The Twenty Minute VC"Cursor is Dead" is Total BS: Here is Why | Miles Clements
CHAPTERS
- 0:00 – 0:59
Opening thoughts: Growth can hide business fragility
Miles opens with a warning that headline growth and financial metrics can be misleading in today’s environment. He sets the tone for a conversation focused on fundamentals, product pull, and what actually endures as markets move fast.
- •Over-fixating on financial metrics can lead to bad judgments
- •Rapid growth can mask underlying product or business “ills”
- •Sets up the episode’s tension between hype, metrics, and durability
- 0:59 – 2:39
A framework for “true value” in AI: time-to-value vs durability
Miles proposes a simple lens for evaluating AI companies: how quickly users get value, and how durable that value is once achieved. He contrasts slow-to-deploy but sticky vertical AI (legal/accounting) with fast but shallow “vibe coding” apps, and argues coding tools win on both axes.
- •Two-axis framework: time-to-value and durability of value
- •Vertical AI can be slow to deploy but extremely sticky once adopted
- •Vibe-coding can deliver instant wins but may lack durable differentiation
- •Coding is today’s AI battleground because it’s both fast and compounding
- 2:39 – 4:25
Why “Cursor is dead” is misleading: market expansion and agents
Harry cites claims that developers are leaving Cursor for Claude Code and that Cursor is overpriced. Miles argues this is a false zero-sum narrative: the market is expanding rapidly, revenue growth is heavily consumption-driven, and Cursor’s product is shifting from IDE novelty to agent-centric workflows.
- •Claude Code’s rise is tied to model momentum (e.g., new Opus versions)
- •Success is not necessarily at Cursor’s expense—market is expansionary
- •ARR growth is driven significantly by consumption, not just per-seat deals
- •Cursor is evolving beyond the IDE framing; agents are the core shift
- 4:25 – 7:03
Cursor’s agent metrics and multi-model advantage
Miles points to public usage stats showing Cursor’s agent product is now central to how users work. He also argues multi-model support is strategic because developers switch models constantly, making Cursor an “index” of model innovation with compounding product improvements.
- •More users engage with agents than classic tab completion features
- •High daily usage of agents; rapid agent growth rates cited
- •Cloud agents contribute meaningfully to merged PRs
- •Developers frequently switch models; multi-model is the default future
- •Multi-model creates a flywheel: model improvements compound product capability
- 7:03 – 7:43
Why Cursor building its own models can still make sense
Miles rejects the idea that Cursor was wrong to pursue its own models, framing the goal as specialization rather than general-purpose AGI. Specialized coding models can serve enterprise needs and deliver differentiated workflows without needing broad consumer capabilities.
- •Generalists vs specialists: Cursor’s aim is specialized coding models
- •Enterprise coding tasks benefit from domain-specific optimization
- •Differentiation doesn’t require being good at non-coding tasks
- •Model strategy can reinforce product moat and enterprise adoption
- 7:43 – 11:33
Underwriting Cursor at $9.5B vs $27B: engineering platform aspiration
Miles explains the investment upside by comparing Cursor’s ambition to platform companies that “own” a domain. He argues engineering lacks a true platform owner across the stack, and Cursor’s financials—while stunning—are ultimately treated as a reflection of extreme product-market fit rather than the primary underwriting driver.
- •Thesis: first true platform company for the engineering vertical
- •Prior winners built huge value from slices (issue tracking, monitoring, etc.)
- •Multiples matter less than the product pull in hyper-growth situations
- •Revenue planning is a tool for assumptions, not for “managing to plan”
- 11:33 – 19:34
Do smaller outcomes still matter? Fund size, nuance, and ServiceTitan lesson
Harry challenges whether anything under massive scale can matter for large funds. Miles argues outcome sizes have expanded (including trillion-dollar comps), but also stresses portfolio nuance—avoiding the trap of only chasing extremes—and shares the ServiceTitan miss as a lesson in being overly price-disciplined when the market depth is real.
- •Outcome sizes have grown substantially over the last decade
- •Pure late-stage momentum investing is hard; multi-stage helps
- •Avoid “middle” positioning; embrace either clear consensus or clear non-consensus
- •ServiceTitan miss: rigid multiple rules caused Accel to lose a great company
- •“Winning” doesn’t imply monopoly; multiple large companies can coexist
- 19:34 – 25:31
Missing Rippling: marginal ease of ARR accumulation and rule-breaking
Miles explains why Rippling stings: Parker Conrad’s ability to create compounding growth levers made the business unusually scalable. He discusses investor errors like extrapolating from anomalous quarters and how reputational concerns, speed, valuation, and ownership rules can cause firms to miss generational founders.
- •Concept: “marginal ease of ARR accumulation” as a growth mechanic
- •Rippling’s adjacent modules create compounding distribution and retention
- •Reputation and speed can be decisive in competitive rounds
- •Ownership thresholds and “breaking rules” are sometimes necessary
- •Growth can obscure weaknesses; usage intensity is increasingly critical
- 25:31 – 31:33
Coverage, accountability, and win rate: how Accel reflects on misses
The conversation shifts to how firms assess performance—coverage of top companies and ability to win them. Miles describes rigorous partnership introspection (including global offsites), an aspirational coverage mindset, and a realistic view that losing deals is part of competing for the best opportunities.
- •Internal discipline: track “investor of record” status in top private companies
- •Global offsite retrospectives to diagnose misses and course-correct
- •Coverage is essential; 100% coverage/100% win is aspirational, not realistic
- •Miles suggests an ~80% win rate might be “healthy”
- •If you never lose, you may not be competing hard enough
- 31:33 – 35:30
Ownership has changed: laddering into 20% via multi-stage tactics
Miles argues ownership math is inverted versus earlier venture eras: rather than buying big early and letting dilution happen, firms now ladder toward meaningful ownership through tenders, growth rounds, and IPO participation. This favors multi-stage platforms that can support companies across phases.
- •Old model: high early ownership, accept dilution into IPO
- •New model: build ownership over time via multiple entry points
- •Tools: founder/employee tenders, growth rounds, IPO rounds
- •Multi-stage structure is increasingly strategic for ownership outcomes
- •Avoid “hope and pray” as a core ownership plan
- 35:30 – 39:15
Anthropic at $180B and the Pentagon controversy: principles vs pragmatism
Miles discusses investing in Anthropic and why a few model companies may operate on a trillion-dollar trajectory. He then addresses the tension around defense-related controversy, emphasizing respect for founders who adhere to stated missions and principles even when commercial incentives pull the other way.
- •A small set of companies may plausibly become trillion-dollar outcomes
- •Danger: incorrectly applying “Anthropic-like” logic to ordinary Series A deals
- •Pentagon episode highlights ethics under commercial pressure
- •Miles respects mission adherence, while acknowledging uncertainty in outcomes
- •Consumer traction can sometimes increase amid controversy
- 39:15 – 44:22
What happens to 2021-priced SaaS (Miro, Snyk): LBO landings and founder leadership
Miles describes the post-2021 market as humbling: great businesses can still be great even if entry prices were too high. He predicts many companies will find outcomes via private equity or alternative paths, and argues founder-led leadership has a special resilience—while noting exceptional professional CEOs can also work.
- •Many “great” companies are down; valuation pain doesn’t erase product quality
- •Investor behavior matters most when teams’ backs are against the wall
- •Likely beneficiaries: Thoma Bravo/Vista-style buyout platforms
- •Founder-led advantage is real, though not absolute
- •Professional CEOs can be world-class in the right situations
- 44:22 – 47:14
Why go public now? The sub-$5B IPO trap, liquidity, and secondary markets
Miles explains why top private companies can now get liquidity and M&A currency without IPOs, which reduces the incentive to go public. He notes that going public too small can cap breakout potential, and discusses how cash-flow valuation resets may be rational but have also over-rotated in the market’s “SaaSpocalypse.”
- •Private markets now offer employee liquidity and M&A currency via secondaries
- •IPO motivation remains stronger for companies outside the top-tier private set
- •Sub-$5B public outcomes can struggle to break out; many wait for clearer scale
- •Valuation reset reflects changed cash-flow discounting; some oversold cases exist
- •Example cited as oversold: Figma (from the outside)
- 47:14 – 1:03:53
Secondaries, taking chips off the table, evergreen funds, and boardroom dynamics
Miles lays out a principle for secondaries: prioritize what’s best for the company, but diversify when the company explicitly supports a tender. He contrasts situational liquidity decisions (WeWork vs CrowdStrike-style compounding), addresses whether VCs should manage public positions, and closes with practical guidance on boards, career advice, and Accel’s internal “who’s best at what.”
- •Chips-off decisions are situational; company-first is the governing rule
- •Counterexample: staying invested through IPO can compound massively (CrowdStrike)
- •Evergreen/public-hold strategies should be selective, not blanket policy
- •Board insight: the most vocal board members are often the least helpful
- •Accel “best of”: sourcing, picking, and winning relationships; plus career professionalism advice
