Turning Peter Thiel's $100K into $10M Angel Portfolio & Why VCs Can Be Sharks | Josh Browder

Summary

• Josh Browder runs a one-man accelerator: pre-seed checks at sub-$5M valuations (fund four: 33 deals, median entry $5M, range $1.5M–$21M, typically 5–7%), with founders sleeping in his Four Seasons-residence spare room on $50-a-night beds — “it’s like Hotel California, where you can’t check out until you’ve raised your institutional seed.” His answer to adverse selection at low prices: his best investments were all founders he lived with.
• The track record behind the model: he turned his Thiel Fellowship $100K into what “will be in eight figures” on paper, and Micro 1 — the staffing founder he made reincorporate in Delaware, move into his spare bedroom, and pivot to software — is “well over a thousand X.” His conclusion: “it 100% pays to be early,” so fund four holds no reserves — a 15% reserve into Owner’s Series A was good, but its opportunity cost was 20–30 pre-seeds.
• The macro reversal: “I thought 12 months ago it was a bubble. Now I don’t think it’s big enough… I think Anthropic will get to a trillion in revenue. We’re not in a bubble.” He mentions an investment in the AI-infrastructure stack; separately, his Thiel Fellowship roommate founded Fluidstack.
• But the wealth distribution is combustible: “For every Anthropic employee who’s making 20 to 100 million, there’s 7,000 Block employees being laid off… You can’t have 50,000 people with all the money. I think actually there could be a revolution in our lifetime.” Absolute goods hold price; positional goods — “only eight seats in Delta first class” — go off the charts.
• His personal allocation: all of his money goes into Nevada land — no stocks, no bonds, no cash, because “the dollar doesn’t have a good future.” Returns are “only like 10 to 20%” but safe, Nevada uniquely combines no income tax, low property tax, and rising population, and land hedges both AI outcomes: post-economic abundance where land is the last scarcity, or tech going to zero.
• “VCs are sharks”: pitching is poker — never reveal your price; take a kingmaker (Founders Fund, Sequoia) at half the valuation of anyone else; dilution sensitivity is “nonsense” (“5% of infinity, you should still do it”); prefer SAFEs because B-minus VCs push priced rounds to manufacture markups for their own fundraise; and never sign on the spot.
• Selection doctrine: never-give-up beats credentials — the VC “race right now to back the math olympiads” over-indexes IQ. Meanwhile fake founders are reverse-engineering his stated criteria with Claude and ChatGPT (“ideological fraud”), so he screens with 11 p.m. meetings and “let’s look up your Stripe right now.”
• The framing lesson from DoNotPay’s near-death seed raise: three minor deck changes — a demo, Intuit/Honey/Credit Karma logos, subscription instead of advertising — flipped uniform Sand Hill rejection into an on-the-spot investment. “Nothing changed about the company… but the most minor differences in framing and strategy made all the difference.”

Deep dive

Browder Runs a One-Man Accelerator

• The game Browder plays is day-one entry: sub-$5M valuations, unpolished first-time founders, and a crash course transferring “10 years of DoNotPay mistakes… in a matter of 3 weeks.” Fund four so far: 33 deals, median entry $5M, minimum $1.5M, maximum $21M, typically 5–7% ownership. Founders live in his spare room — “technically a Four Seasons residence,” but four co-founders can mean four $50-a-night beds in one room: “Hotel California, where you can’t check out until you’ve raised your institutional seed.”
• His answer to the adverse-selection question he knows you’re asking: the best founders he ever backed, he lived with at some point — roommates with Assured’s founder in the house where Facebook started, Micro 1’s founder in the spare bedroom. And the constraint is deliberate: when an LP suggested renting a hotel to run ten at a time, he refused — “that removes the artificial constraint of one.” One partner, one company, no brand dilution.
• What he fixes inside: pre-seed companies die three ways — “they run out of money, they run out of hope, and co-founder disputes.” Money means pitch training; hope means daily progress and ignoring “vanity signals of being a young founder in SF”; disputes he pre-empts by recruiting his own friends in and handling vesting. He has never lost belief in a house guest — the opposite: “I’ve accosted them at 2:00 a.m. … I want to put in another 300K.”
• Portfolio construction changed on cold math: he put 15% of fund three into Owner’s Series A (“normally I would never go above 7%”), and says fund three will be very good because of that reserve investment and the pre-seeds; its opportunity cost could have been “20 to 30 pre-seeds” — so fund four runs no reserves, everything up front. Once a top-tier firm commits, “it’s like a stampede… where were all these people just 3 weeks ago?”

Browder Screens for Fake Founders

• The backdrop: a Stanford Review piece claimed it’s now easier for a Stanford student to get into YC than to get a job, breeding “fake founders” with no connection to their problem. He deliberately scales back investing in summer — “not because I’m in Capri like all the other VCs” — because during the summer he can’t tell who has actually dropped out.
• His screen runs like “a visa interview”: propose meeting at 11 p.m. (the best say sure; the mediocre offer “Tuesday week”), then validate every claim on the spot — “I’m at 5,000 in revenue”? “Let’s look up your Stripe right now.” “What serious entrepreneur doesn’t have the Stripe app on their phone?” A D-minus goal is “a partnership with Anthropic”; an A+ is “I’m going to fly to Milwaukee to meet with a dentist to get them to sign my $500 a month SaaS plan.”
• The arms race worries him most: founders now use Claude and ChatGPT deep research to reverse-engineer his stated criteria — “in 3 weeks, someone is going to say, ‘We were best friends in high school.’” Browder says that, after publicly citing family trauma, gaming, and early entrepreneurial success as signals, “nine out of 10 young founders that we meet pitch those three exactly.” The label they land on: “ideological fraud” — “it’s not illegal to say you’ve had childhood trauma when you’ve actually grown up middle class.”
• What he actually wants is founders who are their own first customer: Adam Guild built Owner’s first product for his mother’s dog-grooming business; Browder is “the type of person to get 10 parking tickets just to test out the service.” “If you’re building for yourself, at least you have one customer.” Illegal fraud, by contrast, is getting harder — “Twitter sleuths will detect anything,” and single articles have killed companies.

Persistence Beats Credentials

• Ali Ansari arrived as a solo founder in Los Angeles running a staffing business inside a California LLC — “uninvestable.” Browder set three conditions: reincorporate in Delaware, move into his spare bedroom in the Bay Area, and build a software-style product. Result: his best multiple ever, “well over a thousand X” — and Ansari “still lives in my building to this day,” working past midnight. Quickfire pick for most underrated CEO: Ansari — “watch this space in the next 12 to 18 months.”
• The generalizable rule: “If you back someone who’s above average IQ… and never give up, of course they’ll succeed.” He thinks VCs are over-indexing credentials — “there’s a race right now to back the math olympiads. VCs are going to spelling bees” — when persistence is the scarcer input. He endorses Ken Griffin’s athletes-with-above-average-IQ frame: athletics is one form of never giving up; “a huge chip on your shoulder is another.”
• His biggest self-diagnosed investing mistake: letting his own imagination write the roadmap. “I never tell the entrepreneurs what to build… unless it comes from them, it’s not their life’s work.” Hard no-go zones: crypto (“best left to the crypto funds”), consumer hardware, and wet science.
• And he runs toward competition: “all of my best investments have actually been in very competitive markets” — Owner dominates crowded restaurant tech; large labs throw data-labeling companies “100 million, even close to a billion-dollar contracts.” The corollary is his jargon filter: “if you go to a pub and say I’m building observability for AI agents, they’ll laugh at you” — but software automating health-insurance claims makes sense at the pub.

Framing Wins DoNotPay Funding

• The formative story: three years into DoNotPay, millions of free users, “a slam dunk pitch” — and Sand Hill said no, one after another, a week after he’d dropped out. Two or three pitches from quitting (“do something really depressing like go work for big tech”), he rehearsed for his outside counsel, Wilson Sonsini partner Damien Weiss, who interrupted halfway: “You’re doing it completely wrong.”
• Weiss’s three fixes: a live demo (“they’re not investing in the PDF deck… the fact that you’re not doing a demo is criminal”) — so he scrambled together a robot appealing bank fees; aspiration logos — Intuit ($200B; “we want to be the TurboTax of consumer rights”), Honey (just acquired for $6B), Credit Karma ($8B); and swapping advertising for subscription in the Cambridge Analytica era. He didn’t believe the last one — it’s now DoNotPay’s main business model.
• The next pitch was “night and day”: the firm wanted to invest on the spot, and the herd “started rescinding their rejections” — which he found “a bit depressing.” The lesson he now drills into founders: “Nothing changed about the company, nothing changed about me… but the most minor differences in framing and strategy made all the difference.”
• On calibrating ambition: “all of the best founders I’ve invested in have had delusional levels of ambition. The more delusional, the better” — though the current state must be described accurately (“I’m not like one of these hype Americans”). He nearly sold DoNotPay for a million dollars early on; the people he asked for advice laughed at the offer but were right — “you can do better than this.”

Founders Should Outsmart VCs

• His frame on VC value-add: “There are three types of people. Those who make it happen, those who watch it happen, and those who wonder what happened” — at best, VCs are the second; they can add value at strategic points. So play poker: “You should never reveal too much information about what you’re seeking,” above all price — “the price is a function of how hot the deal is, which ironically is less hot if you go in swinging for the fences.” Always demo, the CEO pitches alone, and fly in person over Zoom.
• Kingmaker math: take Founders Fund or Sequoia at X over another firm at 2X — “you’re king made and it will save you in the next round.” But founders’ kingmaker lists run too long: “No tier two firm thinks they’re tier two.” Harry adds the revenue angle he sees at Harvey and Legora: kingmaker investors deliver mega law firms as customers — the brand is an acquisition channel.
• Dilution sensitivity is “nonsense”: “either it succeeds or it doesn’t. If it succeeds, at a minimum you’re worth hundreds of millions… If it fails, you’re nothing” — if extra money cuts failure odds even 5%, “5% of infinity, you should still do it.” Structure matters more than points: B-minus VCs push priced rounds to manufacture the markup for their next fundraise, but “SAFEs don’t dilute other SAFEs” — and dodge the 15%-option-pool-at-seed shark move.
• The shark warnings: “The VCs will say anything to get you to sign right there and then” — golf-buddy customer intros that never materialize — so never sign on the spot (he gives his founders overnight, answer by morning; “if they’re running a process, it’s too late”). Same logic on founder secondaries: “the people buying these secondaries are sharks” — if you’re bombarded with offers, “perhaps they have more experience with the market than even you do.” Friends sold, then watched valuations 3x in weeks.

The Fellowship Shaped Browder

• The origin breakfast: at 18 or 19, on the verge of making DoNotPay a nonprofit, Marc Andreessen reached out on X and — over breakfast in Atherton, in “breakfast clothes” — convinced him “the biggest organizations are for-profit entities… you can have 10 times the impact of a for-profit company because the incentives are aligned” (he points to OpenAI’s nonprofit-to-for-profit switch and “the $150 billion lawsuit”). Andreessen is his first investor at DoNotPay, first institutional investor in his fund, and “the most curious of the luminaries.”
• The dropout wasn’t the Fellowship: Stanford’s creative-expression requirement meant a 9 a.m. social dance class — “I’m more of a kind of Claude Code style person. I love the code” — and he chose keeping DoNotPay alive for millions of users, failed the class, and dropped out; the Thiel Fellowship came weeks later. His caveat for today: “dropping out is almost the establishment” now, and doing it for its own sake is “definitely wrong” — college still gives you recruits, a .edu email, and a free pass for mistakes.
• The Fellowship’s real product is 19 peers with the same problems; his assigned roommate, a fellow Brit, founded Fluidstack — “publicly reported to be a tens of billions company.” Selection has cycled: from out-of-distribution individuals (“Vitalik was just a crazy individual doing crypto, which was not exciting at the time”) through a late-2010s flood of externally validated applicants, back to the individual. The fellows’ saying: they have an “expiry date” — his own version was being “terrified of being Macaulay Culkin.”
• He put his first $100K installment straight into Adam Guild and other fellows he met while helping interview candidates: “when it’s all said and done it will be in eight figures… it 100% pays to be early.” And “unfortunately, I’ll make more money from the investing” than from DoNotPay — an edge he credits to being a founder himself: “founders love to work with other founders. They hate professional money managers,” citing Lockie Groom and Dalian.

DoNotPay Runs Lean

• DoNotPay is “in some ways a media business”: 90%-plus of customers arrive organically through SEO, viral stories, and referrals — hundreds of thousands of customers, fully automated, 11 employees, “more money than we’ve raised,” with dividends now going quarterly. He raised $22M total and skipped the 2021 mega-round; a competitor spending “$300 to acquire a customer worth $150” sold at the peak — “good for them. I can’t play these games.”
• Burn as a character test: “founders who burn all the money — that’s not cool. It’s kind of lame. Why didn’t they just cut the burn?” Black-swan failures excepted. To Harry’s needle — you won’t dividend your way to a fund-returner — he concedes nothing permanent: “we are going to take some big swings this year,” rolling up consumer businesses into DoNotPay.
• On the SEO golden goose: Harry cites monday.com’s CEO reporting SEO down 15% — $100M-plus at that scale — but Browder counters that GEO, “the AI version of SEO,” is rising in parallel, and “the world doesn’t move at the pace of San Francisco”: his customers are in middle America and the UK, where Google is still huge. The product thesis: “DoNotPay is an ETF on the world’s problems… fortunately for us, the general trend of problems is up” — bill negotiation is the biggest unsolved rip-off, and in-flight Wi-Fi refunds run a 100% success rate.
• Hiring doctrine: “business school is actually a counter signal” (Harry: “I agree 100%”), strategy hires are meaningless, and the tell is employees who automate themselves — one bought prepaid Visa gift cards at Target so free trials never touched his real card. Solve bottlenecks instantly: his own design “would give our early users headaches,” so hire number one was a designer. His role-of-the-future call: custom evals — “the future of AI will actually be organization-specific,” evals run on your own data.

AI Reshapes Venture Markets

• The changed mind of the last 12 months: “I thought 12 months ago it was a bubble. Now I don’t think it’s big enough… I think Anthropic will get to a trillion in revenue. We’re not in a bubble.” Pressed by Harry on why he doesn’t then buy Nvidia, AMD, Nebius [likely] and CoreWeave, he says he has an investment in the AI-infrastructure stack; separately, his Thiel Fellowship roommate founded Fluidstack.
• The concentration barbell: to Harry’s scenario of eight companies taking “$5 [trillion] plus” of market cap, he predicts “a rise of medium-size businesses, almost DoNotPays, that fill the niche” plus the giants — the plain-large middle “have to be very worried.” The human ledger: “For every Anthropic employee who’s making 20 to 100 million, there’s 7,000 Block employees being laid off.” His decomposition: Meta’s layoffs are AI-driven (“they need fewer engineers”); Block’s are COVID over-hiring.
• San Francisco’s split-screen — 600 OpenAI employees cashing out an average $11M the same morning big tech cuts — gets his absolute-vs-positional goods frame: standard apartments and food hold price, but “there’s only eight seats in Delta first class… only eight houses on the best road in San Francisco. The positional goods will just go off the charts. If you value your success in life by attaining positional goods… good luck.”
• Harry’s structural gripe — Series A is the worst place in the market, “200x ARR if you’re a million in revenue” with little PMF — gets a response: Josh says “there’s this illusion that the valuations are lower than Series A, so it’s less risk… sometimes the best place to be could actually be investing at a billion.” Harry then adds that at that stage the investment is de-risked and cash flows can be modeled. On constrained exits, secondaries are the growing channel, and small funds have the edge: “if Menlo Ventures sells secondary in a company, the company is dead” from signaling — a friendly pre-seed selling a sliver isn’t. His own rule: “the things that are most attractive to the secondary buyers, you never want to sell” — though he’s finally selling some now.

Browder Buys Nevada Land

• His entire personal allocation: “I take all the money I make and I buy land. I don’t put it in the stock market… I don’t keep it in cash because I think that the dollar doesn’t have a good future. I don’t buy bonds.” Inflation “is just going crazy, and it’s just going to get worse… the only way to stay ahead in this AI world is to have real assets.” The Churchill anchor: “Land is the only scarce resource.”
• The land is a two-outcome hedge on AI: either AI creates “a post-economic world where it replaces all big companies and the only thing that’s scarce that’s left is land,” or “maybe it’s all a bubble and all of tech goes to zero, but land will still be valuable.” He buys in Nevada, “far away from the tech bubble” — nail salons, passive, set-and-forget — because three things hold there simultaneously: no state income tax, very low property tax, rising population — “not true, in my opinion, anywhere else in the US” (Florida fails on property tax). Returns: “only like 10 to 20%” but very safe — Harry: “that’s better than I thought” — plus depreciation.
• The tail risk he genuinely fears: “It’s not sustainable. You can’t have 50,000 people with all the money. I think actually there could be a revolution in our lifetime. Something has to change.” Harry calls it his biggest worry too, citing 33% of UK children growing up in poverty. Browder’s out is job creation — AI data cleaning didn’t exist five years ago; air conditioning in data centers is coming — but “the government will have to get a lot better at helping people transition.”
• On politics he doesn’t hedge: whatever your view of the administration, “it’s objective that for tech, the current administration is a lot better.” And: “Lina Khan, in my view, is evil” — several portfolio companies had acquisitions blocked, and in one case a blocked pharma acquisition meant the drug never got developed: “to this day, 50 people die a year because of the acquisition being blocked.”

Browder Explains His Roots

• The origin gag: at 14 he built Pret A Manger’s iPhone app unofficially — “ripped off all the graphics” — and Pret weighed legal action until realizing that suing a 14-year-old would be terrible PR, so they invited him in and made it official. The lesson that stuck: “it’s best to ask for forgiveness versus permission.” (His mother insisted on chaperoning the CEO meeting — “I don’t want you meeting strangers on your own.”)
• The UK was also the perfect first market: “repressed… it’s just fines fintopia” — average speed cameras strike his American friends as a surveillance state. DoNotPay’s actual unlock was UK-born: a Freedom of Information Act request for the top 12 reasons parking tickets get canceled, then a chatbot so consumers “could write whatever gibberish they wanted” and get matched to the correct defense — his answer to why chat UI works for him (and is “terrible” for travel).
• The geography trade: the UK offers big-fish-in-small-pond advantages — less competition for talent, easier media (“anyone can get on the front page of the Daily Mail”) — but “the scale of ambition in the US is a hundred times bigger”: UK’s best outcomes are tens of billions (Revolut), the US is trillions. To Harry’s SF-is-the-worst-place-to-start case he half-concedes an arbitrage — leadership in SF, talent global, like Deel — but defends SF’s serendipity: “In London, the rich keep to themselves in their fancy clubs. In San Francisco, you can play pickleball with anyone.” Europe’s one fix: “get rid of these ridiculous regulations” — he abandoned a German investment over bureaucracy, and “how can you charge a sales tax [VAT] on investments?”
• The spine of it all: mid-poker-game at Stanford, a news alert told him his father — a human-rights activist and “enemy of the Russians” — had just been arrested. He cashed out his chips, called his mother, and as a teenager tried anything to block extradition, down to emailing the Consul General. The legacy is “paranoid and fearless”: “They’ve come to our house… knocked on the door” — so an upset bureaucrat “is all noise. It doesn’t really matter.” The paranoia is the fuel: “If you’re not motivated by the fear of losing, I think you’re asleep at the wheel.”