 years after Walt’s death his spirit is still omnipresent (feature animation, TV, theme parks), while rival studio moguls are forgotten. The counterintuitive claim: the longer a company stays in founder mode, the more it endures after the founder, because he institutionalises the magic and builds a moat deep enough that a ham sandwich can later run it — Apple still runs the iPhone Steve Jobs left. He caveats that technology is ‘a cinema for change’, so tech companies need founder mode for longer than Buffett’s Coca-Colas.

Redesigning Airbnb for the AI era

The forward task is a ground-up redesign. Chesky wants to change Airbnb’s atomic unit from a home to a person: the most authenticated identity on the internet (proof of personhood matters in an age of AI), the richest preference library, a real-world social graph, and a membership programme — a person at the centre with fifty things (homes, experiences, services, eventually flights) ringed around them. Two further fronts: industrialising from three businesses to fifty, and disrupting the app with AI without breaking a public company on whose earnings hosts’ livelihoods depend (hence separate sandboxes — ‘what’s after Airbnb?’). Underneath sits the mindset shift: detaching from adulation (‘a cup with a hole at the bottom’) and status toward intrinsic craft — the artist who, like Rick Rubin’s, makes for himself. He closes on recruiting as the CEO’s first job (he co-hires the top 200, builds pipelines not searches, ‘start with the results, work backwards to people’) and on belief as the kindest gift — seeing potential in people they do not see in themselves.

Speakers

• Brian Chesky — co-founder and CEO of Airbnb; industrial designer by training (RISD), and the founder whose practice gave rise to the term ‘founder mode’.
• Patrick O'Shaughnessy — host of Invest Like the Best; CEO of Positive Sum and chairman of O’Shaughnessy Asset Management.

See also

• Brian Chesky on Airbnb and Product — his prior appearance; the original statement of founder mode, the divisional decay arc, and product-marketing integration.
• IC CEO — the AI-era extension of founder mode to hands-on product-and-code participation; Chesky’s ‘AI founder mode’ is the same claim from the founder’s side.
• Product Craft and the Role of Design — the design-led product philosophy running through Chesky’s industrial-design training and craft-over-status turn.
• Culture as Product — treating the organisation, brand, and community as the durable artefact when the software is ephemeral.
• Patrick O'Shaughnessy — the host.