Investing in People Years Before They Start Companies

Summary

Kyle joins Sam Silvershein on Alpha Partners’ Driving Alpha for a wide-ranging conversation on how Contrary generates alpha — anchored by the firm’s people-centric sourcing model: identifying exceptional people and investing in them years before they start companies. Kyle frames his own path as a “Goldilocks” tour of investing styles — TCV’s private-equity “diamonds in the rough,” Coatue’s hedge-fund “big-idea arbitrage,” and Index’s “venture classics” relationship-building — and contrasts all three with Contrary’s edge. Where his prior firms could only play “hurry up and wait” (meet sharp people early, then pounce the moment they raise), Contrary has built a flywheel of “SKUs” — a network, exposure to companies, scout checks, talent placement — that lets it add value across an entire career. The payoff: 5–10% of the first 50–60 employees at companies like Ramp and Base Power came out of the Contrary network, and by the time a founder starts a company, Contrary has often known them five-plus years and skips the fundraising process entirely.

The middle of the conversation is about writing as the engine. Kyle separates his two surfaces — Contrary Research, born from a need to market the firm without the “typical VC fodder” of self-congratulatory blog posts (a public research arm as a service, letting ideas “speak for themselves”), and his personal newsletter Investing 101, which he calls a “secret public journal” à la Mike Birbiglia. Prodded by Rex Woodbury at Index, he committed to publishing once a week no matter what — going on four years since 2022 — and the discipline forces him to “see the world in articulatable bites.” On AI, he’s a heavy Claude and Claude Code user (building a personal wiki à la Andrej Karpathy), but draws a sharp line from a debate with Jack Raines at Slow Ventures: writing does two jobs — thinking and saying. Where you only care about the output (math for a bookshelf, translating a language), let AI shortcut it; but where the “shape of your brain” is the point, outsourcing the thinking is the real danger — his fear is a generation ending up like the people in Wall-E.

On portfolio construction, Kyle reframes the “David vs. Goliath” narrative about megafunds as a strategic, not moral, debate — a sniper rifle versus a shotgun. His sharper framework is the “unholy trinity”: yield farmers (giant LPs who need to park $200–300M a pop at low replacement-rate bars), capital agglomerators (happy to raise ever-larger funds), and capital absorbers (companies whose default operating model is raising billion-dollar rounds and pursuing “world domination”). Elon Musk is the “patron saint” of the absorber archetype — even though SpaceX itself raised comparatively little (~$13B). The danger, Kyle argues, is that “to someone with a billion-dollar war chest, everything looks like a spend-money problem,” when most business problems aren’t — Quibi is his cautionary tale.

The last third turns to defense and durability. Drawing on The Anduril Thesis (three years of research, 100 years of military history), he explains why Anduril counter-positioned against decades of cost-plus contracting, invokes the Lindy effect (if an industry hasn’t changed in 100 years, there’s usually a reason — complexity or captured incentives, not excellence), and points to a new Contrary investment in liquid missile propulsion (Galadyne) as the “find the structural bottleneck, then counter-position” playbook. On the “defense bubble,” he returns to his core thesis: companies built on momentum-first, capital-absorption-first logic are often doomed because they chase a “pretend goalpost,” while durable companies (Tesla through the green bubble, Amazon through the dot-com bust, Anduril through the ChatGPT wave) are built around a deliberate worldview true independent of which way the wind blows. He closes on where Contrary is investing now — deliberately avoiding the “neo-lab-for-X rat race” — split into two buckets: (1) advantaged software with a hard-to-replicate ancillary layer (Carda Health’s in-home cardiac/pulmonary rehab; a privacy-first productivity suite unlocking sovereign AI), and (2) the physical world, which “has not had its ChatGPT moment” (American Housing Corporation’s heavily-automated in-factory homebuilding) — most of it flowing through people Contrary has known since undergrad.

Transcript

Sam Silvershein: Welcome to Driving Alpha, the podcast where outperforming investors share their insights and paths to success. This show is brought to you by Alpha Partners — we’re a growth-stage fund that co-invests alongside top VCs by sharing economics with early-stage investors in their most exciting companies. Today I’m thrilled to welcome Kyle Harrison, general partner at Contrary. Kyle’s venture career spans heavyweight firms like TCV, Coatue, and Index Ventures. He’s now focused on building out Contrary’s later-stage practice. Kyle is one of the sharpest minds writing about the mechanics of our industry today — through his newsletter Investing 101 and his leadership at Contrary Research, which includes a recent 300-page deep dive and a book, The Anduril Thesis. Kyle, welcome to Driving Alpha.

Kyle Harrison: Thanks for having me.

Sam Silvershein: You’ve said Contrary is your renegade of choice for building a later-stage practice. When you’re competing against massive funds, how does a people-centric platform — investing in top talent before they even start companies — fundamentally change how you source and generate alpha?

Kyle Harrison: There’s a bunch to unpack there. I joke that I’ve had a very Goldilocks experience over my career — I’ve tried a little bit of everything. TCV was a very private-equity mindset, at least when I was there: diamonds-in-the-rough hunting. Coatue was much more hedge-fund — big-idea arbitrage was the leading thesis there. And Index was a bit more venture classics. There are a lot of different styles you can compete against.

In terms of Contrary’s approach and how it distinguishes itself from everything else I’d done — one of the most significant things is that the people-centric model of venture can be very memeified. Everybody says “I want to go earlier and earlier,” people talk about sourcing in high school, whatever. But what it really means is building a relationship over time — and the ability to build that relationship requires having relevant “SKUs,” something to offer people even years before they start a company.

I used to joke that at Index and Coatue, we’d often meet really sharp people very early and then play the hurry-up-and-wait game: as soon as they put their hand up, we were ready to go — give them a term sheet, get excited. But before that, there’s not much you can do as a firm besides maybe invite them to a dinner once a quarter and keep them close. There’s no real product offering.

Contrary built the firm around this ethos of building an engine — a flywheel of all the different things we can do for someone over the course of their career. Even if they’re very early — undergrad, grad school, PhD programs — it’s exposing them to a network, getting them exposure to interesting companies, sourcing opportunities to invest, and being able to lead scout checks. It gives them a learning experience, and it lets us get to know them and better understand their judgment.

Then you go into the talent world — helping them find interesting roles and connect to other people. A lot of the companies we’ve invested in, like Ramp and Base Power — 5 to 10% of the first 50 or 60 employees came out of the Contrary network, in large part because we helped facilitate those relationships. So in many cases, when we invest in a founder, we’ve known them for five, six-plus years, and we’ve helped them at every major phase of their career. In some cases they don’t even run a fundraising process — they just come to us, we already know each other, and it works great. In other instances, the world keeps getting more competitive, so exposure to that broader talent network becomes a competitive edge — we help plug companies into it. When you think about the DNA that makes up the Ramps and Andurils and Base Powers of the world, it’s that pool of people we want to help companies plug into.

Sam Silvershein: You’ve got such breadth of network, and it shows in the research you publish. You’ve written before: “I don’t know what to think until I write what I say.” How does deep public research and writing help convert your thoughts into deal flow when you’re competing against other funds for the same opportunities?

Kyle Harrison: I shouldn’t steal credit — that’s a Flannery O’Connor quote, but it’s very much how I think about the world. I need to research, write, and articulate something before I understand how much I know and how much I don’t.

There are two sides of the coin for me. There’s my personal blog and essays, and then there’s the Contrary Research brand. They were born out of very different needs, but both are advantageous in a venture career — and being chronically online, as my wife would say.

Contrary Research was born out of needing to do more marketing as a firm. The first four or five years were very one-to-one relationship-building — finding people and building those relationships — but that’s not scalable. The goal was exposure to a broader set of folks. Eric and I — Eric, the founder of the firm — sat down and said Contrary needs to do a little more marketing. We’re pretty averse to the typical VC fodder of writing blog posts about how smart we are, so we thought: what’s something actually advantageous and additive to the ecosystem? We were already doing a lot of prepared-mind thinking — thinking about companies and markets, articulating our perspective on those categories, the inner workings, what matters. So we put that out into the world, and it lets the world vibe back. That’s one of the benefits of writing online: you put things out there and people reach out. We wrote a report on Stripe, and someone thinking about how Stripe’s business works says our memo was a helpful articulation of that strategy in their founding journey. Or someone deep in a market for a long time says, “It feels like you guys articulated this better than anybody else.” Letting our ideas speak for themselves — providing a public research arm as a service to folks — is a big part of why we do it.

On the personal-writing side, it’s much more — Mike Birbiglia the comedian calls his blog his “secret public journal,” and I think of it very similarly. It’s primarily for me. I actually don’t really care how many people subscribe or read it. It’s more a consistency exercise. When I started, Rex Woodbury — an investor I worked with at Index — pushed me to focus on one goal, and the goal I settled on was: publish something once a week, every week, no matter what. No matter how good or bad, how tired or busy I am. I’ve kept that goal since the beginning of 2022 — coming on four years of consistent weekly writing. What it forces me to do is see the world in articulatable bites. When I look at a market, a problem, an incentive — I hear about another VC’s misbehavior, or a new fundraise and what it means for the world — I don’t just think “interesting” and move on. I look at it and ask, “How would I write about that?” It forces me to chunk out how A leads to B leads to C, and I understand the thing better.

Sam Silvershein: You started in 2022. How has your writing or research changed with the AI tools we now have?

Kyle Harrison: I love Claude. I use Claude constantly, and I’ve started using Claude Code to build a personal website and personal wiki, à la the Andrej Karpathy LLM-wiki idea. But I had a debate recently with Jack Raines — an investor at Slow Ventures — a couple days ago, and the question is: writing does at least two very distinct jobs. It’s thinking and it’s saying.

It’s the same as when people say it’s only a matter of time until we’re all wearing AR lenses that translate languages into our brain via electric waves — what’s the point of learning a new language? There’s so much research that learning a new language, or complex math, isn’t just about being able to compute or translate. It’s literally the rewiring of your brain to think in a specific way, to be forced to wrestle with complex ideas. The most disadvantageous thing people can do is let AI replace that — where you say, “I’m not going to think about it, just give me this,” and it gives it to you, and you say “yep, sounds good” and move on, and there’s no thinking. That’s the most problematic thing.

But think about math and language differently: if I’m building a bookshelf and need to do some geometry to figure out angles, I’m not worried about what my brain will look like after I’ve built the shelf. If anything, I’d like my brain not to be thinking about building a shelf. So if something can do the math for me — great, I’m trying to do a thing, I just need to get it done. Language is the same: I’m just trying to communicate. If the shape of my brain before and after isn’t the primary consideration, AI is an incredible tool to shortcut unnecessary complexity.

The danger is that it becomes intoxicating. It’s very satisfying to say “complicated thing,” hand it to AI, and receive a non-complicated answer. You have to develop the ability to sit with discomfort and complexity — and writing forces you to do that if you do it the old-fashioned way. We’re constantly doing research, synthesizing a ton of sources, identifying where to go deep, pulling ideas from a ton of conversations — there’s so much AI can do, and it’s great. But I worry about people not taking any consideration for what the shape of their brain looks like today, let alone a couple years from now. Your brain becomes like the people in Wall-E, sitting in their chairs just zooming around. That’s the fear.

Sam Silvershein: That’s a great point — everything compounds. You’re constantly referencing old things you’ve written, and without that context the new information wouldn’t stick or fit in the puzzle. I want to talk about something you’ve written about that feels very relevant — the David-and-Goliath dynamic in this industry. How have these massive funds with seemingly endless capital changed how you construct your portfolio at Contrary?

Kyle Harrison: I don’t know that I’ve said it exactly this way — funny enough, I teach my six-year-old’s Sunday school class at church, and we just went through the David-and-Goliath story, so it’s been on my mind. I actually think David and Goliath is a poorly used analogy. What people mean is just “big versus small” — but the story is a moral debate. David represents standing up for a belief system against a threat.

The problem with applying it to what I’ve called capital agglomerators (the “Goliath”) versus cottage keepers (the “David”) is that I don’t believe there’s a moral debate there. It’s not that if you’re a capital agglomerator you’re a bad person, or you want to build crappy companies, or you’re filled with hubris and excess. It’s a strategic difference — choosing a sniper rifle versus a shotgun. It’s not a moral debate; it’s a strategic one. Often people have already decided where they fall on what they see as a moral debate — size versus specialty — and they hate the side that does the thing they think is wrong. I stop short of passing judgment. I really hesitate to say “larger firms equal bad things.” Larger firms can have bad incentives that create bad systems, but smaller firms can too. It’s really a strategic problem.

The more apt framework for the environment we sit in is what I’ve called the unholy trinity of venture capital. It’s not that large firms are inherently problematic — capital as a concept isn’t inherently bad. It’s that there’s an ecosystem feeding off each other that’s becoming problematic, because it’s being slammed against a form of creation that doesn’t scale the way this engine requires.

You have what I call yield farmers: LPs who represent massive amounts of capital — hundreds of billions of dollars, the CalPERS-scale public pensions. They need to park $200–300 million a pop to drive yield, and they’re not necessarily looking for venture-scale returns. They don’t need 3x, 5x, 10x funds — they need 6–7% replacement rates, sometimes 3–4% on even larger pools, with specific annual obligations. Their bar is much lower; they just need to park capital. Then you have capital agglomerators happy to oblige — taking and deploying larger and larger pools. And then you have capital absorbers: the standard way many companies now frame themselves as needing to be built. When you see $300 million seed rounds, or billion-dollar seed rounds or Series As — that massive capital absorption becoming the default modus operandi of company-building. The way you build a company is: demonstrate success, tell incredible stories, go after massive TAMs, raise a giant war chest, hire a massive number of people, and pursue world domination in your chosen field.

Elon Musk is the patron saint of the capital absorber. People love to look at his TAM ambitions — the $23 trillion of intergalactic B2B SaaS — and say, “We need to tell our version of that story.” What’s ironic is that SpaceX as a business did not raise ungodly amounts of capital.

Sam Silvershein: Secondaries, yeah.

Kyle Harrison: Totally. Last time I added it up it was something like $13 billion — chump change compared to what we’re talking about in some cases today. So Elon as a North-Star ambition-setter is the language they use, but the strategy is more defined by the Ubers of the world — raising tons and tons of capital. Uber, WeWork — that was the early generation of these companies, and those gave way to the OpenAIs and Anthropics of the world. Now every company down the stack wants to be that, and they think capital absorption is their tool. You have the unholy trinity lined up: capital agglomerators ready to shove capital in, and yield farmers letting them raise larger and larger funds. Bigger, bigger, bigger.

That’s not problematic at all if OpenAI and Anthropic are, number one, sustainable businesses in the long run, and number two, repeatable approaches to building a business. If you can do that consistently — people talk about the power law and say only a handful of companies matter, that’s fine — but there are hundreds of companies getting funded assuming they can be one of those, which is why they’re raising so much and burning significant money.

This is as close as I get to passing judgment: when the default mentality of how you build a business is shaped by capital absorption as a first principle, it’s very much “someone with a hammer, everything looks like a nail.” To someone with a billion-dollar war chest, everything looks like a spend-money problem. And the reality of most company-building is it’s not a spend-money problem. There are a handful of cases where it is — first-mover advantage, establishing network effects, getting a flywheel spinning, high capital intensity. There are a few business cases where spending more money is advantageous. But it’s not the majority. Most problems in business-building don’t get solved by throwing more capital at them. We have a dozen examples — Quibi — to look at and say: just because you spent $250 million and you’re one of the greatest people to ever come out of Hollywood doesn’t mean you can turn on product-market fit. I actually think Quibi could have crushed it, and could crush today — on the Disney+ app there’s a short-form feature where you swipe through clips of shows and movies to figure out what to watch, and I’m like, “That’s a really good idea, this is ahead of its time.” Quibi could have made it, but it didn’t, because the mentality was “it’s just a money-spending problem, we just need to spend more.”

Sam Silvershein: I’m curious what patterns emerge from that 2021 hype cycle — when everything was spent on growth marketing, and once the tide went out you realized a dollar in on marketing wasn’t a dollar out in revenue; it was really just “VC-fund-my-life.” As people start saying “bubble,” it’ll be interesting. But I want to drill in on high capital intensity and building in massive markets — let’s move to your book, The Anduril Thesis, which you said took three years to write. You mapped out a hundred years of military history — the bureaucracy, the cost-plus contracting that hurt innovation and drove out talent. What are the key takeaways for listeners looking at companies trying to rewrite the rules of entrenched legacy industries, like underwriting founders like Palmer Luckey?

Kyle Harrison: The book was an evolving, moving-the-goalpost thing. It shouldn’t really have taken three years — it was a function of us doing a lot of research investing in companies in this space. That’s where it started. Then we went super deep on Anduril, and we’ve done several long-form pieces — OpenAI, Stripe, Databricks — and you look at Anduril and say, this is a pretty unique setup. Every company has “heritage” they point to: OpenAI has a rich history of ML development over many years. But very few companies are built almost as the rebellious stepchild — literally, deliberately counter-positioned. They look at 50, 60, 70-plus years of military-industrial history and say, “This is broken, so we have to do this. And this is broken, so we have to do this.”

To understand why Anduril focuses on high-volume, low-cost assets — it’s because of all the cost overruns that forced concentration of capital into large, exquisite assets. That’s why they did it: because this was broken. Or why they spend on their own R&D — because you’ve got defense primes who built their entire business on overcharging the government to reinvent the wheel every time. Understanding all that helps you appreciate what’s required.

For companies trying to displace legacy industries, I think a lot about the Lindy effect — the idea that the longer something’s been around… People think, “This industry has operated the same way for a hundred years, like life insurance — it’s ripe for disruption.” Actually, there’s an evolutionary law: if something hasn’t changed in a hundred years, there’s probably a reason. If it sucked, it would have changed. Lots of things changed in the last hundred years because they sucked. So there’s usually a reason — and most of the time it’s not that it’s so good. It’s that it’s either complicated, or held hostage by the muddying interests involved.

In defense it’s a combination. There’s money in politics, jobs in congressional districts, things that keep incentives misaligned. But there’s also the evolving nature of warfare and conflict. What the world looked like post-World War II, in the ’90s, and today are three completely distinct worlds requiring distinctly different approaches to conflict. It’s hard to move such a big, lumbering machine along a rapidly evolving ecosystem. So as we underwrite these things — we just invested in a company called Galadyne, building a liquid propulsion system for missiles. That’s a very specific component of the supply chain; solid rocket motors are a critical input, a massively broken, structurally constrained bottleneck. They have a unique approach, and understanding why that bottleneck is the way it is — then counter-positioning to it — is where there’s real opportunity. People who go in and say “this is broken because people are stupid and lazy, we’ll just do it better” usually fail.

Sam Silvershein: Backing up — you were very early backers in Anduril. What did your LPs think when you started investing in defense tech?

Kyle Harrison: For better or worse, Contrary started as — Eric, the founder, was 23 when he started the firm. He built it as a solo GP for a while and then brought me on; I joined after he’d been slugging it out for four years. In the early days it was basically entirely entrepreneurs — people who’d built businesses investing in Contrary. The benefit of not having a super-institutionalized LP base is it let us be much more belief-driven than checkbox-y. It wasn’t that we were filling somebody’s IC checkbox — founders want to invest in firms that invest in people building great things. So we never had that problem.

Same thing — we were the first investor in Hallow, a large Christian prayer app that’s crushing it now. At certain times of year they beat OpenAI on consumer app downloads. Great business, but we invested in that company — the first time was I think 2017. Investing in religiously focused consumer applications was also a weird thing to do at the time. Different categories go through a vibe shift at different times. If Hallow were started today it’d be completely non-controversial, but at the time it was different. Defense was something a lot of early investors got angry messages or broken relationships over. One benefit is we just didn’t have to deal with that, because we didn’t have LPs saying “we don’t want you investing in specific-denomination stuff, or in weapons.”

I actually think it’s good that capital allocators wrestle with the implications of what they invest in. Not even defense-specific — you’re seeing this breakout in AI more broadly. There are basically two divides in technology: people who see an unbroken chain of globalization and believe it’s just meritocratic competition across the world, and people who acknowledge that China and the CCP operate in a very specific, confrontational way with the US — maybe not explicitly, but implicitly: our companies aren’t allowed in their markets, but their companies are allowed in ours. That’s an adversarial relationship even if we’re not in open conflict. So when you look at things like the Manus acquisition being blocked, Meta’s deals being blocked — that’s problematic and creates these issues. As capital allocators you have to decide: are you going to invest in companies with inherent or implicit ties to the CCP, or not? It’s good for people to have that debate within themselves about what they’ll allocate capital to.

Sam Silvershein: Eventually, as these valuations get bigger and companies scale, we’re going to have to figure out how to get into the Chinese market to justify multi-trillion-dollar valuations for AI businesses. Curious how that evolves. I want to go back to Anduril and Palmer Luckey. Defense tech was deeply unpopular when Palmer started Anduril — he was ghosted by engineers, investors, people he thought were friends. Fast-forward about ten years and one might argue we’re in a defense bubble. What are your thoughts on the evolving psychology of VC and the shift away from software, high-margin businesses toward more capital-intensive, potentially single-buyer markets?

Kyle Harrison: Palmer jokes that 2022–2023 was his “I told you so” tour. Post the Russian invasion of Ukraine, there was this acknowledgement… There’s a great story we tell in the book: in the early 1900s there was a book — a bestseller — all about the “end of history,” the idea that large conflict was inconceivable, that the countries of Europe were foolish to build up arms when the world was so interconnected that meaningful conflict was unthinkable. Then five years later, World War I started. It was the perfect example of “we have graduated past people willing to use violence to achieve their means” — we keep trying to convince ourselves that’s true, and we keep being proven wrong. It happened again in 2022–2023. Ukraine, Israel and Palestine — those conflicts shake people’s worldviews.

On the defense-tech bubble: this is another reason companies built with a momentum-first mentality — loud, massive buildout, raise as much capital as possible, sprint as fast as you can — are, to some extent, doomed to fail. Maybe not, but often it’s because they’re chasing a pretend goalpost. Think about Tesla. Early 2000s, there was the green bubble — John Doerr famously pivoted a massive amount of Kleiner Perkins capital and attention toward sustainability and clean energy, and the vast majority of those things didn’t pan out. Tesla was able to prevail through that bubble. Or Amazon — a quintessential dot-com. Bezos was a Wall Street guy who said, “The internet’s growing fast; what can I ship easily? Books.” A textbook example of a company that probably could have gotten washed out during the dot-com bust but survived.

Some of it is survivor bias — I’m sure there’s a myriad of companies built on good first principles that just couldn’t weather the storm, and that happens whether you’re in a bubble or not. But a good chunk is inherent in the fact that companies built with a specific worldview in mind — structuring their product and approach around that worldview — weather storms more effectively. There’s a great line from Palmer: about the ChatGPT moment, someone asks, “Is ChatGPT a huge deal for your business?” and he says, “Our products don’t use ChatGPT.” There’ve been a lot of evolutions in how capable LLMs are and how involved they are in Anduril’s products since. But in the early days of the ChatGPT fervor, it benefited Anduril’s business not because they were using GPT-3.5 in Lattice, but because people who rarely paid attention to cutting-edge AI were using ChatGPT in their everyday civilian life. That awoke them to the possibility — “wow, this automation, this AI, is really powerful” — which turned them on to Anduril’s approach, despite the fact that at that time the two didn’t have a lot to do with each other. So Palmer says, “I’m happy to take advantage of this as long as it comes, but I also know it’s going to come and go.”

The defense bubble is going to come and go. There will be a shifting of the tides. You already see different conflicts becoming heavily politicized — Iran, Venezuela. Conflict has always been politicized, but you’re seeing an uptick in discontent with specific conflicts. Administration changes will change the attitude toward defense buildout. Things come and go. But Palmer — and Elon with Tesla — has this same mentality: the way you weather the storm is by having a deliberate, believed worldview that’s true independent of which way the wind blows. If the wind’s blowing your way, great — take advantage of the tailwind. But as soon as it changes to a headwind, if you’ve built your entire ship so that the only way your engine works is with a tailwind, you’ve not built a durable business.

That goes back to my core problem with capital absorption as a first principle: it’s a very dangerous way to build a business, because most people don’t have a deliberate worldview of what their product should be in the long run. They just have a sense of the game they should be playing — the fundraising game, the markups game, the big-press-release-partnerships game — but no worldview they’re building toward. That typically ends problematically. It happened in 2021 with crypto companies — there wasn’t really a worldview behind those companies. It was a moment in time.

Sam Silvershein: They sold a nice story. Web3, crypto — a view for the moment in time, like you said. I want to end on this: you wrote about the “hero generation” of VCs — those building instead of extracting, trying to build things worth building in the biggest markets — and you were looking to fund them. What areas of the market are you most excited about today?

Kyle Harrison: The core hype cycle we’re avoiding is the neo-lab-for-X rat race — anybody who leaves a major AI research lab and starts their own new bespoke research lab for a different focus area. That just doesn’t make sense for us to play in, in part because we’re not deploying billions of dollars every year, and we want to make eight investments a year, not 40 or 50. We generally avoid that knife fight, with the occasional exception.

Our work has bifurcated into two distinct directions. One: we’re ardent believers that a lot of software can be disrupted — existing software companies can be easily disrupted. But software as a business and a product does not die just because AI can write it more quickly. There are advantaged ways to build software; it’s more about building an ancillary component around that software that’s very difficult to replicate. We have a company called Carda Health — a digital-health company providing in-home therapy for cardio and pulmonary rehab. The company is Claude, AI, end-to-end — tons of things they’ve done to advantage their business — but most people wouldn’t describe it as an AI-native company. It’s not like they have AI agents replacing therapists. But that company is 3x-ing to almost $100 million of revenue — growing massively, doing very well, and doing it profitably. Building a network of therapists is difficult; building deep relationships with insurers as distribution partners is difficult. But once you’ve built those things, software is an advantaged distribution mechanism to deploy your product.

We have another company doing privacy-first productivity — basically rebuilding messaging, email, and docs from scratch with an almost Signal-chat-esque protocol around privacy. They can sell into highly regulated industries — defense, healthcare, financial services — and the benefit is it’s a foundational layer that unlocks sovereign AI in the long run. Those companies aren’t mind-bogglingly complicated per se; they sound like software companies, but they build something very specific and complex.

The second bucket: we really buy into the idea that the physical world has not had its ChatGPT moment. Despite what people say about robotics and humanoids, the physical world is still massively complicated and difficult to traverse. Companies building specific, capable engines around physical-world processes are very advantaged in our view. One of our companies is American Housing Corporation — building in-factory housing units that aren’t cost-optimized cheap modular housing, but very high-quality, well-designed, capable building projects that are heavily automated and can dramatically reduce the cost of construction. That’s a difficult, complex, messy environment, and they’re tackling it head-on. Those types of companies have a real advantage. So we’ve split at the labs level and gone either into the software or the hardware layer — really, how do you get this thing deployed into a messy, complicated world?

Sam Silvershein: And how do you build moats around it. I assume a lot of those folks come through your pre-existing networks — you may have known them since undergrad, and they’ve scaled within the Contrary ecosystem, spun out, and now you’re able to invest.

Kyle Harrison: Yeah, exactly.

Sam Silvershein: That’s awesome. Kyle, I really appreciate your time today. Thank you so much for joining the Driving Alpha podcast.

Kyle Harrison: Thanks for having me. This was great.

Connections

The firm & its people

• Contrary — the multi-stage firm Kyle joined to build out the later-stage practice; the “renegade of choice.” Its people-centric edge is a flywheel of “SKUs” that adds value across a person’s whole career, years before they start a company. See Contrary — My Renegade of Choice.
• Contrary Research — the public research arm, born from a need to market without “typical VC fodder”; builds the “prepared mind” and lets ideas “speak for themselves.” Home of the 300-page Anduril deep dive.
• Eric Tarczynski — Contrary’s founder; started the firm at 23 and ran it solo for ~four years before bringing Kyle on. Credited with the belief-driven, entrepreneur-heavy LP base that let Contrary back defense and faith-based companies early.

Kyle’s path & prior firms

• TCV — private-equity “diamonds in the rough” training.
• Coatue — hedge-fund “big-idea arbitrage.”
• Index Ventures — “venture classics,” relationship-driven; where Rex Woodbury pushed Kyle to commit to publishing weekly.

Writing & AI

• Investing 101 — Kyle’s newsletter; his “secret public journal” (à la Mike Birbiglia), published once a week every week since the start of 2022. “I don’t know what I think until I read what I say.”
• Flannery O’Connor — source of the write-to-think line Kyle credits.
• Claude / Claude Code — Kyle’s daily AI stack; building a personal wiki à la Andrej Karpathy. His frame: writing does two jobs — thinking and saying; outsource the “saying,” never the “thinking.”
• Jack Raines (Slow Ventures) — the sparring partner for the thinking-vs-saying debate.

Portfolio & companies referenced

• Ramp · Base Power — 5–10% of the first 50–60 employees came from the Contrary network.
• Anduril — Contrary’s deep research subject and the book’s spine; the counter-positioning-against-broken-incumbents model; ChatGPT awakened buyers to AI even though Anduril’s products “don’t use ChatGPT.”
• Palmer Luckey — Anduril founder; his 2022–2023 “I told you so” tour; the “deliberate worldview true independent of the wind” thesis.
• Hallow — Contrary’s first check (~2017) into a Christian prayer app, now beating OpenAI on consumer downloads at times; an early “weird” bet the LP base allowed.
• Galadyne — new Contrary investment in liquid missile propulsion; the “find the structural bottleneck (solid rocket motors), then counter-position” play.
• Carda Health — Contrary’s digital-health company (in-home cardiac/pulmonary rehab), 3x-ing toward ~$100M revenue; advantaged software with a hard-to-replicate services + insurer-distribution layer.
• American Housing Corporation — heavily-automated in-factory homebuilding; the “physical world hasn’t had its ChatGPT moment” thesis.
• Stripe · Databricks · OpenAI — companies Contrary Research has published long-form work on.

Capital & market structure

• The Unholy Trinity of Venture Capital — Kyle’s central frame here: yield farmers (giant low-bar LPs like CalPERS), capital agglomerators (ever-larger funds), and capital absorbers (companies whose default is raising and burning huge rounds).
• Capital absorption as a first principle — “to someone with a billion-dollar war chest, everything looks like a spend-money problem”; most business problems aren’t spend-money problems.
• Elon Musk — the “patron saint” of the capital absorber for his TAM ambitions — even though SpaceX raised comparatively little (~$13B).
• Uber · WeWork — the earlier capital-absorber generation, now succeeded by OpenAI / Anthropic.
• Quibi — the cautionary tale: $250M and Hollywood pedigree can’t “turn on product-market fit”; a good idea (short-form discovery, now on Disney+) that lost by treating it as a money problem.
• Tesla (green bubble) · Amazon (Bezos, dot-com bust) — durable companies built on a worldview that survived their era’s bubble; contrast with momentum-first companies chasing a “pretend goalpost.” John Doerr / Kleiner Perkins cited on the green bubble.

Concepts & themes

• David vs. Goliath (as analogy) — Kyle reframes it as a strategic, not moral, debate (sniper rifle vs. shotgun): capital agglomerators (“Goliath”) vs. cottage keepers (“David”).
• Lindy effect — if an industry hasn’t changed in 100 years, there’s usually a reason (complexity or captured incentives), not excellence.
• Counter-positioning — great companies build counterintuitively against a broken status quo (Anduril vs. the defense primes and cost-plus contracting).
• The Anduril Thesis — the three-year, hundred-years-of-military-history book underpinning much of the defense discussion.
• Defense bubble / momentum-first building — bubbles “come and go”; durability comes from a deliberate worldview, not tailwinds.
• China / CCP — the adversarial-relationship divide capital allocators must wrestle with (Manus blocked, Meta deals blocked); sovereign AI as the long-run stake.