Sep 18, 2023 · 1h 43m · news
Miles Grimshaw: The 5 Pillars of Venture Capital & Why Co-Pilot is an Incumbent Strategy | E1061 · 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 episode of 20VC, venture capitalist Miles Grimshaw of Benchmark joins Harry Stebbings to share valuable insights on user-centric investing, the transition from Thrive Capital to Benchmark, and why startups must build outcome-based architectures rather than simple AI co-pilots.
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 18.6% of the talking time here. How this is scored →
Harry as informed peer 4.3 Guest teaching 4.3 Guest disagreement 2.2 Harry pushing back 2.9
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 1:05:39 Miles rejecting co-pilot framing for AI software
Miles forcefully argues that co-pilot features are an incumbent strategy and asserts that true startup disruption requires moving completely away from selling seats to selling work outcomes.
Hardest push from Harry ▶ 23:29 Harry calling out basic SaaS illiteracy in foundersHarry refuses Miles' lenient framing regarding an engineering founder not knowing LTV, stating flatly that basic SaaS mechanics are non-negotiable with today's information availability.
Biggest teaching moment ▶ 50:50 Miles dissecting his Figma pass and revenue per seat dynamicsMiles educates the host on how Figma expanded far beyond core designer seats to become the central collaboration hub for product, marketing, and engineering, generating massive spend.
Harry holds his own ▶ 23:29 Harry pressing on core business fundamentalsHarry directly challenges the guest's perspective on founder evaluation by asserting that basic metric knowledge is mandatory for any SaaS founder.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Entry into Venture Capital and Stepfather’s Influence | 3 | 2 | 1 | 1 | Miles shares his origin story moving to the US and seeing his stepfather build an e-commerce startup. Harry chimed in with brief contextual references like Tumblr and Foursquare. | |
| Market Dinosaurs, Terminology, and Amazon's "Working Backwards" | 5 | 3 | 3 | 5 | Harry challenges the idea of market size comparison by arguing that being first vs second generates vastly different value. Miles counters that both served different user needs and emphasizes focusing on the user. | |
| Formative Lessons from Josh Kushner | 3 | 4 | 1 | 2 | Miles outlines three core lessons learned from Josh Kushner at Thrive Capital. Harry asks an insightful follow-up on how Josh detects investor talent. | |
| The Concept of Being "Impatiently Patient" | 3 | 2 | 1 | 2 | Miles discusses the concept of being impatiently patient and how he decided to join Benchmark. Harry relates to the struggle of reconciling urgency with long-term firm building. | |
| Benchmark’s Unique Partnership Meetings | 4 | 4 | 1 | 2 | Miles explains Benchmark's agenda-less Monday partnership meetings and introduces his 5 S's framework. Harry engages by proposing his own 4 pillars before Miles adds the fifth. | |
| The Engineer's Pitch & The SaaS Jargon Debate | 7 | 5 | 5 | 8 | Harry strongly pushes back when Miles suggests it is fine for a founder not to know LTV, arguing basic business metrics are non-negotiable. Miles defends his stance by pointing out rate of learning and engineering backgrounds. | |
| Do VCs Add Value & The Hippocratic Oath | 4 | 4 | 2 | 3 | Miles outlines his belief in VC value-add based on trust, context, and decision quality, citing a Hippocratic oath approach. Harry brings up Sam Lessin's view that top founders don't need VCs. | |
| The Rubric: Founder, Traction, and Market | 5 | 4 | 3 | 4 | Harry asks how Miles weighs founder, traction, and market. Miles reframes it as an integrated system, warning that early traction data can often be a trap or empty calories. | |
| Market Dislocations and What to Look For | 4 | 4 | 2 | 3 | Miles explains looking for market dislocations and shifts as key investment signals. Harry equates this to market timing risk versus adoption risk. | |
| The Equal Partnership Model at Benchmark | 5 | 4 | 2 | 4 | Miles discusses Benchmark's equal partnership model and minimal reserve strategy. Harry questions how Benchmark positions itself when taking checks from seed stage to Series C. | |
| Lessons From Successes: The Story of Benchling | 3 | 4 | 1 | 1 | Miles contrasts his experiences with Benchling and Monzo, highlighting how two very different founder archetypes achieved massive scale. | |
| The Regret of Market Hesitation: Figma, Plaid, and Scale | 5 | 7 | 2 | 2 | Miles shares detailed post-mortem lessons on passing on Figma, explaining how he misjudged market size by ignoring non-designer seats. | |
| Analyzing Plaid and the Illusion of Rational Markets | 4 | 5 | 2 | 2 | Miles analyzes his miss on Plaid, pointing out how over-intellectualizing market rationales blinded him to banking lock-in incentives. | |
| Specialist vs. Generalist: The Biologist and Physicist Analogy | 4 | 5 | 2 | 3 | Miles discusses being a generalist and uses the analogy of a biologist observing adaptations versus a physicist enforcing fixed rules. | |
| Why the Mobile Analogy Fails for the AI Era | 5 | 6 | 3 | 3 | Miles breaks down why comparing the AI shift to mobile is flawed, arguing mobile didn't change core application architecture or business models whereas AI does. | |
| From Co-Pilot to Control Center: Outcome-Based Software | 5 | 6 | 3 | 2 | Miles details the shift from seat-based co-pilots to outcome-based control centers selling work and SLAs on results. | |
| The Co-Pilot Trap and Incumbent Dominance | 4 | 5 | 3 | 3 | Miles explains why co-pilot features are inherently incumbent strategies, forcing startups to seek orthogonal architectures. Harry questions if co-pilots are a transitional step. | |
| Incumbent Inaction and the "No More Software" Rallying Cry | 6 | 4 | 4 | 6 | Harry challenges Miles on adoption speed, arguing enterprises will take 10 years to move past co-pilots, creating a timing trap for outcome-based AI startups. | |
| Societal Ramifications of AI and the Impact on Media | 4 | 4 | 2 | 2 | They discuss societal impacts of AI on content creation, discovery, and trust. Miles argues trusted legacy publishers like The New York Times gain centralized value. | |
| Debunking AI Tropes: "Wrappers" vs. New Architectures | 4 | 5 | 3 | 2 | Miles debunks the narrative that AI apps are just thin wrappers on LLMs, putting current developments into historical context regarding software paradigm shifts. | |
| Where Value Accrues: Application vs. Infrastructure Layer | 5 | 5 | 3 | 4 | Harry posits that value currently accrues to the infrastructure layer. Miles agrees for the short term but asserts long-term value will concentrate in outcome-based applications. | |
| Inside Benchmark's Investment in LangChain | 4 | 4 | 1 | 2 | Miles shares the story behind leading Benchmark's investment in LangChain and how he built a relationship with founder Harrison Chase prior to incorporation. | |
| Ultra-Marathons, Mental Grit, and the Perfect Pre-Run Nutrition | 2 | 3 | 1 | 1 | A lighter segment discussing ultra-running, mental grit, and sleep deprivation during 60-hour adventure races. | |
| Quick-Fire Round: Personal Growth, Board Dynamics, and Benchmark Lessons | 5 | 4 | 2 | 3 | In the quick-fire round, Miles discusses fatherhood, Toby Lütke's fast decision reversal at Shopify, founder respect over friendliness, and board dynamics with Eric Vishria. |
Statements from this episode (60)
Grimshaw: Tens of millions of software engineers will become AI engineers
“Tens of millions of engineers would become AI engineers versus like hundreds of thousands of ML scientists. That would beget the need for an application sort of framework for AI and tooling for that computing environment.”
Grimshaw: Cloud spend grew from $10B to $400B between 2012 and 2022
“If you go and look, call it 2012 to like 2022 ish timeframe, cloud spend went from like ten billion to three, four hundred billion.”
Grimshaw: User Focus Matters More in VC Than Metrics Frameworks
“And the most important thing is always like, who's the user and why do they care? If you ground yourself in that always, instead of sometimes a sort of like whirlwind of other, you know, framings from investing, it'll probably serve you pretty well.”
Grimshaw: Josh Kushner combines hyperactive daily urgency with long-term patience
“The third one would be this again, The greatest gift of things you sort of carry with you as ways of being would be what I might call impatiently patient. Again, sort of a little, a duality that almost shouldn't exist which is why not now? Why not yesterday? L…”
Grimshaw: Operating without firm hierarchy makes being yourself the hardest challenge
“In many ways it was you be you, like you be, You totally kind of unfiltered and as authentic as you are, and the best version of you will be the best investor. And I think in many ways, when you get rid of structure in this, there's no like level to be had or,…”
Grimshaw: Benchmark uses no deal sheets, memos, or data room presentations
“We have no deal flow list, no sheet, there's no memos there's no data room presentations.”
Grimshaw: Venture investments are only complete upon LP distribution
“It's only a great investment in some sense when the, you know, shares are distributed to a broader shareholder base in terms of distribution or the like. I think people forget about, you know, in, in, in the push for forever funds and forever, you know, people…”
Grimshaw: VC diligence should feel like planning an adventure, not an examination
“In diligence, I think it should not feel like an examination. It should feel more like looking ahead at the future together and stepping into shoes together, which is the sort of work you'll be constantly doing with a founder. Sometimes I'd say it should feel …”
Grimshaw: Waiting for a Perfect Pitch Before Pitching Benchmark Is Silly
“Like some, sometimes there's a, ah, I'll only pitch benchmark when I got the perfect pitch. Like that's silly. That doesn't need to be the case”
Benchmark won't re-judge founders in future rounds based on early pitches
“We're only early. And so you get sort of a chance of it, but also we're not going to be able to, we're not going to re-judge you in the future based on what you told us now, because whereas other funds might.”
Grimshaw: A founder's rapid rate of learning beats knowing jargon
“The fact that they're willing to say, I don't know, and quickly learn, and grok it is really, I think, like the most important thing, that rate of learning, and in fact, I'd say most of us saw it as a strength.”
Grimshaw: The best entrepreneurs continuously make mistakes, but always new ones
“The very best entrepreneurs are always making mistakes. The really great ones are just making new mistakes each time, and I think there's a lot of truth to that, and so that slope and pace is is the most important.”
Stebbings: Not knowing basic SaaS metrics like LTV is unacceptable for founders
“With the freedom of information today, with the transparency of podcasts like this, the availability of SaaS materials, basically, That's not an acceptable thing, I think, if you're gonna start building a SaaS company. This is the core mechanics of any busines…”
Grimshaw: VCs who enforce generic checklists distract and burden founders
“I've been in some board meetings where investors don't really get the product, don't really get the strategy that has to be pursued. They're trying to apply like procedures and checklists, you know, versus strategies specific to the instance. It becomes a repo…”
Grimshaw: Innate founder capability cannot be coached
“I don't think it was coaching, but per se, but you can't coach height. Gotta have it.”
Grimshaw: Venture investing should be commitments, not bets
“If you think of investing as betting I think you spend a lot more, you know, a lot more energy on the funnel and winning the first, like you're placing chips on a roulette table. And I just don't, I don't think of it that way. I really think of the work we do …”
Stebbings: Second-time founders view VC funding as a deal, not partnership
“I find like second time founders, especially, I get it, you're selling cash with some advice, which is great, but it's a deal.”
Grimshaw: Product and customer come before data, which should only validate
“We should always start from like product and customer, not data. And data should be in some sense validation, not learning.”
Grimshaw: Slower, high-quality growth beats rapid growth in the wrong ways
“And often cases, I'd rather take less growth in the right ways than more growth in the wrong ways.”
Grimshaw: Key VC gut check is whether you can recruit for startups
“And so the gut check, the real question I think about a lot which gets to that sort of holistic nature of it is how do I imagine recruiting for this?”
Grimshaw: Venture investing requires finding market dislocations over market timing
“I don't personally think about it as much from sort of a timing lens, so much as the investigation and the thinking of like, what's different, what's new, what's a dislocation, what's emergent.”
Grimshaw: Investing on normative worldviews without market dislocations is dangerous
“I think it's dangerous in many ways, and this is sort of To think, oh, the world should work this way. And now I'm looking for people who also think the world should work this way, but don't really have the strategy for cracking that open, or there's no disloc…”
Grimshaw: Benchmark will accept adoption risk if market behavioral dominoes are falling
“Force of will can do a lot, but you know, it's got to feel like the, you know, you've got to feel like the dominoes got tilted over, you know, you're looking here, But the domino chain is, is sort of happening conceptually in the market or for behavior, whatev…”
Grimshaw: Benchmark's equal partnership aligns incentives across all portfolio companies
“We're an equal partnership, so it's a whole firm commitment. No person is incentivized to do more, help more one thing versus another. A founder really gets the whole partnership.”
Stebbings: VC follow-on reserves are financial mismanagement
“I actually think reserves are financial mismanagement.”
Grimshaw: Almost all Benchmark capital goes into initial investment checks
“If you would hit the spreadsheet of our investments almost all the capital we invest is the very first very first time we partner.”
Grimshaw: Benchmark is first institutional capital in 50-60% of deals
“I think it's something like 50 to 60% of all the investments we make, we're the first partner. I might want to use the word real, right? But like first sort of institutional investor, so to speak.”
Grimshaw: Roughly one-third of Benchmark investments occur at inception
“I think it's something like 30 odd percent. We're like literally the first investment. We don't use the term incubation. I'm sure there's a bunch of things we've done. People would call incubations. We don't use that term, but like, I think probably about a th…”
Grimshaw: Benchmark backs only about three inception-stage startups per year
“The N of that in a given year, even for that percentage is like three companies, right? And so it's not like we're doing 30, ah, that way and again, you know, so, so the N is small.”
Grimshaw names GitHub, Segment, Benchling, and Monzo as top investment returns
“I think if you looked at Realized, it would probably GitHub or Segment. If you look at Unrealized, maybe Benchling or Monzo,”
Grimshaw: Monzo Has Reached 7 Million Accounts in the UK
“Monzo now has seven million accounts. It's like a 10th of the British population has signed up to this thing.”
Grimshaw regrets missing early investments in Figma, Plaid, and Scale AI
“Dylan at Figma, Zach at Plaid and Alex at Scale, and I have enormous respect for all three of them got to know all three of them really early, and I just, I struggled with it.”
Figma was at $4M ARR when Sequoia invested at $400M valuation
“I remember when they, it's public, but they were at four million in ARR when Andrew did them at 400.”
Late-stage startups spend roughly 3x more per seat on Figma than Slack
“And so how that actually plays out is probably a third of the company from my surveying has paid seats inside of Figma, and they pay about three times as much as they would for something like Slack.”
Grimshaw: Five banks control over 50% of US bank accounts
“So like five banks control over 50% of accounts in the UK, in the US.”
Grimshaw: Banks lack incentive to offer open account authentication APIs
“Now they have no incentive to do that because if they let you do that, it's easier to move money out, and they don't want you to move money out because they make money on you storing money there. That's how a bank works. You get lock in to other products, but …”
Grimshaw: Missed investing in Plaid by over-intellectualizing bank incentives
“And I, you know, I probably got, I got too intellectual about it.”
Grimshaw: Applying Stripe's Playbook to Segment Was Detrimental
“At Segment, there was a lot of pressure, and we did also in a positive way, learn lots of lessons from Stripe, but we were very different to Stripe, even though we would develop a sole product. And too much of the Stripe lesson was Detrimental actually to bein…”
Grimshaw: Relying on vertical benchmarks can undermine startup innovation
“There's moments probably where those metrics are the wrong things, right? Like where benchmarking perfectly to the playbook of that vertical, like you're going to break it in some way. It's going to look different in some way.”
Grimshaw: Specialist VCs follow fixed playbooks while generalists seek new adaptations
“I think that the specialist can become a bit more physicist. Like here's the playbook, here's the rules. And I think of myself as a biologist in that you're constantly looking for like new adaptations, new variants”
Grimshaw: AI will yield new software architectures and business models
“And I think if AI is to be the force that it can be, I think you will get a new architecture and new business model emerging from it versus what we only see right now, which is kind of a sustaining architecture, which is just that of a co-pilot.”
Grimshaw: AI software will sell finished work outcomes rather than seat licenses
“The way to encapsulate would be this idea of selling the work, not the software. And that we'll move from a paradigm where you might think of it as moving from what we see right now as a copilot and moving to what I think about as a control center, where we'll…”
Grimshaw: Most software today is just a database with a form
“What's most software? It's a database with a form on top of it for users to manage information, put information, get information out.”
Grimshaw: Unlike AI, mobile was merely a UX shift with identical architecture
“And in that way, I think it will be very different to mobile, which was mostly another UX but the same architecture, the same business models.”
Grimshaw: Building AI co-pilots is an incumbent strategy
“Most of what we've seen has been co-pilot stuff, and I think co-pilot is an incumbent's strategy. Incumbents own distribution, they own data, they own the UX, and they own a business model that all aligns to a co-pilot.”
Grimshaw: AI startups building co-pilots won't create massive value
“And so I think the opportunity offered up to a startup being a co-pilot for something else, like probably won't be that amazing. And there might be pockets of it where it can really work, but the opportunity to disrupt is to be orthogonal to the incumbents.”
Grimshaw: AI Startups Win by Building Business Models Orthogonal to Incumbents
“There's so much wrapped up in, in, in the arc fundamental business As in company architectures and all the systems internally and all their operatings around that, that it's really hard to just change, right? That's why it ends up being asymmetric competition.…”
Grimshaw: AI Enables an Architectural Shift Equivalent to SaaS 1.0
“And I think it's this idea of this more all encompassing change in architecture and business model that AI enables in the way that the internet really enabled a total, you know, enable SAS one point or enabled a totally different architecture of product model …”
Grimshaw: Incumbents will remain focused on lucrative AI co-pilot features
“I think incumbents can have a lot of cake with a co-pilot idea right now, and that'll be the focus.”
Grimshaw: Founders will launch Benioff-style 'no more software' AI campaigns
“I think there'll be some entrepreneur or a few entrepreneurs, hopefully in the coming year or so, let's say, who channel maybe the Mark Benioff, 1999, like no more software, right?”
Grimshaw: Rapid market shifts can be dangerous for early-stage startups
“First of all, you don't need, as a startup, too much market or too much market changing hands really fast can be dangerous. As a startup, you could only take on so much.”
Grimshaw: LLMs will centralize media trust in established subscription brands
“I think as a function of LLMs, trust centralizes, and there's huge economies of scale to subscription businesses and advantages, and so it's a silly, silly fun idea, but if we're in media.”
Grimshaw: Benchmark invested in LangChain, Cerebras, and ex-PyTorch leads
“Obviously with investment in Langtrain and we have a few others at Benchmark Cerebrus, which is you know, an AI chip for training and inference. We backed a team that worked on and was the leads of PyTorch and has got a new company and some others.”
Grimshaw: AI applications will capture massive value and pull through infrastructure
“I think in the arc of time if it is a shift like the internet was in many ways the applications, Which might not look like an applications, right, in the way that we think of them, but might look more like, you know, this control center idea will have profound…”
Grimshaw: Early AI application sign-up spikes resemble 'high fructose corn syrup'
“Like, I think there can be a lot of high fructose corn syrup of adoption. Like, there's an amazing magical moment in a wow, right? When you go experience ChatGPT for the first time, or some of the imagery models for the first time. The rate of adoption of, lik…”
Stebbings: Enterprise experimental budgets for AI software are already churning
“I'm already starting to see actually the churn take place, which is like the experimental corporate budgets, which were like one or two seats, but they were at Walmart. You're seeing them churn already.”
Grimshaw: LangChain Currently Has More Capital Than It Needs
“At LangChain, obviously, like we have more capital than we need right now.”
Grimshaw: Tobi Lütke divested a Shopify acquisition within 18 months
“For Toby at Shopify to make an acquisition and then divest it With, I don't, I think within a year, maybe it was a year and a half, like not very long for an organization to say, to make a big move and say, nope, we made the wrong decision that fast and not ge…”
Grimshaw: VCs should prioritize founder respect over founder friendliness
“One is founder friendly. Like, I don't think you should be founder friendly. Not that you shouldn't be friendly, obviously. But I think that the higher purpose is like founder respect and friendly can imply like cheerleading, yes, person, et cetera. And I thin…”