Episode Summary
Executive Summary: Miles Grimshaw argues that great investing starts with identifying change, dislocation, and authentic founder-market fit—not with rigid frameworks or data worship. He reflects on lessons from Thrive and Benchmark, his misses in Figma/Plaid, and why AI will reshape software from copilots to systems that sell work and outcomes rather than software seats.
Main Topics: Miles’ path into venture (Priority: 5/5): He traces his interest back to growing up around his entrepreneur stepfather and then being drawn into New York’s early 2010s startup scene, which led him to Thrive Capital. What makes a great investor and firm culture (Priority: 5/5): Miles emphasizes kindness plus competitiveness, belief in talent, patience with urgency, authenticity, and Benchmark’s free-form partnership style built on trust rather than process. How Benchmark evaluates companies (Priority: 5/5): He describes a five-part lens—sourcing, selecting, winning/signing, supporting, and summiting—and says the best diligence feels like planning an expedition, not an examination. Founder quality, learning speed, and decision-making (Priority: 5/5): He argues that will often matters more than skill, the best entrepreneurs make new mistakes as they learn, and VCs add value mainly by helping founders make slightly better decisions over time. Data, market timing, and investment mistakes (Priority: 4/5): Miles says data can be a trap in early-stage investing and reflects on regrets around Figma and Plaid, where he underestimated broader adoption and got too intellectual about market structure. AI’s real opportunity: new architecture and business model (Priority: 5/5): He believes AI’s biggest upside is not just copilots or wrappers, but a shift from software tools to control centers that sell work, outcomes, and SLAs on performance. Benchmark’s current AI focus (Priority: 4/5): He explains why Benchmark is excited about infrastructure and application frameworks for AI developers, especially tools that help every engineer build AI-native products like LangChain.
Key Arguments: Great investing begins by asking what is changing, what is new, and where dislocation is opening up. Early-stage data can mislead; product, customer, and user understanding should come first. Kindness and competitiveness are compatible, and strong firms can be both humane and ambitious. Will and passion often outperform raw skill because they compound through effort and learning. A founder’s willingness to admit ignorance and learn quickly is a strength, not a weakness. VCs should do no harm; the best value comes from trust, context, and helping founders make better decisions, not from imposing checklists. The best founders do not truly need investors, but the right partner can amplify their outcomes. Benchmark treats investing as a full commitment, not a financial bet, and does not reserve with the idea of buying more later. AI is more likely to create a new architecture and business model than just another UX layer like mobile. Co-pilot products are largely an incumbent strategy; the startup opportunity is to be orthogonal and sell work/outcomes instead of software seats. Infrastructure for AI is attractive now, but over time application-layer control centers could capture substantial value. The key mistake in Figma was underestimating how many non-design users would adopt it; the key mistake in Plaid was being too rational about why banks wouldn’t offer the needed connection. Great founders often make new mistakes each time; the slope of learning matters more than immediate perfection.
Data Points: Cloud spend growth: $10 billion to $300–$400 billion - Miles cites this as the scale of the 2012–2022 cloud transformation. Timeframe at Thrive Capital: 6 years and 9 months - Referenced during the discussion of his tenure before Benchmark. Benchmark partnership meeting time block: 10 a.m. to 3 p.m. - Describes Benchmark’s unstructured weekly partnership meeting format. Potential recruitment compounding: 10 extra hours per week / 500 hours per year - Used to explain why passion and extra effort compound into skill. Monzo accounts: 7 million accounts - Miles uses Monzo to show how a mission-driven founder can break through at scale in UK banking. UK population share on Monzo: About 1/10 of the British population - Illustrates Monzo’s scale and penetration in the UK market. Benchmark first-partner focus: 50–60% - Miles says roughly half to most Benchmark deals are the first partner relationship. Benchmark first investment rate: About 30% - He says roughly a third of investments are literally the first investment at inception. AI developers expected: Tens of millions - He argues every engineer will become an AI developer. ML specialists today: Hundreds of thousands - Contrasted with the future of tens of millions of AI engineers. Existing software SLAs: Uptime and support - Compared against future AI SLAs on outcomes/work performance. Revenue spend example: 3x Slack - Miles says many Figma users may spend about three times as much as they do on Slack. Founder/user example: 21 vs 35 - Used to contextualize when a founder may not know certain SaaS concepts because of career stage. US banking concentration: 5 banks control over 50% of accounts - Used to explain why Plaid’s bank connectivity problem persisted. Figma pricing/usage example: Paid seats and broad usage across non-design functions - He says he missed that design workflow reaches many roles beyond designers. AI adoption lag example: Salesforce launched about 5 years after the internet came about - Used to argue that transformative business-model shifts can take time.
Pivotal Quotes: "The very best entrepreneurs are always making mistakes. The really great ones are just making new mistakes each time." — Miles Grimshaw: Explaining how learning speed and adaptability matter more than perfection. "It should feel like planning an adventure, not a colonoscopy." — Miles Grimshaw: Describing how Benchmark approaches diligence and founder conversations. "Sell the work, not the software." — Miles Grimshaw: Summarizing his view of the next major AI business model shift.
Implications: Listeners should expect AI to evolve from assistive copilots into outcome-driven systems, while investors should focus less on checklist metrics and more on change, user need, and founder learning velocity.