Craig Newmark on Institutional Maintenance, Giving Away Control, and the Internet We Were Promised
Recorded live at 92NY, Tyler Cowen presses Craig Newmark — founder of Craigslist, now a full-time philanthropist — on why he built one of the internet’s most useful sites and then deliberately declined to grow, monetise, or run it, and what that same instinct to subtract himself has produced in cybersecurity, veterans’ welfare, and journalism funding.
Key ideas
- Simplicity survived because Newmark gave up control early. Craigslist’s famously plain design was originally an accident of Newmark’s own limited design skills, but it became a deliberate principle once Jim Buckmaster took over management in 2000: simplicity and speed as the standing design criterion, resisting attractive redesigns nobody asked for.
- ‘Enshittification’ explains why the rest of the web did not stay that way. Newmark credits Cory Doctorow’s term for the pattern by which successful platforms need venture capital to keep competing, investors demand more extraction, and usability degrades as a result — the throughline connecting his account of 1990s web design decay to Cowen’s later question about restaurants and institutions declining once they get popular.
- Philanthropy as ‘networks of networks’, not personal control. Rather than running programmes himself, Newmark funds capable intermediary organisations — Blue Star Families, the Bob Woodruff Foundation, the CUNY Journalism School, Shadowserver — and calls this delegation his most effective current form of estate planning, since the networks he funds will outlast him.
- Trust is a design problem, and Newmark distrusts easy fixes. His ‘nine-second rule’ (pause before acting on urgent-seeming requests) targets the manufactured urgency at the heart of most scams; he explicitly rejects handing user-authentication entirely to government, preferring a plurality of private verification systems (credit-card networks, driver’s licences, digital wallets).
- A quiet regret: two large, unstructured Ivy League gifts. Newmark’s clearest stated philanthropic mistake was giving lump-sum contributions rather than staged, milestone-based multi-year funding — a lesson he says now shapes how he grants money, alongside an insistence that grantees invest seriously in communicating their own work.
Content
Design, control, and enshittification
Cowen opens with a pointed question: why has web design gotten worse for thirty years? Newmark’s answer starts with an accident — he had no design skills when he built the first Craigslist page, so simplicity was a default, not a philosophy. It became a philosophy once he relinquished management control to Jim Buckmaster in 2000; Buckmaster has kept the site deliberately plain ever since, resisting the ‘bells and whistles’ Newmark says most sites accumulate. Asked what went wrong with the wider 1990s web, Newmark points to competitive pressure: successful sites needed venture capital to survive, investors demanded more extraction of value from users, and the result is what Cory Doctorow calls ‘enshittification’ — a term Newmark credits Doctorow with articulating better than he could himself. He calls Doctorow ‘a bit of a hero’, both for his advocacy on digital rights and for his willingness to ‘put a target on his own back’.
The same decay pattern resurfaces later when Cowen proposes a general model of institutional life: a new restaurant is best in its first six to nine months, then popularity brings less-informed customers, declining quality, and permanent queues — so the rational strategy is to find peaks early and move on. Newmark does not adopt this as a personal rule (his own restaurant test is closer to ‘does it look like a hole in the wall’), and the conversation leaves open how far the restaurant-and-institution analogy generalises, but the throughline back to enshittification — competitive and financial pressure eroding an initially good product — is explicit.
Subtraction as method: from CEO to philanthropist
Cowen quotes a GPT-generated observation back to Newmark: his biggest decisions all look ‘subtractive’ — minimal monetisation, stepping aside as CEO, continuing personal customer service, and now funding others rather than running programmes himself. Newmark pushes back gently on the framing. He does not experience it as subtraction but as recognising his own limits as he has aged, and discovering that he has ‘accidentally’ built networks of networks that get things done more effectively than he could alone — funding capable intermediaries such as Blue Star Families (military families) and the Bob Woodruff Foundation (veterans) rather than operating programmes directly. He treats this as his most effective current form of estate planning: as he ages, the networks he has funded have ‘a life of their own’ and will outlast him, an alternative to either building a personal AI proxy or trying to remain personally indispensable.
Newmark is candid about his own limitations elsewhere in the conversation. He describes his social skills as largely ‘fakery’, sustainable for perhaps ninety minutes at a stretch, and says he has stopped trying to improve at it — focusing instead on positioning people who do have those skills (Kathy Roth-Douquet of Blue Star Families, for instance) to do the parts of the work he cannot. He answered a large personal volume of Craigslist customer-service email for years — a figure he first recalls as 40,000 in a single year, then revises on reflection to something closer to 4,000 — and treats the exercise as the source of his central finding about people in general: the overwhelming majority everywhere just want to get through the day, and genuinely bad actors are a small minority.
Trust, scams, and the nine-second rule
Newmark’s stated views on trust have a deliberate asymmetry: he became more trusting of ordinary users through years of customer service, while becoming more wary of a small class of confirmed bad actors. He describes learning the hard way that an honest person — a self-described ‘ninth commandment’ guy — is structurally disadvantaged against ‘someone who lies for a living’, a lesson that took years to internalise. His practical countermeasure is the ‘nine-second rule’: pause before acting on anything urgent-seeming in email or on social platforms, because manufactured urgency is the common mechanism scammers rely on, whether the goal is a fraudulent purchase or a lookalike domain designed to steal login credentials.
Pressed on why Craigslist never adopted seller-reputation systems like eBay’s or Uber’s star ratings, Newmark says he anticipated, even in the platform’s early years, that any review system would be gamed — a suspicion he traces partly to reading science fiction (William Gibson’s Neuromancer) about how online media would be manipulated. His preferred fix for identity and trust online is plural rather than centralised: he explicitly does not want government to be the exclusive authenticator of who people are, citing his own political wariness of state control, and instead hopes credit-card networks, driver’s licences, and the digital wallets already on most Americans’ phones can combine into a workable authentication layer — a proposal he says he still needs to run past the Electronic Frontier Foundation.
Philanthropy: small, staged, and undone by size
Craig Newmark Philanthropies has no formal employees; Newmark works as a contractor himself, and credits his bookkeeper Mabel — a self-taught former refugee from Myanmar — with the practical infrastructure that makes this possible. Asked what larger foundations, with big staffs and multiple layers of sign-off, get wrong, Newmark is characteristically reluctant to generalise, but does offer a specific mechanism: once an organisation passes roughly the Dunbar number (about 150 people), one part stops knowing what another part is doing, sign-offs multiply, and responsiveness suffers even when the organisation is behaving ‘very responsibly’.
His clearest stated regret is two large gifts to Ivy League institutions, given as lump sums rather than staged, milestone-based multi-year funding — money that was not wasted, in his account, but not used as well as it could have been. This shapes his current grant-making: a repeated, half-joking threat throughout the interview to be ‘really annoying’ about how well grantees communicate their own work, on the view that good work poorly communicated is worth much less than it should be. His major giving areas — veterans and military families (Blue Star Families, the Bob Woodruff Foundation, a food pantry at Fort Wadsworth), cybersecurity resilience (Shadowserver’s malware-scanning honeypots, the Cyber Resilience Corps helping small water utilities protect their systems against ransomware and nation-state threats), and trustworthy journalism (the CUNY Journalism School, and Wikipedia) — each trace to a specific personal encounter rather than a top-down strategic plan: a chance seating next to a veterans’ advocate at a PBS luncheon, a photo-op at a food-packaging charity, and a long-standing worry about scam and cybersecurity threats to ordinary people and critical infrastructure.
On journalism specifically, Newmark distinguishes Wikipedia’s error-correction process from that of institutional print outlets: Wikipedia articles can be wrong but tend to get corrected over time, whereas he says the New York Times makes a good-faith effort to get things right but rarely revisits what it gets wrong. He is careful that this is a claim about correction mechanisms, not necessarily about relative error rates — a distinction the conversation does not fully resolve.
See also
- Craig Newmark — speaker page
- Tyler Cowen — host
- Jimmy Wales on Wikipedia, Neutrality, and the Governance of Knowledge — a second, more structural account of Wikipedia’s self-correcting trust model
- Jimmy Wales — Wikipedia’s co-founder, whom Newmark’s journalism philanthropy also supports