Miles Grimshaw | Dialectic Summary

1. Key Themes

Hunting Change as the Core Investment Philosophy

Miles frames change not as a risk to navigate but as the very oxygen that creates investment opportunity. New companies form precisely because incumbents are caught off guard or fail to serve new demand.

"Change is the oxygen for new companies to get formed because incumbents are caught off guard or incumbents don't serve something well. There's a new demand. There's a new infinite game to be played." [00:00:00.000]

The Sisyphus Mindset: Loving the Return to Zero

Miles explicitly embraces the venture investor's structural disadvantage — that every major outcome forces you back to square one — as a feature, not a bug. He frames it through Camus's happy Sisyphus.

"I love going back down and looking at new summits and going, can we go do that one? And these are the founders' summits and their conquests. But I like to think of Thrive and all of us as joining the expedition team of those pursuits and going back down blank slate almost, you know, at the early stage and dreaming of a new challenge again." [00:10:13.080]

Where the Founder Is Looking Matters More Than What They've Built

Miles's most distinctive early-stage diagnostic: ignore the current product and ask where the founder's gaze is fixed. The IDE was irrelevant; the abstraction of coding was everything.

"In driving and biking, you go where you look. Any sport you're learning, the thing you learn is look where you want to go. And I think it's really true for companies too. You might see only a bit of it. Where's that founder looking? Michael of Cursor, he wasn't looking at the IDE. He was looking at abstractions of coding making coding easier. The IDE was the manifestation of it today, but he wasn't looking at that. He was looking over there." [00:00:27.040]

Early Customers Are Genetic Pressurizers, Not Revenue Units

The quality of early customers matters far more than the ARR they represent, because demanding customers force adaptive fitness in the product and organization. Stripe with Shopify versus Stripe with Delta is a completely different genetic outcome.

"I care far more that you've selected and are working with really thoughtful, avant-garde, demanding customers than just your revenue. Because those are pressurizing functions for you to adapt your fitness, your product, your organization, et cetera, to be well adapted to the most demanding future." [00:51:44.860]

The Wall Clock Time Moat in Software Is Gone

The traditional defensibility of software businesses — the sheer accumulated hours of building a feature-rich "city" — has been largely eliminated by AI. This structurally changes what kinds of software companies can be durable.

"What's challenging about what's happened in AI for software is the wall clock time's kind of gotten eliminated. And so you can kind of build that city. You might not have any residents yet, but you can build that city. If it took a day to build a city, we'd have more cities for sure." [00:57:54.690]

The Barbell Future of Software: Amazon-Scale or White-Glove Luxury

The mid-market "Radio City Music Hall" software company is being hollowed out, just as mid-market retail was by e-commerce. Winners will either go very broad with intelligence at the center, or go extremely bespoke and high-touch.

"I think you're going to need to do a lot because the wall clock time changed. You should have way more surface area. You should cover way more. You'll make less margin on it all... Or maybe even Salesforce, right? But you need to put intelligence at the center. You need to totally reimagine that too for the modern age." [00:59:51.910]

"I think more companies now should say yes to one customer's needed feature. And if you, A, can't get that ship fast, you're probably not using the tools totally effectively." [01:02:04.690]

Impatiently Patient as the Core Investor Tempo

Miles distinguishes between two orthogonal time horizons: relentless, never-off proactive learning and relationship-building (impatient), combined with genuine willingness to wait years for the right organic moment to invest (patient).

"Our challenge is to be impatiently patient. And the impatiently is to be constantly learning, to be constantly exploring in the world, meeting and supporting new amazing people... And patient to wait for the right moments, the ones that are organic and fit best for us." [00:18:00.940]

"I Don't Know" Plus "I'll Figure It Out" as a Superpower

Miles identifies intellectual honesty about ignorance combined with a bias toward figuring things out as the most important disposition for investors. He ties this to a story about a legendarily demanding public market investor who screened specifically for the willingness to say "I don't know."

"The three most important words for an investor are I don't know. When that very esteemed older investor would interview people, he'd give a case study and the main thing he was looking for was at some point you'd say, I don't know... And the four really powerful words after that are, I'll figure it out." [00:00:55.360]

Craft and Attention to Detail Build Trust Asymmetrically

Craft in product is not purely aesthetic — it functions as a trust signal that compounds into customer willingness to rely on you for things they cannot directly evaluate. Patrick Collison articulated this most clearly.

"Patrick Collison said this once... the reason you should care about craft and attention to detail is that your customers instead of seeing and feeling that will trust you more because they'll know that you really paid attention." [01:05:16.750]

Being Demanding in Good Times, Supportive in Bad Times

Sir Michael Moritz's board member framework stuck with Miles and deepened with experience: the counterintuitive insight is that good times are when pressure-testing and raising ambition is most productive, not when one should ease off.

"A great board member is demanding in good times, is supportive in tough times... The reason you need to be demanding in good times is because that's when you can level up to the next area. It's when you have the capacity of the next amazing recruiter." [01:27:34.960]

2. Contrarian Perspectives

The Business Model of the IDE Was Terrible — and That Was Irrelevant

Conventional investors would screen out an IDE company on market size alone. JetBrains, the best-in-class prior, barely reached ~$200M in revenue. Miles's view is that first-principles analysis of what the product could become (intelligence consumption) completely overrides the inherited market ceiling.

"If you look at the IDE market and you said, what are the genetics of this IDE kind of thing? Like it's a terrible market. It's mostly free. The company that had done the best on it, JetBrains, maybe made a couple hundred, like 200, maybe million of revenue. It's not a business you want to be in. But if you get more first principles about it, you go, what if intelligence is at the core of that IDE? You actually now may have a product that has a thousand dollars — it's now actually like $10,000 of value consumption in that experience." [00:45:38.900]

Bespoke Software, Long Taboo, Is Now a Competitive Advantage

The entire SaaS era was built on the assumption that bespoke builds were economic landmines due to maintenance costs. Miles argues this assumption has inverted — the ability to ship and maintain custom features rapidly is now a signal of architectural health, not a liability.

"Building bespoke, if you did that, was a minefield because the maintenance on all of that over the years ahead was going to be brutal... But that assumption suddenly changed. I actually think more companies now should say yes to one customer's needed feature. And if maintaining that slight version of that end of one feature is a minefield for you, your code base is probably not adapted to literally the ephemerality, disposability, ease of building software." [01:02:04.690]

Don't Optimize for Revenue in Early Customers — Optimize for Genetic Pressure

The received wisdom is that early ARR validates product-market fit. Miles argues the identity of the customer matters far more than the revenue, and that chasing revenue from the wrong customer can set permanently bad genetic direction.

"I think Stripe would be a very different company if Delta was its first customer, not Shopify." [00:52:12.260]

Most People Learn the Wrong Lessons from Legendary Companies

Google's 20% time is celebrated as an innovation practice to emulate. Miles points out it only worked because Google had the best business model in history as a substrate, and copying the practice without the substrate is cargo cult behavior.

"Google was so successful in the core thing that in some sense, no matter what they did next, didn't really matter... People tried to copy them for this idea of 20% time... But also the best business model in the history of the world to do it. But the corollary to where I frame that is, oh, you can only work four days a week." [00:35:39.660]

Software Is Not Over — Change at This Scale Creates More Opportunity, Not Less

The conventional fear is that AI commoditizes software and destroys venture returns. Miles's contrarian view: high rates of change are precisely when new company formation is most fertile, because incumbents cannot adapt fast enough.

"I don't think so, in part because there's a high rate of change right now. And change is the oxygen for new companies to get formed because incumbents are caught off guard or incumbents don't serve something. There's a new demand... Why is there such entrepreneurial zeal right now? Because there was change." [01:08:42.510]

3. Companies Identified

Cursor AI-powered coding IDE, acquired by Anduril (originally described in intro as sold to SpaceX — likely a transcript error for Anduril). Built by four MIT founders, went from zero salespeople through tens of millions of ARR, developed their own tab model in-house. Led by Michael.

"It was like four MIT 20-year-olds who were going up against Microsoft in their like bread and butter developer IDE mainline pursuit... and did it better and created amazing magic for the world from that." [01:10:18.070] "Cursor's tab model was their own. And so they were full stack from the get go, from very early on." [00:49:56.420]

Turbo Puffer Vector/semantic search database company based in Ottawa, Canada. Founded by Simon, bootstrapped to near-profitability before taking outside funding. Cursor was their only early customer. Thrive and Lockie are the only outside investors.

"I'd heard about it because Cursor was using it as sort of underlying semantic search and storage... They had Cursor as an only customer. I did one quick call with Simon and he was wildly impressive just on that call. And we said, hey, we'll be in Ottawa next week for dinner." [00:15:22.260]

Benchling Life sciences R&D cloud platform. Miles has been helping with recruiting for years, including close to the time of the interview.

"I'm still helping Sajir at Benchling, like recruit and close, close, close, close to handle or keep people." [00:22:42.980]

GitHub Developer platform. Listed as a key early Thrive investment in software.

"That curiosity and love of learning there translated to me meeting Will Gabrick and then Will joined Thrive and they said, you should come join us... that led us to Segment and GitHub and other things." [01:31:55.820]

Stripe Payments infrastructure company. Cited multiple times as exemplary — for its early customer selection (Shopify), its genetic trajectory (Patrick Collison looking at "GDP of the internet"), and Thrive's ability to step in during the post-ZIRP drawdown.

"You talk with someone like Patrick Collison and he might not have had the words right then, but he kind of has the framing of GDP of the internet. And that will be expansive in nature. And what they have done so far is like a tiny fragment of where they're looking." [00:42:21.100]

Chai Discovery AI for drug/molecule discovery. Listed as a key Thrive portfolio company.

"He has spent most of his career there investing in mostly software companies, including Cursor, Chai Discovery, Turbo Puffer, Benchling, GitHub, and many more." [00:01:44.940]

Monzo UK digital bank founded by Tom Blomfield. Achieved approximately one-tenth of the UK population as daily checking account users.

"Tom Blomfield who founded Monzo... went and said, we could build a better bank in the UK. And that is a nuts thing to say... something like a tenth of the UK population has it as a daily checking account." [01:10:49.550]

Mesh Optical Optical transceiver / photonics manufacturing company founded by former SpaceX engineers who built the laser interconnect for Starlink. Thrive invested within a long weekend of their first meeting.

"Two guys, three people who'd left SpaceX, who'd worked on the laser interconnect for Starlink so that all the sats could communicate data across each other... We met them basically the week they left... by Monday morning, we'd agreed to work together in an area, optical transceivers, that we had absolutely zero business knowing anything about." [01:15:16.780]

Palantir Data analytics and AI company. Miles cites them as the archetype of the "white glove / luxury" software model — bespoke, high-touch, high-value — and as a model for what survives the commoditization of software building.

"I have a ton of respect for Palantir. And obviously they're doing remarkably well. In some sense, maybe they're luxury... They are the original forward-deployed engineers." [01:01:15.350]

Shopify E-commerce platform. Cited as the ideal early customer for Stripe — demanding, developer-led (Toby Lütke), and scale-pressurizing in ways that shaped Stripe's architecture permanently.

"Stripe would be a very different company if Delta was its first customer, not Shopify." [00:52:12.260]

Segment Customer data platform. Cited as an early Thrive software investment alongside GitHub.

"That led us to Segment and GitHub and other things." [01:31:55.820]

Oscar Health Health insurance company founded by Josh Kushner. Cited as one of the early examples of what greatness looks like that calibrated Miles's pattern recognition.

"Josh building Oscar was like really powerful for just seeing how great can be and what amazing success can look like." [01:32:26.760]

Instagram Photo sharing social network. Listed as an early Thrive success under Bobby Parker that helped calibrate Miles's sense of greatness.

"Bobby Parker, Harry's Instagram was a great success." [01:31:55.820]

Braintree Payments company, cited as a company that early on looked phenotypically similar to Stripe but had fundamentally different genetic trajectory.

"You can take Stripe and Braintree early on, kind of look the same, but you talk with someone like Patrick Collison and he kind of has the framing of GDP of the internet." [00:42:21.100]

HubSpot Marketing/CRM platform. Mentioned in connection with David Skok's SaaS business model research at Matrix.

"He was studying this business model and thinking about this business model and working with HubSpot and some others." [01:31:24.960]

4. People Identified

Michael (Cursor CEO) Co-founder and CEO of Cursor. MIT graduate. Described as relentlessly product-focused, running through tens of millions in ARR with zero salespeople. Miles spent roughly a year getting to know him before the Series A.

"Michael was relentless on his own team's feedback on that product. They are the customer of the product... They had a very strong purity to delivering a great product." [00:37:50.600]

Simon (Turbo Puffer founder) Founder of Turbo Puffer, Ottawa-based. Bootstrapped the company to near-profitability before taking any outside capital. Described as wildly impressive from a single call.

"I did one quick call with Simon and he was wildly impressive just on that call. And we said, hey, we'll be in Ottawa next week for dinner." [00:15:22.260]

Patrick Collison Co-founder and CEO of Stripe. Cited as the exemplar of a founder whose vision (GDP of the internet) far exceeded the visible product at any given moment, and as someone who articulated why craft matters for trust.

"Patrick Collison said this once... the reason you should care about craft and attention to detail is that your customers instead of seeing and feeling that will trust you more." [01:05:16.750]

Tom Blomfield Founder of Monzo, the UK digital bank. Cited as an example of a founder Miles was proud to back — someone who said they could build a better bank and then actually did.

"Tom Blomfield who founded Monzo... went and said, we could build a better bank in the UK. And that is a nuts thing to say." [01:10:49.550]

Toby Lütke CEO of Shopify. Cited as a "developer's developer" whose technical rigor and demanding nature made Shopify an ideal early pressure-testing customer for Stripe.

"Toby is a developer's developer. And so you know he's going to be understanding and pressurizing. Mounting you." [00:54:09.470]

Sir Michael Moritz Partner at Sequoia Capital. Credited with the framework that Miles says shaped his board member philosophy: demanding in good times, supportive in tough times.

"A great board member is demanding in good times, is supportive in tough times." [01:27:34.960]

Eric Vishria Partner at Benchmark. Praised by Miles for his ability to connect theory to operations and drive execution within companies.

"Eric Vishria I think is remarkable at sort of connecting theory to like operations and change and execution within a company." [01:40:15.340]

Josh Kushner Founder of Thrive Capital, founder of Oscar Health. Credited with the philosophy of betting on willpower over proven ability — including in Miles himself — and creating the culture of leaning into young, earnest investors.

"Josh really leaned into people who had a lot of will, even if they didn't necessarily have any proven ability... I'm extremely grateful for Josh and Will and Chris and Jared for leaning into a lot of willpower and earnest curiosity." [01:32:26.760]

Jared Weinstein Partner at Thrive Capital. Credited Miles with the description "Miles was a punk," which Miles takes as a high compliment about his fanatical, selective focus.

"A friend of the pod, Jared Weinstein says, Miles was a punk. And I say that in a complimentary and long on Miles way." [00:01:21.020]

David Skok Partner at Matrix Partners. Author of the "For Entrepreneurs" blog, which Miles credits with giving him a foundational understanding of SaaS business models while in college.

"There's a great blog written by David Skok who is a partner at Matrix called For Entrepreneurs that was basically all about the SaaS business model... He was studying this business model and thinking about this business model and working with HubSpot and some others." [01:31:24.960]

Will Gaybrick Partner at Thrive Capital. Miles credits Will with recruiting him into Thrive and helping establish the early software investing direction.

"That curiosity and love of learning there translated to me meeting Will Gaybrick and then Will joined Thrive and they said, you should come join us." [01:31:55.820]

Jack Altman Described as a good friend of Miles's from his time at Benchmark, characterized as an entrepreneur-minded investor.

"My good friend Jack Altman who is sort of an entrepreneur..." [01:40:44.960]

Ray Dalio Founder of Bridgewater Associates. Miles was drawn to Bridgewater not for macro trading but for Dalio's deep thinking on organizations and people — the Principles — even before they were publicly published.

"Ray Dalio's principles hadn't been published, but they were sort of an internal thing. And when I spent time on the team, there was so much curiosity for team organization." [01:29:56.280]

5. Operating Insights

Recruiting as the Internal Gut Check for Investment Conviction

Miles uses a single consistent test to check whether he genuinely believes in a company: can he vividly imagine and emotionally commit to recruiting for it, both tomorrow and ten years from now? If the pitch to the first recruit isn't compelling in his bones, the investment conviction is probably not real either.

"One of the cool questions I asked myself, like my gut check on any investment, is imagining recruiting for it... I'm thinking about the pitch to the first recruit this year, you know, tomorrow. And the recruit 10 years from now." [00:22:23.260]

Investment Committee as Music Studio, Not Jury Trial

The default mental model for investment committees is adversarial — one person defends, others poke holes. Miles argues the productive model is collaborative music-making: everyone comes prepared but the goal is to find the magic together, not render a verdict.

"The best version of investment team conversations are not a jury trial. It's like being in the music studio, making music together... We're kind of in the studio, jamming, riffing, figuring it out, of which yes, we came prepared and we all had sheet music... But we're trying to get at the magic of this, the right answer to this and figuring it out." [01:35:20.040]

Use "What Is It?" and "Who Cares?" as the Grounding Questions

Two simple questions cut through complexity and marketing language to the atomic unit of any business. These are used at Thrive as a standing practice across the team.

"I have two silly but true questions that are have become a bit of a meme at times at Thrive, which is one, what is it? And two, who cares? Sometimes I say who the fuck cares... What really is it? What is a customer really buying? What is the thing that it is doing for them? And two, like, who cares? Who is that customer?" [00:34:41.390]

When Things Are Going Well, Lean In Harder — Not Back

The natural tendency of investors and board members is to relax when a company is succeeding. Miles argues the opposite: success creates capacity that expires. The time to recruit the next great executive, raise the bar on vision, and push harder is exactly when things are working, not when they're broken.

"The times when usually less attention is given is because it's working. And it's actually almost the time when one needs to try and lean in a little more because it is the time to do that." [01:28:31.220]

Stay Close to People You Respect Even When You Can't Back Their Company

Miles explicitly tracks and maintains relationships with founders he deeply admires but couldn't get conviction on. This is not just networking — it generates the warm introductions that produce the most time-sensitive, highest-quality deal flow, as demonstrated with Mesh Optical.

"A founder who I never backed, but deeply respected and admired and tried to be friendly and a resource for... I didn't get conviction in the product and the company he was building, but he was a star. So we stayed very close and he was kind enough to introduce me to the Mesh Optical team the first week they left SpaceX." [01:15:50.400]

6. Overlooked Insights

Cursor Secretly Built Its Own Model — and That Changes Everything About What It Was

Almost everyone discussing Cursor frames it as a great product wrapper around frontier models. Miles briefly drops a fact that fundamentally reframes Cursor's nature: they built their own tab model in-house from very early on. This means Cursor was always a full-stack AI company with model development capability, not merely a product layer. That genetic trait — understanding the science of the model AND the craft of the product — is almost certainly what made it attractive enough for a major acquirer.

"People forget this, but Cursor's tab model was their own. And so they were full stack from the get go, from very early on. And obviously the labs have done incredibly well as the models have gotten a lot better. But the view, which Cursor is still certainly proving out, of the product model combination to unleash the maximum of that intelligence and the team that cared and thought a lot about the craft, you know, and the science of the model." [00:49:56.420]

Thrive Is Making Its First-Ever Hardware/Semiconductor Investment Right Now

In a single passing sentence, Miles reveals that after over a year of studying the AI infrastructure supply chain — an area entirely outside their historical competence — Thrive is at the moment of the interview actively closing their first investment in that category. He gives no name, but this is a significant strategic signal: one of the most successful software-focused venture firms in the world is quietly pivoting a portion of its attention into the physical AI stack. This is not a casual exploration; it followed sustained, uncomfortable domain learning.

"We've sort of been impatient with ourselves to learn about that. Like really, really pursuing it, but hadn't made any investment basically up until now. And we're making one investment at the moment, our first investment there." [00:19:29.620]