Most important take away

Founders and small seed funds own their destiny by minding monthly burn and choosing alignment over brand. Pro rata rights are framed as “the original sin against entrepreneurs” because they hand investors a free option, and the size of the fund relative to the check is the single most predictive signal a founder can use to know whether a partner will actually show up.

Summary

Actionable insights for founders, operators, and investors:

• Own your destiny by minding monthly burn. The biggest sin of the ZIRP era was capital stuffed down founders’ throats (Tiger/SoftBank style 10-on-40, 20-on-80 rounds inside two-week windows) that funded premature US expansion and inflated cap tables. The corrective discipline: track burn monthly so you never lose optionality.
• Take checks where the math actually works. Calculate what percentage of a fund your check represents. If a 10M check from a billion-dollar fund doesn’t put a partner on your board, that is a negative signal, not a positive one. A small fund putting in 7.5% of its capital and joining the board is a stronger commitment than a brand-name logo writing a tiny check.
• Pro rata is asymmetric against founders. You are giving the VC a free option to buy more of you later. When things go great, pro rata barely matters (Coupang made Sequoia waive theirs). When things struggle, pro rata becomes a stalking-horse problem: existing investors push you to “test the market,” signaling to new investors to price you down. Treat pro rata as something to negotiate hard, not give away.
• Watch for non-consensus opportunities in unloved areas. Frankel’s portfolio company Smalls (cat food DTC) hit $50M ARR while raising in a market that hated DTC, hated cats, and hated food. Non-consensus founders, non-consensus markets, and “orphaned” founders (those abandoned by mega-funds) are still where seed alpha lives.
• Second-time founders who failed honestly are better bets than first-timers with massive exits. The latter come with hubris; the former come hungry with team intact (e.g., Tom Leeser at Motorway after a travel loss).
• “Teams over themes” especially in AI. The $5-on-20 seed range is still investable; $25-on-100 AI seeds make seed-fund math impossible. The seed-fund discipline: ask “can we 10x from here?” (Eric Paley’s rule). If not, pass. Bessemer-style anti-portfolio memos prove you always underestimate the size of your winners, so the 10x heuristic still surfaces $20B outcomes.
• Take secondary at peak heat, not at trough fear. When valuations are screaming hot and 20M is left over in a round, take a third off the table. Heat does not correlate with deal quality; in fact, the hottest deals (5-on-125 seeds) are often the worst.
• Cut burn before it’s too late. Founders get stuck on a treadmill of loyalty to team and last money in, and refuse to reduce headcount even when the runway clearly demands it. Be the investor who tells the truth, even if the founder rolls their eyes.
• Take on incumbents (LEECHs: Lethargic Economic Extractors Causing Harm) with a war-on-air-land-sea playbook. The first move from incumbents is always “you’re illegal” (Suno on AI training, SeatGeek on Live Nation, PillPack on PBMs). You must budget for lobbyists, lawyers, and PR from day one. But you can only do this if customers already love the product.
• AI is short-term overhyped, long-term under-hyped. Sequoia’s $600B AI question is real; capex will not match earnings near-term. But discounting AI over a 10-year horizon (like the self-driving “last 5%”) is a mistake. Vertical SaaS incumbents who own enriched data and ship AI features defensively (Salesforce model) will be very hard to displace.
• DPI is dead for 2018+ vintages. ZIRP-era markups won’t convert; M&A and PE rollups (with 1x liq prefs) are how distributions are starting to flow. Permanent loss of capital is the real risk in 19-21 vintages, not just delay.
• Tech-pattern insight: vertical SaaS players sitting on enriched proprietary data are positioned to use AI defensively to retain customers and grow ARPU, while commoditized AI tools mean buyers won’t realistically build their own internal solutions (they can barely onboard Slack).
• Career insight: when judging a deal or a role, evaluate the entrepreneur/operator as red-button vs green-button. If you wouldn’t take their 7pm call mid-dinner, don’t take the deal. “Fall in love with the who, not the what.” Patience and pain tolerance are the two operator superpowers; building anything important is “jumping out of a plane with a box of silkworms and hoping they’re overachievers.”
• Best matchmaking signal: small funds are economically incentivized to introduce you onward; large lifetime-capital funds are economically incentivized to keep the pro-rata option open. Ask your seed partner which specific GP at the next firm they will hand you to.
• Personal lesson reinforced: never sell a share in a true mode company (Trade Desk, Uber). Optimal liquidation rule of thumb: distribute shares (not cash) on large positions when they exceed a meaningful fraction of the fund; sell and distribute cash on smaller positions. But for true moats, never sell.

Chapter Summaries

• Seed in a mega-round world: Traditional 50-100M seed funds can still win by being right-non-consensus on founders (less traditional backgrounds, failed/orphaned founders) and on markets (Smalls cat food at $50M ARR, restaurants via Aloe). Walk away from $25-on-100 AI seeds; the math doesn’t work.
• Breaking the rules: Frankel broke his own rules once for a $100k check because of conviction in the founder. Counter-example: Pinterest was passed due to a conflict with portfolio company Vimeo (via Zach Klein), and the rules of loyalty are unbreakable even at the cost of returns.
• Reading companies fast: You usually know within three months for consumer; enterprise sales cycles obscure judgment longer. Olo looked dead pre-Founder Collective; Noah Glass kept pulling rabbits out of hats.
• Reserves strategy: Fund 1 had zero reserves; Eric Paley broke that for Trade Desk and saved the position. Fund 2 onward uses a 1:1 reserve policy with rules (no reserves above 20M post initially) that the market kept rendering obsolete. COVID-era panic led to over-reserving just as capital flooded in.
• Hard conversations: Telling founders the truth is a duty, not a courtesy. Running Tide example: $3M/month burn on $10M cash with the founder refusing to cut. Founders get trapped by loyalty to teams and last-money-in optimism.
• Founder psychology: Second-time founders who failed and brought back their team (Motorway) beat first-time massive-exit founders with hubris.
• Pro rata = original sin: It’s a free option investors extract from founders that hurts most when companies struggle. Coupang refused Sequoia’s pro rata when raising at $4B with BlackRock; ordinary companies get used as stalking horses.
• DPI is dead for 2018+ vintages: Fund 1 and Fund 2 distributed early via mid-size exits (Cruise, DataLogix, InfoScout) before Uber/Trade Desk/Coupang. Fund 3 (2018 vintage) has zero DPI; first distributions now coming via PE rollups with 1x liq prefs.
• IPO and M&A markets: One Stripe, Starlink, or SpaceX IPO post-election could re-open the market. M&A and PE consolidation become the new liquidity path.
• LP pressure: Endowments dialed back; family offices pulled back more. Some LPs will be too scared to skip vintages (Excel Fund 7 / Facebook fear). Yale/Swensen 30-40% VC allocation model assumed liquidity and selection hit-rates that no longer exist.
• Fund-returners and fund size: A $75M fund can be returned by PillPack alone; the same outcome doesn’t move a $400M index seed fund. Founders should run this math when picking investors.
• Secondaries: Highly elusive outside top-tier names; pre-IPO only. Take secondary at peak heat. Never sell shares in true moat companies (Trade Desk, Uber regrets).
• Leeches (Lethargic Economic Extractors Causing Harm): PBMs, Live Nation/Ticketmaster, OTAs. The incumbent playbook is “you’re illegal,” then lobbyists, lawyers, PR. Founders need war-budget capital plus customers who actually love them.
• AI: Short-term capex won’t match earnings (Sequoia $600B question). Long-term, AI is one of those “slowly then suddenly” waves like self-driving. Open AI at $100B is probably still a 2x bet. Discipline keeps Frankel out of $25-on-100 AI seeds.
• Vertical SaaS + AI: Data-rich incumbents like Salesforce will use AI defensively to keep customers; buyers can’t realistically build their own. Short-term: margin pressure; long-term: ARPU growth.
• Downside / loss ratio: The most unfair feature of capitalism is asymmetric upside (lose 1x, win 3000x). If you’re not losing, you’re not taking enough risk. Bad losses come from falling in love with the what, not the who.
• Heat ≠ quality: 5-on-125 seeds were the worst-performing cohort. Take secondary at heat peaks when offered.
• Board membership: Real economic alignment requires meaningful ownership (rule of thumb: not under ~15% as a board member). Dilution alongside founders is acceptable for early-stage funds.
• Matchmaking: Small funds matchmaker by necessity; large funds keep pro-rata optionality. Founders should demand to know which specific GP at the next firm their seed investor will hand them to.
• TAM mistakes: Bessemer passed on MediaRadar because newspapers and magazines looked dead; the company became billion-dollar by following customers into Facebook, Google, Netflix, Apple. Great founders find markets.
• Quick-fire: Patience as best advice. PE firms now seen as valuable cap-table partners, not enemies. Jake Saper at Emergence as model board member. Frankel’s biggest loss: Running Tide. Reading The Path Between the Seas (David McCullough on the Panama Canal) for moonshot perspective.
• Closing reflection: Conviction decisions are usually 51/49, not binary. Love, friendship, and loyalty matter most.