Garry Tan - Unwrapping the Gift - [Invest Like the Best, EP. 267]

Episode Summary

Executive Summary: Patrick O’Shaughnessy interviews Initialized Capital’s Gary Tan on venture investing as a system, arguing software can turn investing into a faster, more collaborative “local brain” that spots overlooked founders and ideas. Tan emphasizes trust, conviction, and founder quality over pedigree, then discusses how media, structure, and new tools expand access to capital and company-building.

Main Topics: Software as a “local brain” (Priority: 5/5): Tan argues organizations are reorganizing around software-driven local brains, not just global platforms. Venture investing as a system (Priority: 5/5): He describes YC and Initialized as process-heavy, high-throughput decision machines. What great startups look like (Priority: 5/5): He looks for clear inventions, strong builders, and markets made possible by a real why-now. Trust over capital scarcity (Priority: 4/5): He argues the real bottleneck is trust and network credibility, not money. Media as access infrastructure (Priority: 3/5): Media and essays help reveal opportunity to people outside elite networks and clubs. Founder and firm integration (Priority: 4/5): He says better decisions come from self-awareness, emotional integration, and diverse partner perspectives.

Key Arguments: Software is reorganizing every firm into a local brain; spreadsheets were only the first step. Too much capital exists; the scarce resource is trust, not money. Great investing is about seeing the invention, not fetishizing pedigree or pitch polish. Speed matters in venture because delays kill deals and reduce access to the best founders. Diverse partner expertise improves decisions by combining different lenses on the same company. Markets open when enabling technology changes remove old walls, as with Instacart and Coinbase.

Data Points: Y Combinator acceptance rate: 2% acceptance rate or less - Tan cites YC’s batch selection process as highly selective and software-enabled. YC application volume: tens of thousands of applications - He describes the scale of founder applications relative to available spots. YC batch cadence: twice a year - Used to illustrate the assembly-line style structure of the program. Initialized outcomes: hundreds of companies and 25 unicorns - Tan summarizes Initialized’s investing track record. Partner count at Initialized: six total partners - He uses this to explain the firm’s hive-mind decision process. Decision turnaround: 48 hours or less - He says all-partner meetings must be scheduled quickly to avoid losing deals. Decision window: 48 to 72 hours - Initialized aims to make investment decisions within this time frame. Portfolio/ownership reference: 10% to 20% - Tan describes the firm’s ownership in a company after investing. Posterous scale: 100 million or 200 million monthly actives - He cites Posterous as a key failed/partial success from his founder experience. Formic scale: $100 million a year net - Tan says the robotics startup could reach this level with only a few thousand machines. Formic deployment: 20 up now - He notes the number of machines deployed at the time of the interview. BamSec price point: $30 a month - Steve White describes BamSec as an obvious bargain for daily research use.

Pivotal Quotes: "the water is fine" — Gary Tan: He reassures prospective founders that there are plenty of solvable problems and opportunities. "If you want to go fast, go alone. And then if you want to go far, you got to go together." — Gary Tan: He explains why Initialized is built as a team-based decision engine rather than a solo investor model. "We want to attract founders who then want to attract the engineers and designers and product people to go and make the self-fulfilling prophecy of a startup actually work." — Gary Tan: He describes Initialized as a shelling point for talent and a brand that compounds.

Implications: The unanswered challenge is scaling high-trust, high-speed capital allocation without losing judgment, so founders and investors should build credibility, not just tools.