Waymo, Texas Culture, Airline Lounges, OpenAI & Uber Eats - Rory Sutherland

Rory detailed the Eleven Madison Park case: the restaurant discovered that even the world's number one had mediocre coffee and treated beer drinkers as second-class. By appointing a coffee sommelier and a beer sommelier with pairings, they transformed a small part of the experience into something world-class. Over time, that reputation attracts people who specifically value those details. He applied the same logic to airports (London City Airport focused on speed and lack of shops when all other airports were becoming mega-malls) and hotels (Moxy doubled down on the ground-floor communal space, making you feel welcome even after checkout). He contrasts this with the typical 'benchmarking is for losers' trap described by Roger L. Martin.

Rory recounts a personal story: while charging his electric car on the outskirts of Florence, he had 20 minutes to kill. He wandered around a corner, found a leisure center and public pool (not a tourist spot), then a fantastic local cafe. He contrasts this with the 'curse of Tuscany' — too many famous sites within a short drive creates obligation to visit them all, ruining relaxation. He also notes that a bad hotel is less of a disaster if you have a car because you can escape; on an island, the constraint of limited options can be freeing on a second visit because you've already done the 'obligatory' sights. He recommends islands for repeat trips, cars for first explorations.

Rory describes how he 'really hacked this' and gives precise directions: come in from the back, go left toward lower numbers, past business class lounge, all the way down, left, keep going, and there's a small Pret. He notes that retail placement in airports is highly uneven, and knowing these Easter eggs turns an annoyance into a small win. He also mentions a former EasyJet hack where a pillar hid an extra check-in desk, making the queue move twice as fast. The broader point is that airports should deliberately seed secret shortcuts that only frequent users discover, like the London Underground's hidden passages that locals know but can't be advertised because they couldn't handle the traffic.

Rory notes that Amsterdam Schiphol is an 'annoying' airport because it's endlessly expanded into a single long terminal (gates D, E, F, G) rather than multiple terminals, making it a long walk. He discovered that gate D2 has a bench specifically without armrests, unlike the intentionally hostile 'hostile architecture' seating elsewhere that prevents homeless people from sleeping. He says 'I've spent many of time at' this bench, and recommends it as an Easter egg. This aligns with his larger argument that airports should have secret shortcuts and hidden amenities known only to frequent users.

Chris describes the feature: under 'Suggestions' there's car hire, bikes, stuff for teens, then 'Get anything delivered' with subcategories like health, personal care, baby, gourmet, pet supplies, flowers, retail, electronics, and courier services. He notes you can even send an Uber courier to retrieve a forgotten item (e.g., a watch at the gym). Rory compares this to the 'Starbucks PR dilemma' — Uber is known for rides and food, making it hard to expand user perception into other categories, much like Pret struggles to sell more coffee and Starbucks to sell more food. The feature is there but awareness is low because users' mental model is narrow.

Rory advises against paying for features you don't value, like a top school district if you have no children. He suggests looking for houses next to a pub (if you're a heavy sleeper and don't mind noise) because many buyers will avoid it, lowering the price. He speculates that with car electrification, being next to a busy road will become less noisy and less polluted, making it a long-term play. He criticizes property search sites for not allowing users to filter for negatives — things that are deal-breakers for others but acceptable for you. He also suggests the rule 'buy the cheapest house on the most expensive street, not the most expensive house on the cheapest street.'

Rory recounts how he told Expedia this at their headquarters. He says when he sees 'adults only' his first thought is 'perverts' or imagining spending the week in a mask while a German dentist urinates on him. He ran a Twitter competition and received 15 alternatives, the best being simply 'grown-ups'. He emphasizes that the phrase 'mature hotel' would also be bad because it sounds like a retirement home. The underlying principle is that word choice in consumer categories carries massive latent associations that designers often overlook — similar to how 'disabled wheelchair' and 'walk' have no intermediate option at airports for the mildly elderly.

Rory notes that Spotify successfully suggests new bands he wouldn't have picked, saying 'this really know me'. He extends this to the earlier point about iterative preference refinement: you train the algorithm over time, much like you'd want an AI agent to search for a house or a toaster — not just by price and size, but by learning what trade-offs you end up happy with. He contrasts this with dating apps, which he likens to crude property search filters (how many bedrooms, location), arguing that dating should be highly iterative because what you think you want and what makes you happy often differ. He even jokes about the 'air fryer girlfriend' — an experience that underpromises and overdelivers, which an algorithm might miss if it only looks at explicit specs.

Rory critiques modern luggage design for going in the wrong direction: multiple sealed compartments seem organized but create 'what zip did I leave it in?' moments. He says it's 'like making love to a goth — there's just too many zips.' He describes a friend who uses a workman's bag, essentially a tray with handles, where you can see everything from above and pick out items without opening any zips. This is a practical travel hack, but it also ties to his broader theme that features are often added for perceived organization at the cost of usability. He mentions he carries an open bag himself.

Rory credits Roger L. Martin for the critique of benchmarking, then expands it with his own concept of 'reverse benchmarking'. He tells the story of Will Guidara at Eleven Madison Park: after dining at the world's number one restaurant, Guidara asked his team not 'what did they do well?' but 'what was a bit disappointing?' The answers were: the coffee was nothing special, and beer drinkers were treated as second-class citizens compared to wine drinkers. Back home, he appointed a coffee sommelier and a beer sommelier, giving the minority of diners who ordered those an unexpectedly stellar experience. Rory extends the principle to hotels — the Moxy chain doubled down on the ground-floor communal space rather than rooms — and to airports, where London City Airport focused on speed and proximity to gates instead of shopping. The idea is that you only need to be 'top decile' in what everyone else does, then find an ignored dimension and turn it into a flagship feature. He calls it a 'generalized theory of innovation', akin to what Steve Jobs did by making tech beautiful when everyone else ignored aesthetics.

Rory describes ordering Waymos through Uber in Austin and noticing that the estimated arrival time was often 50–100% longer than predicted. He theorizes that because the car has blacked-out rear windows, no visible driver, and no capacity for road rage, other drivers and pedestrians treat it with zero deference. Pedestrians step in front of it knowing it will stop; drivers don't let it out at junctions because there is no fear of retaliation. He contrasts this with the subtle 'social calculus' human drivers perform — for example, your willingness to let someone in depends on how fast you're going; if you're stuck in traffic, the cost is low. He also notes the Japanese invention of hazard-light 'thank yous' and how acknowledgment matters more than the cost of the favor. He worries that a generation of urban young people who don't drive frequently — or only rent cars occasionally, never reaching the 'system one' fluency — miss out on this domestication effect, which he says 'if you're not careful can be quite psychopathic'. He ties this to extended adolescence and reduced car ownership among Gen Z.

Rory illustrates with parking apps: initially, you could pay by app or by coin meter — the app was an option that made life easier if you had no coins. Then parking operators removed the coin machines because digital payment was cheaper, and suddenly the app became an obligation, alienating elderly people without smartphones. He applies the same logic to the two-income household: when women entered the workforce, it was an option to enjoy a higher material standard of living. But over time, house prices rose to absorb the additional household income, so that a single-income family could no longer afford the same home. The option became an obligation — both partners must now work just to maintain what one income used to provide, while losing 40 hours of discretionary time per week. The gains flowed to property owners and the tax base, not to families. He uses the analogy of drinks parties (optional: you can skip if you don't feel like it) versus dinner parties (obligatory: you must have a good reason to cancel) to make the distinction visceral.

Rory draws the analogy between dating profiles and products: an air fryer looks modest but upon use becomes indispensable — if it broke, you'd buy another instantly; that's high repeat-purchase intent. A sleeping bag, however, looks impressively compact in its packaging but becomes a nightmare to repack, especially when wet — high initial promise, low long-term satisfaction. He extends this to the 'IKEA effect' (building furniture yourself increases perceived value, much like pick-your-own strawberries) and notes that companies obsessed with short-term metrics overinvest in acquisition because it's easier to measure than retention. He suggests that platforms like Amazon should show a 'repeat purchase o-meter' — not just how many bought an item, but how many bought it again. This would help consumers identify true air-fryer-girlfriend products. He also connects it to electric cars, which have a high repeat rate because the upfront learning cost (charging infrastructure) acts as a sunk cost that increases loyalty — similar to women playing 'hard to get' in dating.

Rory speculates that the natural direction of travel for an AI-empowered world is the opposite of today's model: instead of companies appointing advertising agencies to find consumers, consumers will appoint AI agents to find products. This is already visible in Spotify and YouTube algorithms that learn preferences and make recommendations the user wouldn't have chosen themselves, but the logical endpoint is an AI that handles all search — not just music but rental cars, toasters, houses. He argues that current property search or dating apps are crude because they're not iterative: you can't refine preferences through the marketplace feedback. An AI agent could do that. He concedes this could make his own job in advertising redundant, though not entirely.

Rory argues that income inequality, while real, is tiny compared to wealth inequality, which is largely driven by land ownership. He points out that rising house prices have been reported as 'good news' for decades, which is monstrous because it ignores that for non-owners it represents exclusion. He introduces Henry George's philosophy: you keep 100% of the fruits of your labor (no income tax), but you pay tax on the land you own because you didn't create the land — you're a custodian. He notes that Texas, often stereotyped as a conservative state, has heavy property taxes (~2.5%) which make land cheaper to buy and discourage speculative hoarding. He recounts a woman in a £3–4m house who has no liquid money and whose children can't afford shock absorbers, illustrating how property wealth is illiquid yet destructive. He recommends the book The Inheritocracy and says even libertarian economists see a role for government in collecting and sharing information to empower consumers, such as repeat-purchase rates.