"Cursor is Dead" is Total BS: Here is Why | Miles Clements

TL;DR

Excel investor Miles Clements argues Cursor's 'death' narrative is wrong because agents grew 15x and the IDE-focused criticism misunderstands their platform ambitions.

Key Points

• 1.The 'Cursor is dead' narrative fundamentally misreads the market. Despite Chamath and Jerry Murdoch claiming developers are abandoning Cursor for Claude Code, Cursor hit $2B ARR; Claude Code's rise is market-expansionary, not cannibalistic.
• 2.Cursor's agent product has completely eclipsed its original tab feature. 2x more users use agents than tab completion; agents grew 15x last year; 90% of users are daily active on the agent product; cloud agents now handle 35% of merged PRs.
• 3.Being multi-model is Cursor's core strategic advantage. 50% of developers switch model families daily and 95% switch models daily, making Cursor's model-agnostic approach a compounding product flywheel benefiting from every AI improvement.
• 4.Excel invested in Cursor at a $9.5M valuation when ARR was roughly $100M. Management projected $500M year-end ARR; Excell haircut it to $300M; the company ended the year in the billions — rendering financial modeling largely irrelevant.
• 5.Cursor is building specialized coding models, not general-purpose AI. Their internal models target professional coders doing professional work, not poetry or recipes, giving enterprise users a differentiated, task-specific experience.
• 6.No platform company has ever owned the engineering vertical despite its scale. Atlassian (issue tracking) and Datadog (monitoring) each built $50–100B businesses on single slices of the stack; Cursor is attempting to own the entire engineering platform.
• 7.Excell missed Rippling because of Parker Conrad's prior reputation and valuation discipline. Rigid rules around ownership thresholds and price caused them to pass; Rippling's 'marginal ease of ARR accumulation' via compounding revenue lines like laptop provisioning made Conrad a generational founder in hindsight.
• 8.Excell invested in Anthropic starting at the $180 round, believing it could be a trillion-dollar company. They praised Dario's decision to reject the Pentagon deal as principled ethics, noting Anthropic subsequently passed GPT-4 in App Store rankings with 1M daily net new consumer signups.
• 9.Miles fundamentally changed his view that all generational AI investments had already been made. A year ago he believed the best bets (Scale AI at $14.9B exit, lab investments) were locked in from 2016; he now believes the innovation flywheel is just starting.
• 10.Figma was cited as the most oversold public company despite Dylan Field being a generational founder. The SaaS apocalypse is partly an overreaction, but also a correction from valuing companies at inflated multiples on mediocre 18–20% growth rates; Figma at $11B 'doesn't return your growth fund.'