Mar 9, 2026 · 1h 3m · news
"Cursor is Dead" is Total BS: Here is Why | Miles Clements · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this deep-dive interview, Accel growth partner Miles Clements discusses the mechanics of modern venture capital, debunking rumors surrounding AI startups like Cursor while exploring Accel's investment philosophy, market valuations, and the institutional culture that sustains long-term venture success.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 28.1% of the talking time here. How this is scored →
Harry as informed peer 4.9 Guest teaching 4.8 Guest disagreement 3.3 Harry pushing back 4.2
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 4:25 Guest rejects third-party Cursor critiques
Miles aggressively counters criticisms from Jerry Murdoch and Chamath, calling out their lack of real metrics with the line 'who are these people to make these judgments?'
Hardest push from Harry ▶ 33:30 Host rejects 'singles and doubles' VC strategyHarry explicitly pushes back on Arthur Patterson's advice, arguing that Series A venture capital is strictly about taking massive swings for hundred-billion-dollar outcomes rather than playing for singles and doubles.
Biggest teaching moment ▶ 18:35 Guest counters monopoly fallacy with enterprise dataMiles educates Harry on market dynamics by citing AWS's 35% market cap limit and pointing out that Deel's $1B ARR is still only 1/20th the size of legacy player ADP.
Harry holds his own ▶ 14:06 Host forces concession on fund math and outcomesHarry presses Miles on Accel's multi-billion dollar fund math, forcing Miles to agree on outcome requirements and share Accel's expensive pricing lesson on ServiceTitan.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Podcast Welcome & In-Person Greeting | 5 | 5 | 6 | 4 | Harry cites prominent industry voices like Chamath and Jerry Murdoch to frame Cursor as declining or overly expensive. Miles responds aggressively, defending Cursor's product velocity and dismissing third-party commentators for lacking real internal metrics. | |
| Cursor's Multi-Model and Specialized Model Strategy | 6 | 5 | 5 | 7 | Harry challenges Accel's willingness to invest in moderate-growth companies when 15x AI opportunities exist. Miles defends portfolio nuance and multi-stage strategies, while Harry repeatedly presses on opportunity cost and fund size constraints. | |
| Fund Size, Outcome Scaling, and the Service Titan Lesson | 6 | 6 | 4 | 6 | Harry insists Accel's large fund size requires multi-billion dollar outcomes, rejecting lower hurdle rates. Miles concedes the point by bringing up Accel's past mistake with ServiceTitan where over-rigidity on pricing metrics cost them a massive winner. | |
| The Fallacy of "Winner-Take-All" Monopolies in Tech | 6 | 7 | 5 | 5 | Harry points to mega-caps like Nvidia and Salesforce to argue tech operates in winner-take-all monopolies. Miles counters with market share data from AWS and compares Deel's $1B ARR to ADP's $20B ARR to prove market expansion allows multiple massive players. | |
| Stinging Misses: Rippling and Generational Founders | 4 | 6 | 2 | 3 | Harry asks about missing deal opportunities like Rippling. Miles introduces his concept of 'marginal ease of ARR accumulation' and reflects openly on missing Parker Conrad due to past reputation concerns and strict rule adherence. | |
| Sticking to Rules vs. Breaking Them & "Marginal Ease of ARR" | 5 | 5 | 3 | 5 | Harry teases Miles's elaborate VC vocabulary and asks whether investors should break ownership rules for hot AI Series A rounds. Miles shares a quote from Jim Breyer on investing as art versus science and warns against constantly breaking discipline. | |
| The Competitive Thrill of Venture Capital | 5 | 5 | 6 | 5 | Harry mentions investor fatigue and asks how Accel handles missing generational companies like OpenAI or Anthropic. Miles vehemently rejects the fatigue narrative and details Accel's rigorous internal offsites where partners hold themselves strictly accountable for misses. | |
| Growth Fund Scaling & The "Singles and Doubles" Debate | 7 | 5 | 5 | 7 | Harry directly rejects Arthur Patterson's classic VC advice of focusing on singles and doubles, arguing Series A investing demands swinging for hundred-billion-dollar outcomes. Miles defends Accel's disciplined approach against pure momentum chasing. | |
| Underwriting Anthropic and the Power of Mission-Driven Founders | 5 | 5 | 4 | 5 | Harry probes Accel's decision to underwrite Anthropic at an $18B valuation and brings up the controversy around military applications. Miles justifies the valuation by framing Anthropic as a candidate for a trillion-dollar platform and praises the founders' principled stance. | |
| Post-Bubble Valuations and the Power of Founder-Led Companies | 5 | 5 | 2 | 4 | Harry examines late-stage 2021 bubble portfolio companies like Snyk and Miro that are struggling under inflated valuations. Miles explains the shift toward private equity buyouts for maturing SaaS businesses and emphasizes staying backed by committed founders. | |
| The AI Innovation Wave and Accel's Scale AI Exit | 5 | 6 | 2 | 4 | Miles candidly admits he was wrong a year ago when he thought generational AI bets were already finished. Harry brings up Scale AI's $14.9B outcome and asks how to properly value data and services businesses in the current wave. | |
| Secondary Liquidity and Managing the Public Book | 6 | 5 | 3 | 5 | Harry questions why private companies stay private so long and challenges holding public stocks given market volatility. Miles uses CrowdStrike as a case study, showing how holding from early stage through public listing generated massive compounding returns. | |
| Accel's Superpowers and the Art of Being a Helpful Board Member | 4 | 5 | 2 | 3 | Harry asks about internal partner strengths at Accel and queries whether top-tier founders actually need VC assistance. Miles clarifies that great founders do not need micromanagement, but rely on strong board members for key yearly 'bumper decisions'. | |
| Top VC Recommendations: Seed, Series A, and Growth | 3 | 2 | 1 | 1 | Harry moves into a quickfire round asking for top non-Accel VC firm recommendations across stages. Miles praises Liquid 2 for seed, Meritech for growth/Series A, and Thrive for late-stage conviction. | |
| Venture Career Guidance: Preserving Long-Term Professionalism | 4 | 3 | 1 | 2 | Harry asks about Miles's career guidance, biggest losses, and favorite deal wins. Miles shares a personal story about winning the opportunity to invest in Linear by staying near the founder in San Diego during a difficult personal time. | |
| Hype Cycles and the Realities of Media Perceptions | 3 | 2 | 1 | 1 | Harry and Miles discuss intense investor hype around companies like Lovable and Linear. The episode concludes with Miles expressing optimism about the emerging generation of young investors at Accel. |
Statements from this episode (44)
Clements: Rapid growth can obscure underlying structural problems within a business
“Growth can obscure and blind you to a lot of underlying ills in the business.”
Clements: VCs can achieve success in the current market backing consensus deals
“You can actually be successful in this market investing in consensus.”
Clements: The art of investing lies in knowing when to break rules
“Investing is an art and a science. The science is understanding how to Properly value a company in the yard is understanding when to break the rules.”
Clements: AI coding leads other verticals due to quick and durable value
“The reason that I think coding has become like the vertical in AI is because it shines on both dimensions. Like you can start using cursor in an afternoon and by that evening, like you're 10 times more productive. The time to value is very short. And then the …”
Clements: Cursor has 2x more agent users than tab autocomplete users
“There are two times more people using agents in cursor than using the tab feature.”
Clements: 90% of Cursor users are daily active agent users
“90% of cursor users are daily active users of the agent product.”
Clements: Cursor's AI agent product usage grew 15x last year
“The agent product grew 15 X last year.”
Clements: Cursor cloud agent accounts for 35% of merged PRs
“You know, the cloud agent product, which was new as of like October 30th, so it's only been in market for three months, is now responsible for 35% of merged PRs in cursor.”
Clements: 95% of developers switch AI models daily
“And one of the things that we're learning is like 50% of developers switch model families on a daily basis and 95% of developers switch models on a daily basis.”
Clements: Cursor will build specialized coding models for enterprise users
“Cursor is going to build specialized coding models that are going to serve specialized coding tasks, especially for a lot of enterprise users.”
Clements: No company has ever built a single platform for engineering
“There has never been a platform company for engineering as a vertical and engineers are like, I mean, this is the fastest growing, most dynamic vertical there is, and no one has ever owned that.”
Clements: Accel's initial Cursor investment was priced at 4-5x year-end ARR
“Actually, this company is growing so quickly that on a multiples basis, you know, our first investment was at like four times, five times year-end ARR.”
Clements: VCs get hammered sitting in the middle of consensus
“I think you can actually be successful in this market investing in consensus. And I think you can actually do really well investing in like non-consensus. I think you get hammered sitting in the middle.”
Clements: Public trillion-dollar market cap companies grew from zero to twelve
“A decade ago, there were zero companies worth a trillion dollars. Five years later, there were six public companies worth a trillion dollars. Today, there's a dozen companies worth a trillion dollars in the public market.”
Clements: Accel lost ServiceTitan over rigid valuation caps before $9B valuation
“We had fallen in love with our envy. We were chasing this round. It was going to happen in the two 50 or three hundred million dollar range. And, you know, we had these rigid rules about like, you definitely can't pay more than six times, six to eight times fo…”
Clements: AI verticals will sustain multiple major winners, not monopolies
“But I also think the way that a number of these verticals are going to play out in a number of the AI categories, there's going to be a couple of really big companies and several of them.”
Clements: Nobody builds ARR compounding levers better than Parker Conrad
“And I think nobody in the world does that better than Parker Conrad.”
Clements: Breaking VC rules for Rippling would have been worthwhile
“And I think like, I don't regret not breaking the rules in general, but you know, this would have been a time when it could have been worthwhile.”
Clements: Extrapolating temporary ARR spikes as product-market fit is a trap
“There have been investments where, you know, a company went from, they had a million dollars of ARR, and then in the period that you know, before they fundraise, they had like a four million dollar quarter, and it's like, they've got it. Like the product marke…”
Clements: Traditional VC startup benchmarks are now largely obsolete
“I think this is why, you know, the benchmarks that used to give us all comfort are largely obsolete now. And so, like, you have to be really clued into the usage intensity of your product and really understand how people are using it.”
Stebbings: Jason Lemkin says SaaS VC investing is exhausting compared to AI
“I was talking to my dear friend Jason Lemkin the other day, and he's like, fuck this. I've had enough of this. I just want to do an anthropic SPV and go home. I don't want to pick the winner in a SaaS company.”
Clements: Accel missed investing in early AI foundation model companies
“A lot of firms miss the model companies early, and we're guilty of it.”
Clements: A healthy VC deal win rate is around 80%
“I think a healthy win rate would be, like, 80%.”
Clements: VCs that never lose deal fights aren't competing hard enough
“I think if you're not putting yourself out there in losing, sometimes you're not chasing competitive enough things.”
Clements: VC funds must ladder ownership over time to reach 20%
“Today, you have to back into 20% the other way. Like, you do what the market will allow in the earliest possible investment. You sponsor a tender, you do a growth round, you do an IPO round, and you can ladder your way up to 20% ownership. You have to be a mul…”
Clements: Targetting $100B exits at Series A leads to failure
“If you're just constantly stepping to the plate and trying to, you know, I can see at the series A that this is going to be a hundred billion dollar exit, you will just overswing and you will fail.”
Clements: Anthropic and select private AI companies could reach $1T valuations
“I think that company and a small handful of companies in the private market today are operating on a different plane. You know, I think it is not bombastic to say that some of those businesses could be trillion dollar companies.”
Clements: Anthropic's app surpassed OpenAI's ChatGPT in the App Store
“Yeah, I mean, they passed GPT in the app store.”
Clements: Top public software companies are currently oversold
“I mean, the greatest companies of three, four, five years ago, many have gotten totally beat up in the public markets. I believe many are oversold, but you know, that's a separate conversation.”
Clements: Private equity firms will acquire overvalued 2021 tech startups
“I think it's probably a good time to be in the LBO business. I think it's probably a good time to be in the Toma Bravo Vista Blackstone KKR business. There will be homes for a lot of those companies, you know, who get themselves to a sustainable place and they…”
Clements: Never bet against Atlassian co-founder Mike Cannon-Brookes
“Never bet against Mike Cannonbrooks. Absolutely not.”
Clements: Generational AI opportunities and outcomes are just beginning
“Those companies will be bigger than the outcomes that I probably envisioned a year ago, and there is still time to be a part of some of them, and because, like, the innovation flywheel is just getting started. We are barely scratching the surface.”
Clements: Accel avoids major celebrations even after $14.9B exits
“There's a an appropriate congratulations and acknowledgement to Dan. There is a huge, loud, full-throated thank you to Alex, and then everybody gets the fuck back to work, you know, and like, it's a humbling industry, and you are only as good as the next thing…”
Clements: Only top 10 private tech companies can avoid IPOs
“With that said, I think that applies to like the 10 best private companies in the world. Like Databricks can do those things. Stripe can do those things. There's a lot of companies that just do need to get public.”
Clements: Startups delay IPOs to avoid $5B market cap murkiness
“I don't think that's actually the reason. I think it's because generally speaking, you want to go public and you want to be able to have like fairly clear line of sight to hitting the five billion dollar threshold and trading beyond that. Cause it's murky belo…”
Clements: Public market repricing of SaaS software stocks is an overrotation
“Fundamentally, people are valuing the future cash flows and the future terminal value of these companies differently, and I don't think that's wrong, but I do think this has been an overrotation.”
Clements: Accel is not in Figma, calls company oversold
“We are not a part of Figma, but have a lot of respect for that company. But I know that Dylan is a generational founder, and it's a very important company with an incredible financial profile, and it just feels for a lot of ways, for a lot of reasons, oversold…”
Clements: Accel did not sell secondary shares in Miro at $17B valuation
“We didn't take liquidity out of Miro at seventeen billion”
Clements: Accel invested in CrowdStrike at $160M valuation with $1M ARR
“The example I would point to is, you know, CrowdStrike. Samir Gandhi and John Locke intersected CrowdStrike when it was, you know, there was like a million dollars of software revenue and there was a nine million dollar consulting business, and like, that was …”
Clements: Highly vocal board members are inversely correlated with actual helpfulness
“I think there's generally an inverse correlation between how vocal somebody is and how helpful they actually are. So the person who just has to get the first and last word in and shows up at the board meeting and has to teach you something that you didn't alre…”
Clements: Accel regrets missing out on investing in ElevenLabs
“I think 11 Labs is a clear company that we wish we had been a part of. As you know, we haven't spent enough time with the founder, which is, you know, our loss. I think we really regret that one.”
Clements: Losing a competitive deal stings more than missing it
“They're equally bad, but losing stings more. Having had the opportunity and failed, stings.”
Clements: Abridge founder Shiv Rao is a generational talent
“I don't know that I would characterize it as we lost, but the company that I, where I really, really loved the founder and we didn't get there was Shiv and a bridge. We actually hosted this AI dinner a couple of weeks ago and I was like, I'm going to manipulat…”