Most important take away

Seed investing is fundamentally a human, relationship-driven game where most outcomes are decided by power-law winners — so price discipline, group voting, and “value-add” theater all become liabilities. The actionable thesis from both Tisch and Rohan: build long, authentic relationships with founders before the transactional moment, let individual conviction (not committee consensus) drive yes decisions, and never pass on a likely outlier because the entry price feels high.

Summary

Actionable insights and career/operating advice from the episode:

• Pick a style and be consistent. Tisch’s biggest regret was spending years trying to be a type of investor he wasn’t. There is no single “right” way to invest — find your authentic mode and compound on it.
• Stop trying to “prove value” in the pitch. The pitch window is a fake, compressed timeframe. Real value is delivered over a decade-long relationship of ups and downs. Optimize to be the founder’s “favorite” (the friend in the room), not their “best” (a checklist of services rendered).
• Meet founders before the transactional moment. The earliest, highest-leverage relationships are formed before someone is raising — at university, at their current job, while they’re still mulling an idea. That’s where seed investing is actually won.
• Treat seed picking as gut/conviction, not consensus. Both guests argue group voting at seed produces consensus-safe picks that miss outliers. Box Group has nine investors and any one of them can say yes; they don’t vote. Founders investing through Otherwise operate the same way — distributed individual conviction.
• The cost of omission >> the cost of commission at seed. Saying no to a generational winner blows your returns; saying yes to a dud is a rounding error. Find the one reason to say yes, not the 32 reasons to say no.
• Price doesn’t matter at seed; conviction does. If it’s the right company, today’s “egregious” price is the lowest you’ll ever pay. Price discipline starts to matter at Series B, where the game shifts from art to math.
• Reserves and follow-ons may actually hurt seed performance (Rohan’s view). They depress DPI, suffer from adverse selection at Series A pro-rata, and complicate the founder relationship. Tisch counters that if you have reserve capital you must be great at follow-ons, but agrees over-reserving is the bigger mistake than under-reserving.
• Don’t sell secondaries early. The compounding from the late stage (e.g., $5B to $10B) is often more valuable and more predictable than early compounding. Selling 20% at $10B can make sense; wholesale exits at Series B are a return-profile mistake.
• Signaling is largely a myth (per Tisch). Multi-stage funds don’t need to do every Series A from their seeds. Real “signaling” only shows up when a deal goes stale in market — at which point it’s already a failed fundraise, not a signaling problem. Rohan partially disagrees: the question still gets asked, and arming founders with a clear answer about insider posture is part of the job.
• Brand is the most under-discussed, most potent currency in venture. It compounds slowly, is durable once built, but can degrade. It’s the actual reason multi-stage funds can win seeds — engineers being recruited care that “Sequoia did it.”
• Founders make great seed investors because See/Pick/Win all skew in their favor. They see deals via natural founder community, they win without competing, and they pick well because they invest in their own networks/domains rather than chasing demo-day heat.
• Cold inbound works. Both take cold email if it’s human, personable, and shows attention. Many deals they regret missing were sitting in their cold inbox.
• Uncapped SAFEs are fine. Tisch uses them with conversion provisions. The job is to invest in the best companies through whatever mechanism, not to enforce rules for their own sake.
• Founders’ best move: get to capital-optionality. Rohan’s advice to his founders is to raise initial capital, then become so capital-efficient (especially with AI tooling) that venture becomes optional. That preserves dilution, board control, and growth pacing.
• For new fund managers raising LP capital: expect a totally different pace. 12–18 months is normal. Hundreds of nos are normal. The front-office speed you’re used to does not apply.
• Have a vision for why you’re starting a fund. Don’t do it because it looks easy — the ROI timeline is brutal. Define why “1 + 1 = 1” or “1 + 1 = 3” for your firm.
• The job of an investor is to invest. Say yes or no quickly and transparently, give the money, get out of the way, and do what founders ask when they ask. Don’t waste their time.
• Outside the top ~25 individual VCs, investors are commodities at best and harmful at worst. Founders overweight the willingness to lead as a positive signal — the median lead is neutral.
• Don’t believe in “coaching” as a category. Relationships, mentors-as-friends, and selective wisdom transfer work. Hierarchical “coach/player” framing applied to investor-founder relationships is wrong.
• The success of a company comes from employees 1–100 and the founders. Outsiders amplify what’s there; they don’t create it. Even Sequoia’s HubSpot impact came after HubSpot was already working.
• Operate in a state of panic (productive). Rohan: the best investors are intensely hungry and swarm. Treat every check like the one that will determine whether your firm survives the power law.
• Stay relevant or die. The next network of great founders is somewhere — find it. AI eating the world means you have to compete with experts in spaces you don’t natively know.
• Seed in 10 years looks largely the same. AI won’t replace it because seed is about anomalies, not pattern recognition. The human-in-the-room becomes more important, not less, as the rest of the market gets more efficient.

Tech patterns mentioned: shift from physical/in-person pitching to fully remote/Zoom-based deal velocity post-COVID, multi-stage funds collapsing decision timelines, AI making startups dramatically more capital efficient and venture-optional, secondaries as a compounding alternative liquidity path as M&A markets contract, registered VC funds increasingly able to buy common stock.

Chapter Summaries

• Intros and firm models: Terrence Rohan runs Otherwise, a discreet fund that gives capital to top founders to make investments primarily at seed (he writes 250K checks personally). David Tisch runs Box Group, a $212M pre-seed/seed fund (six funds in, 15 years), 500K–1M checks, collaborative non-board-seat model, with a paired follow-on fund (Box Group Picks).
• Why founders pick you: Both reject “value-add” theater. The pitch is too short to prove value; the real product is a decade-long human relationship. Goal is to be the founder’s “favorite,” not their “best.”
• State of the seed market: Rohan calls it a “jump ball” — fund expansion/fragmentation, generational change at storied firms, founder optionality at all-time highs, plus the recent AI disruption. Tisch adds that post-COVID accessibility to multi-stage firms has permanently equalized speed.
• Multi-stage funds at seed and price: $5M seeds existed before; what’s new is more accessibility. Both argue you can’t be valuation-sensitive at seed — if it’s the right company, any price was the lowest. Price matters at Series B, not before.
• Conviction, luck, and humility: Tisch pushes back on the word “conviction” — most picks will be wrong, and seed investors aren’t 10x better than each other at picking. A lot of VC Twitter is content marketing. The hard work is done by founders and first-10 employees.
• See/Pick/Win: Seeing is the actual skill — meet founders before the transactional moment. Picking can’t really be optimized because feedback takes 5–10 years. Winning follows from genuine relationships.
• Decision making and group think: Both reject voting/consensus at seed. Box Group’s nine investors can each say yes individually. Negative grenades and consensus crush fragile early ideas. Distribute decision-making to whoever spent the most time on the deal.
• Founders as investors: Rohan’s Otherwise model — founders see and win at a standard deviation above professional VCs because of community access and reputation; they also pick well by staying in their networks/domains.
• Reserves and follow-ons: Rohan argues no-reserves outperforms — depresses DPI, suffers adverse selection at A pro-rata, and complicates relationships. Tisch defends reserves if you’re great at them but agrees over-reserving is the bigger error than under-reserving.
• Secondaries: Both largely don’t sell. Late-stage compounding (e.g., 5x→10x) is the easiest math in the business. Selling tiny slices at $10B is fine; wholesale exits at Series B are wrong.
• Staying relevant: The daily panic is “are we seeing things and understanding them” — networks and brand compound but require constant refresh as software gives way to AI, science, hardware.
• AI’s impact on seed: First-money-in rounds aren’t going away. AI makes companies more capital efficient and venture-optional after the first round, which is a good outcome for founders.
• Quickfire — best advice: Tisch: “Most investors invest” — your job is to say yes or no quickly and get out of the way. Rohan: the best founders need you the least, so underwrite truly special people.
• Quickfire — advice for new managers raising: Rohan: brace for the LP world’s glacial pace (12–18 months, hundreds of nos). Tisch: have a clear vision for why you’re starting the fund.
• Worst no / hardest mistake: Both: the misses you regret most are the ones where you weren’t true to your own investing style. Specific names withheld.
• LP picks across stages: Tisch refuses to name names (“collaborative”); the bigger point is that top firms stay top firms via power law and brand, with second-tier VCs being uninteresting and third-tier being harmful. Brand is the most potent under-discussed currency in venture.
• Signaling debate: Tisch declares signaling a fake word — no one expects multi-stage funds to do every Series A from their seeds. Rohan partially disagrees: the question still gets asked, especially when a deal goes stale, and arming founders with a clear insider posture is part of the job.
• Other BS to debunk: Most VCs aren’t great; the average VC actively wastes founders’ time. Forwarding TechCrunch articles is not value-add. Cold emails are accepted and good ones are a form of hustle. Decks are fine. Uncapped SAFEs are fine.
• Misalignments: VCs sell LPs a product (20% ownership, board seats, reserves) that doesn’t always match what founders want to buy. The misalignment is biggest when the VC is below-average — average VCs are worse than absent ones.
• Coaching and value creation: Both reject “coaching” framing. Outside the top ~25 individual VCs, investors are commodities. Company success comes from founders and the first 100 employees. Sequoia-tier brand can move needles via talent and customer access, but only after PMF.
• Seed in 2034: Both expect it to look the same — more funds, more fragmentation, more capital, but the core remains a human relationship business. AI won’t replace it because seed is about anomalies, not patterns. The human in the room becomes more, not less, valuable.