Sandy Kory

In my 12 years as an angel investor, I put 60%+ of my post-tax income into 100+ investments. I'd do it again over investing in public markets every time. History tells a clear story here: a diversified stock portfolio, held long enough, does well. Often very well. I don't think there's any arguing that, and for most investors, it's probably the right call. But with public markets, you have no impact on the companies you're in and no real relationship with the people running them. That was always a sticking point for me. With startups, even a small check can mean something, though it depends on the dynamics—you're probably not moving the needle with a $5K angel check in some giant startup. But in the right situation, you can have influence and be genuinely helpful. Even more gratifying, you can build meaningful relationships with amazing founders who are truly inspiring. And beyond that, I was always fascinated by the learning and the opportunity to transform that into alpha. Where some investors don't even open their investor updates, I don't think I ever skipped reading a single one. I always thought that every update was training for the "AI" between your ears, which (if you're doing it right) is ultimately what drives returns. I've always been thirsty for that information. There's also the fun of it. I guess if you're a billionaire investing in Apple today, maybe you can build a relationship with someone up there—but otherwise it's pretty hard. With startups, you're there from the beginning—and being a first believer is something founders actually remember. And then—when things go well—there's the magical feeling you get from being early into a startup that actually changes the world. That feeling goes hand-in-hand with the alpha, which is also pretty nice. In those cases, it's really a win-win-win. (Pictured with one of the founders I'm fortunate to work with, Oguzhan Atay.)

We're in a bull market for distractions. If you're building a startup right now, the Theory of Constraints is the best framework I know for cutting through the noise. When markets tank, layoffs spike, and social media becomes a wall of anxiety, there's a temptation to react to all of it. The Theory of Constraints (TOC) says don't. It was originally conceived for manufacturing operations, but the core idea applies far beyond the factory floor: every complex system has a single weakest link. Find it, fix it, and the whole system improves. Leave it alone, and nothing else you do matters much. This resonates with me personally more than almost any other framework. There are always a hundred things on my to-do list, but the TOC forces a clarifying question: which one of these is actually the constraint? That's where the focus should go. For early-stage software startups seeking tight PMF, it's particularly useful. I'd argue the key goal at this stage is having customers happy enough to refer their friends and colleagues. And getting there is a system of linked problems, with one acting as the bottleneck at any given moment. That could be: → Onboarding → Time-to-value → Scalability The constraint isn't always obvious. But the process of searching for it—really interrogating where the system is breaking down—tends to surface the tradeoffs that matter most in product development. That kind of focused dialogue is, in my experience, what separates teams that execute well from teams that stay busy. Great execution at early-stage startups is almost always only visible in hindsight, which is why advice like "execute like Airbnb" or "execute like Canva" isn't particularly useful. You can't see the 1000s of small prioritization decisions that got them there. The TOC won't make you execute like Canva. But it'll stop you from working hard on the wrong thing, which is a large part of the battle.

In 2007, a close friend told me he was joining an early-stage company targeting government and financial services with 2 distinct products. I, ever the wise ass, told him that was a dumb strategy. That company was Palantir, and I was very wrong. But the reason why wasn't so simple. Their hedge fund product eventually died on the vine. The government product went everywhere and became the foundation of everything Palantir is today. So the brilliant, coherent strategy people now admire? It didn't really exist. What looked like a 2-pronged plan became a 1-pronged business after 1 prong failed. And yet Palantir became one of the most remarkable companies in the history of enterprise software. What I missed in 2007 was that my friend was joining Palantir because of the people, not the strategy. He'd been around serious talent his whole career, and he said he had never seen anything like the caliber of people assembling around Peter Thiel. I think about this a lot when founders ask me what they can learn from Palantir. There's a real irony in that question. Palantir is arguably the most contrarian company in tech—a company that built its culture by doing the opposite of what everyone else did and ignoring conventional wisdom at almost every turn. And yet the way founders try to emulate Palantir is almost always about copying the surface-level features, like their public profile. The culture—the extreme talent bar, the mission-driven posture, and the willingness to be contrarian when it wasn't cool—was there from the beginning. The bombastic public profile only came 15 years later, after the culture had compounded into something that could support it. Founders who look at Palantir today and try to reverse-engineer what they see are studying the output of 2 decades of culture, not the input. If I'd been paying attention to what really mattered in 2007—the quality of the people my friend was so impressed by—I might have understood what was actually being built. (Fortunately, a few years later, in the darkest days of the GFC in spring of 2009, I had a chance to invest. I honestly still couldn't figure out their strategy. But I had come to appreciate that Palantir was one of the most talent-dense and mission-driven groups of humans on the planet. The world seemed like it might be about to end, but this was the group most likely to thrive in any circumstance. I put about 1/3 of my net worth into the company.) Strategy is legible. You can explain it in a slide. Culture is harder to see, which is probably why it gets less credit than it deserves. But at Palantir, culture was the whole game from day one.

@armandcognetta Congrats @armandcognetta!!!

Announcing General Proximity's $12M, 2x over-subscribed, pre-Series A round. Medicine is in the dark ages. Our understanding of biology far exceeds our ability to perturb it therapeutically.

@ssankar is a legend. Great interview! One key takeaway that's unsaid: A strong mission is a collective superpower. Good luck trying to compete vs Palantir if you don't have equal conviction in your mission.

@ycombinator @lucent_ai @RaeAlisa_ 🔥🔥🔥

Lucent (@lucent_ai) is an AI that automatically watches your session replays and finds bugs and UX issues. They’re already used by top product teams including Reducto, Julius, Finta, and Happenstance. Congrats on the launch, @RaeAlisa_! ycombinator.com/launches/Pfq-l…

As we're squarely in fundraising season, I thought I'd share 10 tips gathered from the 1,000s of seed-stage pitches I've heard in my 15+ years in early-stage investing: 1. Picking at the seed stage is hard. Canva got rejected by 100+ seed VCs. Don’t take it personally. 2. Treat feedback from VCs who pass on your round as noise, not signal. 3. Look for true believers. Don’t spend precious time trying to convince the skeptics. 4. To learn a VC's actual strategic North Star, ask: “What’s the best investment you’ve ever made?” 5. VCs prefer to invest in lines, not dots. They'll note your predictions and your follow-through. 6. The best founders keep making progress while they are fundraising. 7. Great founder-VC relationships are often built upon complementary strengths. 8. How much value investors add before investing is highly correlated to how much they add after. 9. It's all about power laws. IYKYK. Arguably the highest and best use of a VC is sharing first-hand insights gleaned from backing outliers. Remember, VC advice to founders about fundraising tends to make the VC giving the advice look good. Founder advice about fundraising tends to make the founder look good. My 10th and final tip: take it all with a grain of salt.

The average VC might invest in roughly 1 of every 50 companies they meet with. For growth equity firms, that number is closer to 1 in 1,000. The dynamics of VC vs growth equity often bewilder founders. That's not accidental. To understand why, it helps to clarify what these firms actually are. VCs traditionally back early-stage startups with high technical risk. They're unprofitable for years and usually fail, but can create huge returns when successful. Its opposite is private equity, where investors back mature, stable companies for shorter periods and use financial engineering to juice returns. Between them sits growth equity. They look for companies with strong growth and proven unit economics. If the VC targets a 1000x return over 10-15 years and the PE 3x over 3, the growth equity investor might target 10x over 5. Historically, startups needed VCs to survive. In contrast, growth equity investors realized their best investments were in companies that didn't need any money. Thus, unlike the history of sales dynamics in VC, where startups sold to VCs, early growth equity investors such as Summit and TA found sales execution key to generating alpha. These days, the growth equity firm that's most active in reaching out to Horizon's portfolio is Insight. I recently chuckled when I heard one of Insight's founders on a VC podcast boasting that "we've really built one of the great sourcing engines of the world" and that "to get through the gauntlet of Insight's due diligence" is "a pretty big accomplishment." It's true: Insight and their growth equity peers, such as Left Lane and Lead Edge, do far more due diligence than typical VCs. For every company that receives growth equity investment, 100s go through deep due diligence and don't make it "through the gauntlet." But why would so many companies share their information with Insight & their peers when the chances of investment are so slim? They are masters at the art of seducing founders. They hire young, hungry ex-bankers and consultants, train them in outbound sales, and have them reach out to every promising company in a given space. They get smart on the sector and make founders feel like they're being pursued--even when they're not. Now, I want to be clear that I don't think this is unethical. I understand the business model, and when these firms do invest, founders generally like working with them. The problem is the information asymmetry it creates. For founders who don't know they're 1 in 1,000, the experience is a distraction at best and genuinely demoralizing at worst. So, if you're going to talk to these firms—which are so good enough at getting in front of founders that you probably will—go in with the right frame. They're using you to get smart on your space, so use them back. Get their analysis and understand what a sophisticated numbers-driven investor sees in your business. Just don't mistake the meeting (and the likely follow-up texts, etc) for a relationship.

I retook the MCAT in January and scored a 519. Up from 514 in September. As a 46-year-old VC with 2 kids, I think it's a long shot that I'll go to medical school. But part of me is strangely drawn to the idea. For context, I slept through science in high school and didn't take any in college. But I later found myself doing quite a bit of biotech investing. It's an area I knew almost nothing about and, somehow, where I've made some of my most successful investments. Then the pandemic hit, and science was suddenly at the forefront of everything. I'm also getting older, my parents are getting older, and I find this stuff interesting and relevant. That said, I don't know if med school is the answer. I've certainly thought about it. And while my MCAT score wouldn't hold me back anymore, there's a ton of work I'd have to do (volunteering locally and research, primarily) to put together a compelling application. I actually think I would enjoy a lot of that, but it wouldn't be easy to find the time as a full-time VC and a dad. I also have the first-world problem of living on an island, where it's at least a 45-minute drive from everything. I pathologically hate commuting and actively try to minimize unnecessary stress in my life, and the volunteering piece would require 1.5+ hours in a car every time. That sounds absolutely brutal to me. And even if I did all that and got into a great school, I'd have to move my family or live away from them. At 46, I'd be starting at 47 or 48 and finishing in my early 50s. It's also physically demanding, and I have some aches and pains from old sports injuries that might honestly keep me from being on my feet all day. On top of all that, AI advancements might make medicine as we know it obsolete. So while I continue to weigh my options, I ordered basically the entire first-year curriculum for medical school. It's complicated, dense, and a zillion pages. I'm loving it! For now, my plan is to continue to study whenever I have gaps in my day--partly because I haven't yet let go of the idea of being the old dude at med school, partly because I'm genuinely fascinated and want to learn more. But regardless of where I go with this, it's been a remarkably interesting journey.

@nikiscevak Hallelujah!!!

Mimetic desire is a powerful force and with Elon's decision to IPO SpaceX, a wave of venture backed companies will go public over the next five years. SPACs may even return to all time highs (with the dozen big winners from 2021 providing the allure). Savour those capital losses and get ready to deploy

There's a persistent belief in the startup world in the importance of strategy. I think it's because successful companies, in hindsight, always seem to have great strategy. Strategy is legible. Culture, on the other hand, is easy to overlook. The streetlight effect operates in full effect when judging the importance of strategy vs culture. The light shines brightly on strategy, so it must be important! But a startup's culture is very hard to decipher from the outside. Worse, we've all heard poor performers talking up their culture. It can be very hard to separate corporate bs from meaningfully positive culture. But having invested in Canva and Palantir at the earliest stages, I'm convinced strategy is overrated. Culture is the true source of alchemy in startups. A good example of a focus on strategy going sideways is the common obsession with startups building moats. VCs push founders to define their moat, as if there's some clever positioning that locks in a defensible future. But a moat is an output--the natural result of a culture that consistently does the hard things over a long period of time. And with AI shifting competitive landscapes by the second, today's moat can be tomorrow's AI roadkill. When Canva and Palantir were at the earliest stages, they had no moats. What they had were really smart people working really hard, building the beginnings of really good products. That relentless execution was the culture. The moat is just what it looks like from the outside, years later. I don't think it's a coincidence that Canva and Palantir are also two of the best case studies on adapting to AI. Take Palantir, whose share price has appreciated over 15x since the launch of ChatGPT in Nov of 2022. Sure, we can say that Palantir has had a good AI strategy in the past few years. But read the earnings announcement or CEO letter from Q4 of 2022, and you'll see no mention of ChatGPT, LLMs, or anything relating to AI. Few foresaw the importance of ChatGPT, so I would not fault Palantir for that. My point is that what enabled Palantir to adapt to AI so much better than their peers is their culture. It was built around incredible talent density and an intense mission orientation that enabled them to rapidly shift trajectory to leverage AI. Their "correct" AI strategy was very much downstream from their remarkable culture. And while Canva's valuation hasn't skyrocketed like Palantir's, they also leaned into AI much better than their peers and saw revenue accelerate accordingly. Like Palantir, they lacked a perfect crystal ball for strategy but had immense talent and a strong culture that made them amazingly flexible and resilient. Although Canva and Palantir are exceptional companies, they aren't exceptions to the rule. I encourage every founder to follow their lead and worry less about strategy and more about establishing talent-obsessed and customer-obsessed culture. Done well, everything else will follow.

Over the years, I've watched dozens of startups hire the "can't miss" candidate from a direct competitor. It usually fails--often spectacularly. Base rates help explain why. In statistics and medicine, a base rate is simply the proportion of a population that shares a given trait. They're the critical starting point for any meaningful comparison. Imagine we compare the height of a randomly selected pro basketball player and a randomly selected pro soccer player. We'd expect the basketball player to be taller for no other reason than that the populations are different. The same logic applies to industry knowledge. The average candidate from a competitor will know a lot; the average candidate who didn't work at one will only know a little. So when you compare them on industry knowledge, the competitor candidate will win every time--but that tells you almost nothing useful about a candidate's ability to become a star performer. Any decent employee will learn their company's industry. The problem is that this comparison happens implicitly all the time. Industry knowledge is rarely the most important factor in a hire--it's usually considered a "nice to have." But when a candidate has 100x of a "nice to have," it stands out and often outweighs more important factors like attitude, intelligence, and motivation. There's also the selection bias issue. If your competitor is doing well, they'll try to keep their best people--so why is this person job hunting? And if the competitor is struggling, that may not say great things about the quality of their people. Hiring from competitors can work. Sometimes, competitors have star players who want to join your company. But, in my experience, that's quite rare. Founders get in trouble when they are wowed by a candidate's industry knowledge and forget to hone in on the more important signals around ability and motivation. That's why even when a candidate from a competitor knocks your socks off, you should ask yourself: Would they stand out if everyone in the room had the same industry background?

Two things can be true at same time. One, AI was largely a distraction for enterprise software pre ChatGPT. Two, PLTR had an extremely prepared mind to seize the post ChatGPT AI opportunity. I vividly remember Shyam explaining neural nets to me when we were in grad school in 2003. He knew it would lead to world changing tech.

Palantir's AIP was not built over 20 years. it was put together in weeks as a response to ChatGPT after years of internal denial that AI was more than a distraction. These are two former Palantirian FDEs in a slack chat recently responding to my original 10,000 word Palantir analysis.

This sounds more than familiar. $PLTR somehow kept its name out of this story. OpenAI describes Frontier as a “semantic layer for the enterprise”—a unified platform that lets AI agents navigate business software, execute workflows, and make decisions across an organization’s entire technology stack, such CRM systems, HR platforms, and internal ticketing tools. Early enterprise customers include Intuit, State Farm, Thermo Fisher, and Uber. Meanwhile OpenAI says its own “forward deployed engineers” will work alongside the teams from the consultancies in client engagements. Under these new partnerships, which OpenAI has deemed Frontier Alliances, each consulting firm is investing in dedicated practice groups and building teams certified on OpenAI technology. The consultants will help heir clients redesign workflows; integrate AI agents with software tools and systems; help clients with change management; and provide industry-specific expertise OpenAI doesn’t have, Frontier, which OpenAI debuted earlier this month, is a system that allows businesses and organizations to build, deploy, supervise, and govern AI agents .fortune.com/2026/02/23/ope…

Growing up, my dad always said that if you want a serious opinion, you need to know both sides of the argument. Whether we were talking politics, religion, or sports, it didn't matter. It's now one of the most important rules I follow as an investor. If I had an opinion, any opinion, I had to be ready for my dad to tell me I was wrong. He was a high school debate champion--as were his two younger brothers--and was raised in a house that never missed an opportunity for debate. Hearing "you're wrong, here's why..." all the time isn't for everyone. But it didn't bother me. I always knew my dad was smart, wise, and genuinely wanted me to sharpen my thinking. Over the years, I got comfortable admitting it when someone called me out for sloppy thinking. And I built the habit of exploring both sides of just about any issue that mattered to me. I realized that for many questions, the answer wasn't clear--particularly in investing. So I developed the ability to hold a bull case and a bear case at the same time. My debate mindset is working in the background when I'm getting to know a founder. I have no interest in getting into a debate in a first conversation--that would make me a very annoying VC--but I'm always thinking about both sides simultaneously: what's the bull case, what's the bear case? Most importantly, I'm looking for signs of exceptional potential. Is there a bull case with extreme upside? No founder is perfect, and no investment is either. My job is to figure out whether the strengths are strong enough that the weaknesses don't ultimately matter. That said, there's an important distinction worth drawing out. Balancing competing hypotheses is a tool for decision-making, and it should stay there. As a founder or operator or VC, once you make a decision, you have to go all in. You can't sit there questioning yourself forever. The same rigor that should inform a decision can become a liability once the decision is made. The best advice generally reflects this tension: here's a trade-off and there's a trade-off, now pick the path that resonates most. That's what my dad was really teaching me, even if neither of us would have put it that way at the time. I hope it's been a net positive. Looking back, I like to think it has.

@nick_alexander @ouraring 💯

@sandykory 100% agree (we're also @ouraring holders by way of Proxy acquisition, but never claim it as one of our unicorns or top investments). I believe the dishonesty catches up to you over time, you can fool someone in the short term but a reputation is forged over years

The energy it takes to put bullsh*t on the internet is x. The energy to refute it is 10x. That asymmetry is why intellectual dishonesty is so rampant in VC, and why hardly anyone pays a price for it. I've seen this first-hand: I once angel-invested in a startup that was eventually acquired by Oura. A few days after the deal closed, I got an email saying I was an Oura shareholder. A solid exit. Good news. Recently, I saw a fellow early-stage investor from that same deal name "Oura" as one of his top investments. C'mon! We got stock in Oura after it already achieved a multi-billion dollar valuation. It's not like we invested in Oura at the seed stage. It was effectively the Series D. He didn't mention any of that on the website, of course. And look, we all do this to some degree. There's a famous case of a guy in the industry who attended the Harvard extension program and called himself a Harvard graduate. Technically true, but we all know that's ridiculous. The problem is that no one really cares when, in hindsight, it turns out you were totally full of sh*t. And it's not hard to see why. It's a lot easier to hit follow than unfollow, and we get rewarded for engagement bait, whether or not the story holds up. There's just not really a mechanism to hold anyone accountable, and ultimately, the entertainment is winning. All the world's a feed. But that's exactly why the distinction between honesty and intellectual honesty matters. The investor who listed Oura as a top investment wasn't necessarily lying -- he was an Oura shareholder, after all. But intellectual honesty requires consistency in your thinking and a genuine reckoning with what actually happened. Claiming credit for luck isn't lying, but it's also not being intellectually honest. Unfortunately, in an environment where no one goes back to check, that distinction rarely gets made. And our culture makes this worse. You get pilloried if you make a mistake and admit it. The top PR people for any politician say: "Don't admit you made a mistake, go on to the next thing, and flood the zone." Many VCs have taken notice. The path of least resistance is always to claim the wins and let the losses quietly fade. Along the way, intellectual honesty becomes rarer and rarer. If it wasn't already clear, I think I care a lot more about intellectual honesty than most people—or at least more than what our culture rewards. And maybe that's old-fashioned, but I can't imagine doing it any differently.

Secondary liquidity for founders has long carried a stigma in VC, with many investors believing that letting founders take a few chips off the table would kill their hunger. But in my experience, it actually enables the very best to take even bigger swings. The stigma wasn't irrational, as there were cautionary tales. One infamous example was an early social networking startup whose founders took liquidity at the Series A. One of them bought a Ferrari, and the company tanked from there. Causation? Unclear. But that kind of thing gave investors genuine cause for concern, and it shaped how the industry thought about founder secondaries for years. I saw exactly that dynamic play out around 2014, when I was informally advising a startup raising their Series A. The founders wanted to sell around $500K alongside it. One of them had parents who lived overseas and wanted to buy them a new home. Shasta was the VC, and letting the founders take some liquidity was considered a serious concession at the time. I think there's been a real enlightenment here, though. The founders whose companies are doing great at Series B and Series C aren't going to lose their hunger over a few million dollars. If anything, taking a few chips off the table removes the personal financial pressure they've been carrying. It allows them to buy a house and stop worrying about private school for their kids or a home for their parents. Suddenly, there's not really much of a downside anymore. And with that weight off their shoulders, it can be a lot easier to take big bets--which is what us VCs want. You could argue that the very best founders are so obsessed that the liquidity doesn't move things one way or the other. Maybe, but that only furthers the point: the downside risk of giving them liquidity is low, and the upside is real. What the industry treated as a concession for years, I think, is actually a win-win. And it's why supporting the secondary interests of our founders at Horizon is something I'm happy to do.

@fendien 191

Wild stat from StepStone about emerging managers: In 2021/2022 there were 938 managers who raised a Fund I Since then, only 190 subsequent funds have been raised by this group

If you're building a self-serve or PLG product, you need a wow moment: that instant where users think, "Wow, this is f***ing great!" The problem? The bar is rising faster than ever. What felt magical 10 years ago is completely stale today. What's a "wow moment?" It's that initial product experience where value hits immediately and unmistakably. As a product builder, you want minimal friction to get them there—minimum clicks, minimum friction, boom, boom, boom...wow moment. There might even be some dopamine involved. The key is that it happens fast. Think seconds, not minutes. Now, not every company should be optimizing for this. Some products fundamentally require collaboration and thus lack wow moment potential. For Palantir, getting value means bringing together 16 vice presidents from warring factions and getting them all to align and sign up for major changes. Sure, there's a wow moment eventually, but it doesn't come after 2 clicks. Companies with much higher friction and longer sales cycles like this should prioritize different metrics. But if you're building self-serve or PLG, your wow moment should be a central concern. Here's what makes this tricky: what wows is relative, and it's a moving target. The infrastructure you build today to deliver that wow moment becomes a baseline expectation tomorrow. This means you're always building just to stay in place. Add in the fact that the pace is accelerating—what used to take years to become stale now takes months—and the challenge these companies face becomes clear. As you might imagine, delivering that wow moment isn't simple. There are massive amounts of underlying infrastructure you have to build and all sorts of things you have to do behind the scenes. It's like going into a fine dining restaurant and having a wow moment—you're not exposed to the chaos and madness in the kitchen and their whole supply chain. It's simple for the user, complex for you, and the kicker is that the diner's expectations keep rising, which means the kitchen never gets to stop improving. For founders building these products, this approach to user experience should be the priority. The question should be very clear: How do we enable that wow moment as quickly as possible? My take? Reduce friction as much as possible. That's ultimately what this game is about.

@whoisnnamdi Great post! The stickiness of tech products is easy to underestimate.

Open source LLMs are 90% cheaper than closed source models at the same benchmarked intelligence. Yet they capture less than 30% of token share. If price and smarts were all that mattered, this gap should have closed by now. It hasn't.