I’ve been thinking about this manga a lot as I start a new chapter as a founder, about the similarly long journey and struggle I’ve chosen to walk down. Anytime you listen to any of the success stories, there’s a similar throughline: Jensen talks about how he would never have started NVIDIA knowing what he does now; Zuckerberg’s motto is pathei mathos, a Greek phrase that connotes learning through suffering.

So let me tell you, dear internet stranger, a personal story about why I choose the struggle.

My maternal grandfather should not have lived as long as he did.

When he was a boy during WWII, Japanese soldiers swept through his village in one of the countless massacres that marked the occupation. He survived because he was working in the fields, and hid in a ditch when the soldiers were doing their patrol. They were so close he could hear their boots on the soil, but their commander called them back just in the nick of time. He got lucky. Yet, till the day he died, he always referred to the Japanese as 日本鬼子, which roughly translates to Japanese devils. Seventy years later, he still remembers the ditch he hid in.

My dad’s side of the family lost everything during the Cultural Revolution. This is a generational trauma that most Chinese carry but rarely talk about. I don’t think those of us in the West can fully appreciate what it means to lose everything—status, money, freedom, safety—in a matter of months at a society-wide level. My dad’s family was descended from a notable general and diplomat during the Qing dynasty. But since my grandfather was a university president, the entire family was blacklisted and turned upside down. Everything was taken, including the old family estate. My dad grew up with nothing, eating white rice (with some chilis if he was lucky) and sharing one pair of shoes among three siblings.

It is against this backdrop that both my parents made it into Beijing University - the Harvard of China - and then got into Harvard itself for their PhD studies. Studying hard and coming to America with nothing but $100 to go pursue the American dream. A classic Hollywood immigrant story but lived firsthand. So when I think about the personal traumas that I’ve been through - getting bullied physically and emotionally, not having friends, being excluded - it pales in comparison to what my family went through.

There’s this concept in venture capital: bet on founders with a “chip on their shoulder.” Something to prove. I used to think betting on people was more about their intelligence or grit or pain tolerance. But I’ve come to believe that those are necessary but not sufficient. The truly best founders have a why that doesn’t make sense to anyone else.

So here’s mine: I want to get the old family estate back.

It’s objectively ridiculous. The estate is in the outskirts of Changsha. The house is long gone, just a field of weeds on some old ruins. I have no practical use for a plot of land in Hunan province and I haven’t even seen it in person. And yet the idea of reclaiming it sits somewhere deep in my chest, irrational and unmovable. Call it filial piety. Call it symbolic. Call it my own way of righting the wrongs of history.

The founders who build the greatest empires have something like this. Some core motivation that goes beyond money and status, though those are also powerful motivators. If you’re reasonably talented and born into decent circumstances, you can get a good job and live comfortably. Most people do, and that’s totally reasonable and actually very rational. But the ones who choose to struggle, really struggle, they’re usually running toward something only they can see. Or running from something that they can’t forget.

I’m driven by my immigrant story and being the son of a people wronged by history. I’m driven by competition: I’ve been playing games since I was a kid because I love to win. And if I’m honest with myself, some of that competitive fire traces back to being bullied. To being fourteen and weak and forced under the bleachers by people I thought were my friends, and swearing I’d never feel that way again. I’m driven by my many failures: the dream school I didn’t get into, the many jobs and internships that passed me over, the relationships I lost, and the deals and markets I failed to see (more here).

So you might be wondering why I chose to pen this essay right around Thanksgiving. Actually, I’m not thankful in spite of these things. I’m thankful because of them. It’s because we grew up poor that I appreciate the little comforts of life. It’s because I went through so many rejections that I appreciate the monotonous grind it takes to succeed. And it’s because I formed and lost so many relationships that I appreciate the people who have stuck by me throughout all these years.

At the end of Berserk, Guts is still swinging. Still carrying the weight of everything that happened to him. Still moving forward. The manga was never finished—the author passed away in 2021—but maybe that’s fitting. The struggle doesn’t end. You just keep going. So find your ditch, the thing that almost buried you but didn’t. And climb your way out.

Happy belated thanksgiving.

—

If this resonated with you, consider subscribing. I will be writing weekly: sometimes personal essays like this one, sometimes breakdowns on venture, careers, and biohealth. Most of my writing will stay free. If you want to go deeper, premium subscribers can DM me directly, and I’m always happy to chat. Till the next one.

I came here almost four years ago now off of a cold email. I’ve done many different things: learned consumer investing from some of the best, spun out to establish the games fund, and built out a16z speedrun as part of the founding team from the very beginning. I’ve been an investor, salesperson, invite tsar, software PM, program manager, growth hacker, product advisor, and therapist. It’s been the most challenging and fulfilling way to spend my late 20s.

As for what’s next, I’ve been thinking about founding a company for a while and wanted to take the jump. But it wasn’t until recently that I stumbled upon a stellar idea and opportunity, one that fits with my roots and passion for science, consumer, and biotech… where I started my career as a biomedical engineer many years ago. We’ll be in stealth for a while but if you’re building in the bio or consumer health space I’d love to chat.

However, the purpose of this essay is to serve less as a memoir and more a source of knowledge. The goal being to condense four years of learning at a16z and convey some of that hard earned wisdom to you: the reader, whether you’re a founder, executive, student, or part time technophilosopher. Some of these lessons are simple, and may even seem trite, but the simplest of lessons are often the hardest to internalize.

So with that, here are five life lessons I learned from a16z.

Don’t follow your passions

The worst advice I’ve gotten is to “find what you’re passionate about.” First of all… what does that even mean? I love science but hated working in pharma. I love golf but would never become a professional golfer. Passions are helpful guideposts but don’t let them blind you.

What brought me to California and a16z was my passion for gaming, but what I learned here is that it makes it easy to miss the forest for the trees if you’re too passionate. It can lead you astray, blind you to reality. Gaming is a great industry and I thoroughly enjoyed my time there, but I didn’t accept the truth that it’s going through a long bear market because of how capital intensive it is, how oversupplied it is, and how much demand there is for consumer attention. The structure of the market makes it much more challenging (but not impossible) to be great.

The other factor is opportunity cost. For a while, I stubbornly resisted learning about AI because it felt too consensus and too hypey. Better to be contrarian than hop on the bandwagon I thought. I came across the country to work in games after all! But my passion made me emotionally contrarian rather than rationally contrarian. And in fact, it made me late to what I now believe is a generational and fundamental shift in technology which was a critical mistake as a venture capitalist.

What I really needed to figure out is what I’m really good at, where my skills are, and how to apply that to the market. Where does the world need me?

I would contend that if you’re really good at something that gives way more satisfaction than anything else. It is important to have an underlying reason for action of course, which could be passion but could be something else. Status, competition, money, family… whatever it is, having some deep desire and something on the line makes you work harder to win and be maximally competent.

I lost that motivation at one point, the “dawg” in me. That’s because I was passionate about gaming, and that was the whole reason I had moved across the country and left everything I knew in NYC. But my passion wasn’t panning out. So what now? I wasn’t doing this for the money, and I didn’t care about the status. It was a weird feeling to be in a dream job but feel the drive fading.

So I had to find other reasons for action, ones that were durable. A few that resonate with me:

• Helping founders, my people, succeed. Nothing better than seeing the team you work hard with and care about so much do well.

• Competition. I’ve been playing games since I was a kid because I love to win, and that drive is still inherently there to this day. Even if I lose, it’s fun to play against extremely competent and intelligent opponents.

• Solving interesting problems. As a former engineer, I miss challenging intellectual problems that stretch my thinking and allow me to fundamentally understand how things work at a deep level.

• Motivating and guiding others. I write these blogs to help the not-so-strangers on the internet, and feel a lot of joy when people tell me I’ve helped them in some small way.

That’s why I’m going back to biotech to become a founder. Because I have a personal history with diabetes and Alzheimer’s in my family, I’m technical from years in the biomedical field, and if I can get this product to market I’ll help a lot of people in a meaningful way. It’s not necessarily a passion, but a mission I care about and can execute on. Losing my passion made me rethink why and how to become uniquely great at something.

Share

Becoming n of 1: a monopoly of the self

There is a focus at a16z on strengths rather than lack of weakness. This was quite a change from both college and McKinsey where the focus was more on becoming well-rounded individuals, especially at my alma mater Columbia where the Common Core encourages students to become a worldly scholar.

But focusing on strengths has several main advantages:

• If you’re at a job, having things that only you can do makes you irreplaceable and gives you more leverage with your bosses

• More broadly, markets are competitive. Commodity skills will degrade in value. So having uniqueness allows you to command a higher valuation (salary) in market

• If you’re a founder, you have the deck stacked against you. Larger companies have more people, more money, and more distribution. So the only way you win is by leveraging your team’s unique talents to gain an edge.

Think of yourself as a startup in the market. The goal is to become a monopoly. This will enable you to have a moat (job security), pricing power (higher salary), and leverage (move where you want and work how you want).

So then, how to become a self-monopoly? I think the best analogy is actually skill trees from old roleplaying games like World of Warcraft or Runescape. You might have stats like strength, charisma, intelligence, wisdom, or luck. These actually map to modern skills pretty well too: maybe you have high strength so you become an athlete or high charisma so you become a comedian or salesperson. The goal is to become “best in class” and have the skills and stats to become a top level player.

But classes come in different types, some are glass cannons and some are a mix. A glass cannon might look like a wizard or brute who’s very high on intelligence or strength respectively. But a blended role might be an assassin or a paladin who’s pretty high on a few different skills. So for the real world that might look like a software engineering savant (wizard) vs. an excellent CEO (paladin).

For me some of the skills were learned, some were realized, and others were unearthed. I had to learn good PM skills from my colleague Josh Lu and consumer investing / psychology / growth hacks from Andrew Chen. I realized that one of my strengths was forming close emotional bonds and trust quickly with founders, which would help me win deals and help the company grow. And to my surprise, I use a lot of the writing skills that were drilled into me in high school and the technical biomedical understanding I learned in college.

So figure out where you are in the market, and decide how you become irreplaceable. The way to do that is to think for yourself instead of letting others dictate for you.

Independence of choice and independence of thought

Most people believe they think for themselves when they actually rely on the decisions of others. You buy the stock that everyone else is talking about even though you know nothing about the company. You work at a job you don’t really like because it’s high status and a lot of other smart people wanted it. You buy a fancy car even though you don’t like driving.

Decision by mimesis: we want and think based on how other people want and think. I’ve been prey to this myself many times. Examples would be my choice to go to McKinsey after college, and even several deals I’ve done here at a16z. And most of the time, people who give you advice are intelligent, thoughtful, and want the best for you.

But this is a mistake.

I understand now why older people prioritize “critical thinking.” It is very challenging to form opinions that are truly your own, synthesizing what other people say but not listening to them blindly. I found this very challenging in early stage investing coming from an engineering background, because there was often very little data to go off of and you were wrong much more often than you were right.

I needed to learn how to get to conviction quickly, gather what data I could from questions, references, and research, and form my own opinions about both the founder and the business. People think about VC as a contrarian business but it’s actually the opposite.

It’s all group think.

In VC there are structural rewards for mimesis. Early stage investors rely on other later stage investors to mark them up if their company does well. And so you’re always guessing what might be “hot” in the market in a year or two. VCs also often get rewarded for winning hot deals that everyone hears about, sometimes regardless of whether the business itself is outstanding. The incentives are strong. That’s what makes this whole VC game challenging, because market consensus and an oversupply of capital means that prices and risk go up.

Slowly developing the ability to think critically for what I believe, having the conviction to take action, and then being willing to back it up in the face of criticism. That’s what I learned. It’s honestly really fucking hard. When you stick your neck out, when other very smart people tell you no, when even your friends question your judgment. That’s when you know whether you have real conviction.

Self-discipline and self-measurement

Two truths. Very simple but very powerful.

1. Great things are built from small steps

2. What isn’t measured won’t improve

I learned these well from my former colleague Jordan Mazer who I respect immensely. We all know that we should have more discipline: wake up earlier, go to the gym every day, watch less brainrot tiktok. But putting it into practice requires much more mental fortitude. What I observed was that every day, Jordan would be at the office at 6AM with a social media post up by 9AM. He didn’t particularly like social media or talking in public, but he knew it generated results: views that resulted in signups for his job board turned into hires at our startups.. So now he has 100K+ followers and tens of thousands of people on his newsletter.

I believe that how you do anything is how you do everything.

It’s important to practice the discipline everywhere in your life. If you skimp at the gym you’ll probably skimp the last few emails at work too. So whether it’s running every day or writing every week or reading on the go, I learned to build up small habits everywhere that I believe will translate to real results.

And the second piece around measurement is equally important. I needed to keep myself accountable. When no one’s watching and it’s just me, especially now that I’m a startup founder. The company is myself. There’s no manager, no boss, and not even any customers yet. So whether you’re using substack analytics or whoop fitness scores, I find it so crucial to be honest with myself on how I’m doing.

In order to do that, you may need to change your attitude.

War-time vs. peace-time

a16z is a very serious place. First class business in a first class way.

I think this seriousness is best embodied in the war and peace metaphor. It was a very stark transition for me coming from Riot, which was most certainly a peace time company when I was there. There’s a pirate themed cafe. There’s a massive PC bang with N64s for Super Smash and Mario Kart driving machines. An excellent top end cafeteria. Statues of all the champions. A typical day might look like: come in at 9, grab a coffee at Bilgewater, work till 12, break for lunch with friends and colleagues, play a game of League, head back to work till 6:30 when dinner is served, and then play a few more games after dinner.

It was embodied in the culture too at Riot. People take work seriously but there’s less urgency. There was a sense of “bringing your whole self to work” because you’ll game together and get to know each other outside of the work context. You’ll care about each other and get to know each other on a deeper, personal level. It was honestly… very nice.

On the flip side, I learned at a16z what it meant to be in “war-time” mode, which has honestly prepared me to become a good founder. High urgency. High standards. Things must get done quickly and flawlessly. Punishment and reproach for getting things wrong. There’s camaraderie for sure but the team functions more like the military in the communications (efficient), structure (hierarchy), and strategy (play to win).

The benefits of having a war-time attitude are clear especially if you’re playing in a very competitive space. You need to have the entire team playing to their strengths and pulling in the same direction to generate the best results, Steve Jobs style. However, it also means that you won’t be as close to your colleagues by design, and the culture itself will be aggressively meritocratic.

The most important piece of the war-time attitude ties into what I mentioned before: generating results. Real results, not fake shit. No KPIs or random PM metrics, but things that move the business forward. Because in war, intentions do not matter. What matters is who won and who created history.

Onwards and Upwards

I’ve made a lot of mistakes here, but the main thing I took away from my four years at a16z was wisdom. The wisdom to know that I was fearful of making decisions and hard choices. The wisdom to find conviction and choose. The wisdom to understand the game I was playing and the wisdom to leave and pursue the next journey.

If there’s one stat on my life character sheet I prioritize about all else, it’s this. Decisiveness. The ability to make choices.

I hope you make good ones too.

—

Thanks for reading! If you made it this far, there’s a few things you can do to stay in touch. Also a little easter egg for you from the team for SR001 all the way back in 2023.

• Please consider subscribing to support my work! I plan on writing weekly and will have some paid-only posts that are more in depth.

• I’ll be moving to SF very soon so if you hear of good spots / rooms opening up in Marina / Pac Heights / Russian hill do let me know! SF friends looking forward to seeing you more soon

• If you need career advice, fundraising guidance, or feedback on consumer / growth / healthcare / games, areas where I have a lot of expertise, I’ll be doing a few paid consultations (30 min, 60 min) to pay the bills while I build my startup. If I can be helpful to you feel free to book.

But then you grow up. You graduate. You find a real job that pays the bills. You go into the same office every day, eat the same food, drink the same coffee. You get into routine. You work your way up the corporate ladder. You have a mortgage, bills, spouse, and kids to worry about.

Where did the time go?

I remember the exact moment I knew I was slowly dying. Week ten at a pharmaceutical company in central New Jersey. Another Wednesday. I was eating Farmer’s Fridge again, a small Caesar wrap dispensed from a modern millennial vending machine, for the fourth time this week because that was the best option at the cafeteria. My laptop displayed slide 47 of a deck titled “Organizational Synergies Post-Merger.” I was adjusting the formatting on a text box about “leveraging cross-functional capabilities.”

This was it. This was what twenty-four years of preparation had led to. A very prestigious purgatory. I was making $100K a year to slowly transform into furniture.

I originally chose consulting because I thought it was the adventure. Fly around the world. Give advice to some of the top F500 executives. Work with some of the smartest people around.

All true.

But also what was true was endless slides, endless meetings, and endless optimization. I had this nagging feeling I wasn’t… building anything real.

Maybe this was just modern life? And I’d have to get used to it? Maybe the fantasies I had as a kid of galactic exploration, horseback rides through the vast plains, pitched battles with mysterious foes was just that… fantasy.

Call it idealism or call it naiveté but that’s why I moved to California. To chase the modern gold rush. To work in gaming, a literal fantasy. I wanted to answer the call to adventure.

What I wasn’t prepared for was the struggle, the pain that is required for any adventure. I was always prepared to work hard, but I wasn’t quite ready for the market turning for the worse, the uncertainty about what to do next, the lack of meaning, the depression, the feeling of inadequacy, the responsibility of carrying important decisions.

The hero’s journey always has that moment. When Frodo stands at Mount Doom. When Luke faces Vader. When you’re three years into the California dream, unsure about what’s next, wondering if you made a catastrophic mistake leaving everything you knew: friends, family, job, safety, and home.

And then recently one of my founders said to me: “Speedrun feels like summer camp. Three months of crazy adventures and now we’re all going home.”

And what I realized over the last year was that my journey here in California was about helping others realize their journey. To build a structure, program, capital, and community to enable others to make it through.

See, what actually kills adventures isn’t the dark lord. It’s the Tuesday at 2 AM when you’re alone, your runway is shrinking, and nobody understands why you left your tech company to do this crazy idea. It’s the isolation and self doubt.

But when you’re with five other people that are all having the same Tuesday at the same time? That’s different. Someone’s been there. Someone has an idea. Someone reminds you why you started.

The specifics matter less than the shared struggle: whether it’s debating human behavior until 2AM at the Airbnb, or that moment when a founder from Zurich realizes a founder from New York is dealing with the same customer challenges, or when someone gets their first 1M in revenue and the whole cohort celebrates like they just summited Everest.

We built speedrun to be base camp for the greatest adventure of our time: creating something from nothing. Every three months, a hundred people remember that the ten-year-old who thought they could change the world wasn’t wrong, just early.

So to all those still sitting in their corporate companies, adjusting text boxes on slide 47, eating the same chicken Caesar wrap:

The adventure isn’t gone. It’s waiting.

It will cost you everything comfortable.

You will lose sleep, money, and certainty.

You will become someone your current colleagues won’t recognize.

But the alternative is worse. It’s 3 PM on a Tuesday in ten years, and you’re on slide 48, and the wrap tastes exactly the same, and the ten-year-old version of you has stopped trying to get your attention because they know you’re not listening anymore.

My hero’s journey ended where it was supposed to begin: helping others start theirs.

Before we get into the essay, this week we’re hosting Nigel Eccles (founder of Fanduel, valued at $31B) for office hours to talk all things consumer / sports betting / crypto. Link here. Very pertinent to the essay.

Friday the 19th at 1PM PT. If you missed it subscribe to our substacks so you don’t miss the next one!

On to the essay.

Subscribe now

Four years ago I walked away from a guaranteed promotion at McKinsey and a $300K private equity offer to go work in gaming for a third of the salary. I moved across the country from NYC to LA, leaving behind friends and family I'd spent my entire life with. That's called going all in.

I've been fascinated with poker since I was a kid, playing games with my parents at 12, grinding through late night college sessions, studying game theory optimal strategies and probability tables. But it wasn't until I'd been in VC for a while that I fully appreciated what poker really teaches us about life. See, in chess, the "correct" move is correct regardless of outcome. In poker you can play perfectly and still lose, or play like an absolute donkey and stack everyone at the table. That gaming job led to connections that got me into venture capital, which led me to building speedrun. Which by the way, is accepting applications!

For finance jobs like VC, poker is directly applicable. You're literally betting on incomplete information. But the real value comes from applying poker thinking to life itself. Chess assumes a controlled environment: perfect information, one opponent, predictable outcomes. Life is much messier. Life is multiple players with hidden cards, changing rules, and luck that can absolutely demolish skill in any given hand. Once you start seeing life through the poker lens, you realize most people are playing chess in a poker world. No wonder they're frustrated when the "right" moves don't pay off.

So let's talk about some lessons from poker, and I'll assume if you're reading this you understand the basic mechanics of No Limit Texas Hold'em.

Probability and Process

This is the most important lesson from poker: thinking in terms of probabilities rather than outcomes. Many systems are stochastic, not deterministic. You can do everything right and still get crushed by variance.

This is the shit happens law. Even with pocket Aces, you'll lose 15% of the time against random cards. But once I internalized this, I stopped torturing myself over "failed" decisions that were actually correct. This was the biggest mental shift I had to make in "real life" because up until college, you're basically playing chess. There's a fairly well-trodden path laid out where the route to victory is clear. Study hard, get good grades, go to a target school, land the prestigious job. Checkmate. Then what?

It also helps you make peace with the past. You stop playing results-oriented games ("I should have bought Bitcoin in 2013!") and start playing process games ("What's my framework for evaluating asymmetric bets?"). You stop letting bad outcomes invalidate good decisions. Focus on the process by which you made each decision. The outcome is just one sample from a distribution of possibilities.

Table Selection

Table selection might be the most underrated life skill. In poker, you can be the 6th best player in the world, but if you're sitting with the top 5, you're the sucker. There's an old saying: "If you can't spot the fish at the table, you are the fish."

That's why I watched so many quant friends pivot to crypto in 2017. They weren't suddenly smarter or better traders. They just realized that battling retail investors was way easier than competing with quant savants for basis points. Same skills, different table, dramatically better returns.

The hard part about table selection IRL is that choosing easier tables feels like admitting weakness. There's more social status in losing at the hard table than winning at the easy one. How many people stay at McKinsey getting absolutely crushed when they could be the star player at a mid-size company? On one hand, surrounding yourself with the best people makes you better. On the other hand, sometimes you need to be realistic. You’ll have to level up or find a different game.

Bankroll and Bet Sizing

Understand how much you're willing to risk and adjust based on your conviction and the situation. In life, the bankroll isn't just your bank account. It's your time and energy, your reputation and career capital, your relationships and mental health.

For me, I've tended to be too conservative with financial bets (classic loss aversion from a middle-class upbringing) and also too conservative with relationships (worried about getting hurt, so I hedge instead of committing). Some of my friends are the opposite, consistently “madly” in love. Or onto their next hair-brained business idea. We all have those friends. Know your tendencies and how they evolve over time.

By the way, your risk tolerance should scale with your various bankrolls, not stay fixed. The bet sizes you’re making at 25 should not be the same as when you’re 50.

Variance and Time Frame

Here's where life diverges from poker in a beautiful way. In poker, variance is symmetric. You can only win what's in the pot. In life, variance is asymmetric. The right job, right investment, or right relationship can return 1000x. This means you should actually seek more variance in life than in poker, not less. You want to maximize your exposure to positive black swans. But only if you can survive long enough for the odds to play out.

That's why it's important to figure out what you want to do earlier, even if it's just directionally correct. Pivots are totally fine and happen all the time, but each pivot means starting the compounding clock over. Even for myself, I'm still hunting for my real 1000x, trusting in myself and the game to eventually deal me the nuts.

Reads and Ranges

In poker, you're not trying to guess the exact hand your opponent has. You're putting them on a range of possible hands and narrowing it based on their actions. Life works exactly the same way. People aren't single data points; they're probability distributions across multiple dimensions. Your coworker isn't "trustworthy" or "untrustworthy." They're 90% reliable on small things, 60% on confidential information, 30% when they’re under pressure.

Everyone has a range. The mistake I see is binary thinking. Someone disappoints us once, we write them off forever. Someone helps us once, we trust them with everything. That's like putting someone on pocket aces because they raised pre-flop. You need way more data points.

In negotiations, relationships, and hiring, you should constantly update your range estimates (or "update your priors" as they say in Silicon Valley). What hands could they be playing given their actions? That candidate who negotiated super aggressively: confidence or desperation? That investor who's suddenly responsive after ghosting you: FOMO or new information? Read the range, not the exact hand. You'll never have perfect information, but you can narrow the possibilities with every action.

Imperfect Information

Unlike chess where every piece is visible, poker forces you to make decisions with incomplete data. This mirrors basically every important decision you'll ever make. You'll never have perfect information about that job offer, that cross-country move, that key hire.

The key is getting comfortable making high-stakes decisions with 70% of the information rather than waiting for 95%. Because by the time you have 95%, the opportunity is gone. The startup has already scaled, the person has moved on, the market has priced everything in.

Analysis paralysis is just another form of folding every hand. Yes, you never lose when you fold. You also never win. The players who succeed aren't the ones with perfect information. They're the ones who make better decisions with imperfect information.

I've started using this heuristic: once I have enough information to make a decision that's 70% likely to be right, I pull the trigger or fold. Waiting longer is just anxiety management, not risk management.

Bluffing

The threat of the bluff is often more powerful than the bluff itself. If people know you never bluff, they'll run you over. If they know you always bluff, then they'll call you down light. The optimal strategy is to bluff just enough that they have to respect it.

In life, this translates to strategic ambiguity. Sometimes it's better to let people wonder about your capabilities rather than proving them. Sometimes you need to project strength when weak (interviewing while unemployed). Sometimes weakness when strong (negotiating when you have leverage but want a long-term relationship).

The key is being intentional about what you reveal. Your capacity to bluff is an asset. Don't waste it on small pots. Save it for the moments when strategic ambiguity can change the entire trajectory of the game.

Position and Timing

In poker, acting last is a massive advantage. You see everyone else's moves before making yours. Position is so powerful that professionals will play mediocre hands in position and fold premium hands out of position.

Life rewards position too, but people rarely think about it. The second company in a space learns from the pioneer's mistakes. The last person to make an offer in a negotiation knows everyone else's range. Sometimes there's a first-mover advantage, but often the real edge comes from moving last.

But position isn't just about acting last. The nuance is that it's about controlling when you act. Sometimes you want to lead out and define the game. Sometimes you want to check and see what develops. The power is in choosing. Let others reveal information, make mistakes, show their ranges. Then make your move.

Tilt and Emotional Control

Some of your best decisions will come right after your worst beats. Can you maintain process-focused thinking when you're down 50% of your bankroll? When you just got fired? When a relationship implodes? Most people can't. They revenge-trade, rebound-date, panic-accept the next offer that comes along.

Reading Marcus Aurelius helped me develop a framework: when I feel the tilt coming, I stop playing. Not forever, just until I can evaluate decisions based on expected value rather than emotional recovery. The best players know when they're not playing their A-game.

Counterintuitively, your worst beats often create your best opportunities. Everyone else is on tilt too. Markets overcorrect. People overreact. If you can stay rational when others are emotional, you have an enormous edge. The goal is to recognize when emotion is driving your decisions and have the discipline to step away from the table until it isn't.

Stack the Deck

The final lesson breaks poker's cardinal rule: in life, you can stack the deck. Moving to SF for tech or NYC for finance is literally rigging the game. Being in college to find a partner is maximizing your hands dealt. Building in public is marking the cards.

But once you see life as poker, not chess, you realize something. Chess has a correct answer. There's always an objectively best move. Poker doesn't. Poker is about playing your hand optimally given your bankroll, your position, and your read on the table.

This means there's no universal playbook for life. The optimal strategy for you might be terrible for me. And that's liberating. Stop playing chess, searching for the "right" move. Start playing poker. Make the best decision you can with incomplete information, size your bets appropriately, and trust the process over enough hands.

Because in the end, life rewards those who understand the game they're actually playing.

And it's not chess.

If you liked this essay then go read Annie Duke’s canonical work, Thinking in Bets!

Link to signup here, and follow me on X/LinkedIn/Substack for more of these office hours with future luminary founders.

Subscribe now

And with that on to the essay!

—

Most of what you’re learning in college classes right now is useless, and you know it.

That introduction to science class? That sociology paper you need to finish tomorrow? Probably not the best way to spend your time. The truth is, if you want to learn how to build products people want, create content that matters, or identify who to befriend and date, these are all skills you can develop now outside the classroom.

What taught me the lessons I know now was getting rejected from hundreds of jobs, building products that didn’t ship, surrounding myself with the right people, and forcing myself to stay curious. It was all the stuff outside of classes that taught me the things I know today. Given college is kicking off again this semester, I thought it would be a good time to consolidate the advice I’d give to my younger self.

In Part 1 I covered the philosophy of navigating college and mental models to do so: understanding the system around you, that getting lost is a feature not a bug, and that status is a tool not a trophy. So Part 2 will be the tactical playbook I wish someone gave me when I was 20, sitting in Butler Library grinding on p-sets for problems I’d never see again.

Prepare for a fun little rant with a lot of personal anecdotes.

Learn how to build or sell (what to do)

There are two roles that generate the most value: building and selling. This is what I realized working throughout my time in tech and consulting. At McKinsey the core of what we were selling was: brand (get the McKinsey stamp of approval), knowledge (we’ve done this a hundred times), and manpower (we have a smart team that will work 80 hours for you).

If you’re a consultant, lawyer, or banker, you’re selling knowledge and services. If you’re in tech, you’re creating code and selling code. If you’re in the arts, you’re creating and selling emotions, stories, and beauty. With how our society is structured, we often start with building (essays, decks, code, paintings, products) and move into selling as you get more senior (visions, products, teams, projects, yourself).

A lot of what I learned at McKinsey was how to build trust. How do you make a strong first impression, how do you always overdeliver, how do you understand their incentives so that you’re not just helping the company but the individual person, how do you relate to them on a personal level even if their background is very different to yours. That’s really at the core of selling, is having the other person believe in you and the value you bring.

But you can learn a lot of these skills in college. I used to be the social chair for my fraternity, and I remember curating lists, creating CRMs, tracking open rates, understanding referrals, all techniques that I still use now when understanding app metrics. Digital spaces mirror what happens in real IRL spaces. For some clubs, I went door to door to each restaurant to sell them advertising placement. The key insight there was going further downtown, where foot traffic for Columbia students wasn’t as strong and figuring out ways (like referral codes) to properly attribute lead generation.

Or, I remember my senior project was creating an automatic stroke detection device with a wearable photoplethysmogram. The idea was to detect blood pressure changes in stroke victims, which was one of the earliest signs of stroke. It didn’t quite work during the demo day, but the lessons I learned coding the Arduino and writing signal processing algorithms are still relevant today.

Get leverage on your work (how to do it)

I don’t believe in work life balance. If you want to win and be great at something, you have to put in more hours. It just comes down to priorities. But I grew up in New York, and in New York there’s a lot of work that keeps you busy but not productive. Email work. Appendix slide work. Random memo work.

The thing I didn’t appreciate in college was the concept of leverage. How do you create compounding effects? How do you do 10x with what you currently have? There’s a few forms of leverage:

• Human capital - Convince three friends to go work on a side project with you

• Debt capital - Raise some small dollars for a student venture / fund

• Software - Write code that’s infinitely extensible

• Content - Find your niche and scale your content

• AI - Create automations, search through knowledge, and use software as labor

• Knowledge - Understand and specialize in a sector deeply

This is the leverage I understand now and wish I knew then. I could have started writing and publishing earlier. I could have built and shipped more apps. I could have been more risk seeking and gone deeper in on crypto during the 2017 wave. I could have assembled some friends to go on a great startup adventure.

And even though I chose a specific sector to work on and compound my knowledge, I pivoted several times, from biotech to gaming to AI software. I actually wanted to be on the fast track and had a strong thesis for each new direction, but sometimes you don’t have all the data when you make a bet and you change as a person as you get older.

For me, the bet on biotech was because I loved science, but then I realized how slow moving the pharmaceutical industry truly was and the work that was actually involved. There wasn’t really any science in the day to day, unless you were in the wet lab. The bet on gaming was to be closer to a product I loved. And to this day I still love games. But the bet on AI software was because I simply believe this is the most generational change for the next decade. Consensus to say at this point but it takes a while to fully, deeply appreciate it. And I’ve come to understand with the wisdom of years that these market waves do not happen very often, and they happen extremely quickly.

Anyways the best thing you can do now is choose. Choose a direction and go compound. If it doesn’t work, that’s ok, at least you tried and didn’t stay still.

Find your people (who to do it with)

You may have seen that many of the greats met in college. Larry Page and Sergey Brin (Google) met as PhD students at Stanford. Reid Hoffman (LinkedIn) and Peter Thiel (PayPal) were classmates and philosopher buddies. Bill Gates lived down the hall from Steve Ballmer at Harvard. Why does this happen?

The most talented people that you see around you now are also going to be the most talented people in thirty years. The only thing that separates you from them is time and choices. When I was growing up, I thought there was an enormous gap between the people you read about and the people around me. But as I got older, I realized that the smartest folks around me were going to be the greats of the future. We are the new generation.

A lot of the folks around me now that I used to sit in math class with and go drink beers at the local pub are now researchers at OAI, exited founders, quant hedge fund savants, small VC GPs, etc. Those people you look up to are the same as the people around you, they’re just further along.

But there’s one key in finding the best people. You need to find the right environment. Get to the right city, to the right school, to the right company. Keep pushing until you’re not the smartest person in the room. That’s when you’ll know you’re in good company. Even in my life, I didn’t feel fully satisfied at Columbia, at McKinsey, at Riot, until I landed here at a16z and met some truly top 0.001% people at what they do. The reason there are “mafias” (Paypal mafia, Uber mafia, etc.) is because great people gravitate towards one another.

This might be a weird take, but imo it’s also so important to go on some crazy adventures with those people once you find them. As they say, an interesting person is only born of an interesting life. Ray Dalio road tripped through Latin America after college; Steve Jobs had several spiritual retreats to India. Some of my most memorable experiences were a result of meeting great people and making great memories: hiking Mt. Fuji in the pouring rain, powdery ski descents in Colorado, or late night drives through the hills of LA.

Date (also who to spend time with)

Can’t write a college post without talking about dating. After all, college is a bundle of products and services: credentials, academics, career development, community, and dating. Nowhere else will you meet so many likeminded people, for work, friendship, and for romance. The variance of who you meet will go up after college and your free time will go down. And you’ll likely have to spend a lot more time sifting through noisy signals (aka, swiping left a lot on apps).

It’s also a period where everyone is exploring and finding themselves. As important as it is to cultivate your professional life, your personal life and maturing as an individual are equally valuable. You’ve probably seen the graph of who you spend time with after college, so understanding how to pick the right partner is really important. If you apply the secretary problem strategy to dating, then you must get enough data points to form a reasonable baseline. But I would contend understanding yourself is as important as knowing how to pick a good partner.

A few things I learned about myself through my college dating life as an example:

• Trusting my gut: Easy to say, hard to do. Subconsciously I always knew whether a relationship was going to end, whether one had potential, and whether they liked me back or not. But sometimes I just didn’t act on that feeling. I would often rationalize the situation, convincing myself that it’ll never work out even though I liked her, when really I was just afraid of rejection. Or I would be too infatuated with the idea of somebody to recognize the reality that she didn’t like me back. Either way, I think I intrinsically knew what was going on but didn’t listen to my subconscious.

• Fear of rejection: I hate disappointing people, both by being rejected by others and rejecting them myself. It’s meant that I will procrastinate decisions I know I have to make (breakups) and not take risks because I’m afraid they’ll tell me no (lost chances). The principle I apply to myself now is: the more I want to put off a decision, the more important it is to make it happen right now.

• Hyper-rationality: I’m an extremely rational person and approach many things as a problem to be solved. But in relationships, and especially during emotional moments, what the other person wants is to be heard more than anything else. Having empathy for others and listening well - truly listening and not just waiting for your turn to speak - are both important skills that I had to learn to stop my natural tendency to problem solve.

• The right way to communicate: Words carry a lot of power and should be chosen carefully. You might have good intentions but if said in the wrong way or the wrong tone, they will not come across. Even if you say the same thing, different people will react differently to it. Understanding the right thing to say to the right person in the right way is something I’m still trying to master to this day.

I’m sure you’ll get relationships wrong sometimes, and I’m sure it’ll be painful. But dating in college was very helpful for me to understand what kind of partners I work well with, who I am, and how I handle things.

Take the leap

I was a pretty risk averse person growing up. I think it comes with my Chinese culture; my parents didn’t need me to be a doctor or lawyer but my mom works in risk / AI and my dad’s a professor. Pretty stable career professions. We’re a family of engineers, very pragmatic and EV maximizing. I was afraid of losing in golf so I never competed. I was afraid I wouldn’t get into Stanford so I never tried. I was afraid of losing money in college so I only invested in very stable securities. I was afraid of committing to startups so I went to consulting. So what I’m saying is, I’m really the last person to be giving advice about risk taking.

When I talk with my friends - most of whom just turned 30 so we’re having the “omg we’re 30” moment now - the biggest throughline and regret is not giving it a shot. Asking the guy or girl out, giving it your all for a job, stepping off the golden path, moving new cities. When they say time is short, it takes some time to appreciate really how short it is. There’s cognitive time dilation as you get older, since less experiences are new and each year is a smaller proportion of your memories.

If I had stayed in New York, stayed consulting for big pharma companies, I’m sure I’d still be doing well. But I’d be asking myself what if, what if I moved to California, what if I took the jump to chase after something new. I’ve resolved to never ask myself what if, and just make it happen with conviction. Not all of it has shaped up how I’d like it - it was very lonely in Los Angeles, the gaming market took a turn for the worse, I miss my family - but I have no regrets.

The mentality I’d like to leave you with is, we grow up with many choices made for us: where we live, where we go to school, what we study. And as you get older, the most important skill you can develop is to make your own choices with conviction. No one else is looking out for you. So you better damn well figure it out yourself. The irony is that by the time you have the wisdom to know what you should have done in college, it's too late to do it. That’s why I tried to leave some here. The only question is: will you do it?

New York is a city defined by energy. The briskness of footsteps down the street, the sound of money flowing through digital terminals by the billions, the late-night walks along Riverside Park and the strange, liminal adventures at two in the morning.

San Francisco is a city defined by outliers. The great software empires built from humble beginnings - proof that outsiders can rewrite the rules of the world - the once bullied nerds finding a city that celebrates their own kind (myself included), the liberality of culture and expression that allows you to seek whoever you’re brave enough to become.

New York’s culture lies in human creation. The Broadway shows that push the boundaries for performance art, where performers embody someone new every night. A symphony of music from the refined elegance of the Philharmonic to the raw authenticity of Brooklyn house. Paintings that span the eons as humanity explored itself through small changes in color and brush.

San Francisco’s culture is found in nature’s creation. The soft lapping of the nearby ocean shores in Big Sur, bringing calm to the most turbulent of minds. The peaks of Yosemite that pierce the sky in all its barren glory and remind you that anything worth having must be earned. The great blue expanse of Tahoe that summons the weekend rush for those seeking a sunny refuge or winter escapades.

New York gathers many different souls in one place. Step into any subway car and you’ll encounter a hundred different stories from across the world. Bike across the neighborhoods and you’ll see culture shift around you: new languages, different dreams, and the varying colors of restaurant awnings and smells of the food within.

San Francisco gathers the same souls in one place. Those who believe the world is malleable, highly rational about their journey, and relentlessly ambitious about realizing it. Who refuse to accept that things have to be the way they've always been, including themselves. Never have I met so many individuals from across the world united by the tribe of technology.

New York lives in the present, with fortunes trading in microseconds and ephemeral gossip about weekend plans and the dating market.

San Francisco lives in the future, thinking deeply about what new innovations might disrupt the way we live today.

NY builds up, erecting monuments of steel and glass.

SF spreads out, populating its valley.

NY works hard and parties harder.

SF works hard and works harder.

NY runs on finance and leverage.

SF runs on software and creation.

“Technical” in NY means you understand numbers.

“Technical” in SF means you understand words with structure.

And yet while many will argue which city better than the other; the two are more similar than they are different.

Both are monotheistic to their industries and the high priests that lead them.

Both create those serendipitous 2AM conversations that reshape how you see the world.

Both are relatively meritocratic, caring more about what you deliver.

Both have status games that exhaust and exhilarate in equal measure.

Both make you feel like you're at the center of something that matters.

And both taught me that home isn't just where you're from; it's where you become who you're meant to be.

On to the essay.

Every turning point in my life traces back to a single, irreversible decision. I became a consultant at McKinsey straight out of college because I hadn't yet figured out what I wanted. It seemed like the smart thing to do—prestigious firm, good exit options, interesting work, smart people. But really, it was a non-choice. An unconscious admission that I didn't know my own values yet beyond continuing on a "good path." My values were actually other people’s values, a mirror of what other people also wanted.

That non-choice actually ended up changing my life. I was randomly assigned to a consumer retail project even as a healthcare / biotech guy, and my manager on that project Lena was close friends with another manager, Sylvia. Sylvia happened to have interned at Riot Games, and left soon after to join the company full time. I’ve grown up playing games my entire life (top 200 in PUBG lfg!) and so when working in big pharma was not fulfilling my inner calling there was one choice on my mind.

I got the job at Riot, gave notice, moved to LA, and went all in on my passion for gaming. And LA is where I met my girlfriend of four years. None of which I could have predicted as a young scrappy 22 year old right out of college. But here's what I've learned since: while my non-choice to work at McKinsey paid off, I got lucky. I happened to have set myself into the right place with a lot of smart people and increased my serendipity, but it was the conviction in leaving, in moving, and in taking a somewhat unconventional path that led me to where I am today.

The most important skill I've developed over the years is making actual choices with direction, vector choices—quickly and with conviction. So let’s talk about why that matters.

Author’s Note: This is a personal essay and any views expressed here are my own and not reflective of a16z or representative of the firm.

Everything is Compound Interest

Compound interest isn’t a finance concept; it’s a life mechanic. Every choice you make either compounds or it doesn't. Every day you spend in the wrong market, building the wrong skills, or with the wrong people is a day you're not compounding in the right direction.

Take my move into gaming. Now looking back, I joined during the COVID bull market when gaming was the hottest thing on the block. The Quest 2 was making record sales, investment into games was flourishing, and consumer spend in games writ large was at an all time high. But it wasn’t just hype; there were a lot of smart, talented people creating generational products like Elden Ring and Baldur’s Gate 3. And I spent three years going all-in on that bet. Built deep relationships, learned the industry inside-out, landed multi-million dollar deals, and led substantial projects for the company.

But markets matter more than effort. No matter how hard you work, if the market isn't there, the compound interest doesn't always accrue. Gaming has been flat for the last few years after covid (you can check out some of the data here), PC and console in particular. There’s been mass layoffs across the industry. For those that stay in the gaming industry, jobs are less stable, career progression is not guaranteed, and projects that devs have poured many years into might get cancelled at no fault of their own. That’s not to say I don’t believe in the future of games; I still play a lot of games in my free time, cherish the friends and mentors I’ve met, and deeply appreciate what I’ve learned. The games market will recover. But I learned a painful lesson: choosing what to compound matters as much as the compounding itself.

This principle applies everywhere:

• Relationships compound: Some of the people I met back in college are now running successful startups or have become important executives at notable companies. And more importantly, your relationship with your family and significant grows and strengthens over time too.

• Skills compound: Ten thousand hours matter only in markets still growing. The world’s best flash dev in 2025 earns nostalgia, not leverage.

• Geography compounds: Every year in a central hub (NYC, SF, etc.) adds to your network in ways that are hard to replicate elsewhere. The people you meet, the relationships you build and the roots you grow slowly alter your life in remarkable ways.

Vector Choices

I've come to realize there are two types of choices that matter:

1. Vector choices: Pick the hill

2. Commitment choices: Climb it hard

Most people, including younger me, obsess over commitment choices while avoiding vector choices. We'll optimize our morning routine while staying in the wrong job. We'll work 80-hour weeks climbing a hill that we may not want to summit.

At McKinsey, I was great at the commitment choice. Top percentile of my class, working late nights, taking on the hardest cases. But I was climbing the wrong hill, at least for me personally. Consulting teaches you to analyze other people's businesses, rather than build your own. And the leverage is wrong: you trade time for money, and your best work helps someone else capture the value. Better to have real equity and ownership over the things you build.

The vector choice is hard because it requires admitting you might be wrong, or that your values have changed over time. I chose to study biomedical engineering and work in biotech because of my love for science and understanding how the world works at a fundamental level. But I realized that the day-to-day of working in a wet-lab or a big pharma co was not appealing to me. I chose to not study computer science because I wanted to work on “real-life” engineering (mechanical, electrical, biomedical, etc.). But now working in technology it’s definitely a crutch to not be as technical as I would like. Hindsight is tuition.

The Cost of Hedging

My biggest weakness has always been hedging. Keeping options open feels safe, but it's actually the riskiest strategy. Consulting is the epitome of hedging; you can try new industries and verticals for quite a while until you make a real choice about where you want to dedicate time. And while you're hedging, someone else is committing fully and capturing all the compound interest.

I see this constantly in founders now. The best ones aren't hedging. They're not keeping their day job while exploring their startup. They're not flippantly jumping from idea to idea when it doesn’t work. They pick a big market, make a bet, and commit fully. If they're wrong, they're wrong fast and can make another bet. If they're right, they capture all the upside. The opportunity cost of not finding the right hill to climb compounds silently.

The Paradox of Choice

Here's what's counterintuitive: making more decisions makes future choices easier. Every time you make a real choice—leaving a job, moving cities, starting something new—you build the mental model. You learn to distinguish between real risk and perceived risk. You learn that the cost of a wrong choice is usually lower than the cost of no choice.

When I was 24, leaving McKinsey felt impossibly risky. Now, having made that jump and many others, I realize the real risk was staying. I don’t regret moving to LA or working in gaming or now working in VC. Sure not all my bets panned out (see discussion about the gaming market above), but that’s life. And it’s taught me a lot about how to make better bets and better choices in the future. You only gain wisdom through making decisions, and more wisdom through wrong decisions.

Making Better Choices

So how do you make better choices? A few principles I've learned:

On vector choices:

• Look for fundamental shifts, not incremental improvements. The internet, mobile, and AI were all vector changes that created entirely new opportunities.

• Avoid consensus hills. If everyone agrees it's the right hill to climb, the opportunity is probably already captured or there’s not that much alpha.

• Talk to people one step ahead, not ten steps ahead. They remember the actual climbing and are still in the trenches.

• Rank your own values—industry, lifestyle, culture, location, impact—then test new commitments against that list.

On commitment choices:

• Small daily decisions compound more than big yearly ones. Writing every day beats planning to write a book.

• Automate habits: decide once, then let defaults carry the weight.

• Conviction follows action. Prototype, ship, adjust—thinking alone never closes the loop.

• Work hard once the hill is chosen. There’s no substitute.

Where I Am Now

I've made a choice to go all in on tech, and I’m committed to learning, writing, and building my way to success. I was a student of science and now I’m a student of software. The nice thing is, there are some transferable skills I can tap into. I have a background in Biomedical Engineering and healthcare, which proves useful for understanding the healthcare GTM and business models. Gaming has a lot of crossover lessons around growth, monetization, and design for any consumer facing applications I look at. After all, both games and apps are really tapping into core human behaviors and motivations and incentives. I choose to go in on these areas not because it's safe—I’m actually at more of a disadvantage being non-technical and not from the Bay Area—but because I see fundamental shifts happening. AI is creating new surface areas for consumers. The compound interest potential is massive.

But more importantly, I'm committed. No hedging. No keeping one foot in "just in case." Every day I'm not building in the right market is a day of compound interest lost. A wrong decision made with conviction is better than indecision and staying put. See you in SF.

So what does it actually take to stand out? After evaluating thousands of pitches, I’ve distilled my evaluation framework into a three-legged stool. Note these are my personal views from my experience, and not indicative of a16z as a firm or the broader ecosystem. Every investor out there will have different tastes, theses, and tactics, but this should be helpful as a general framework.

And if you’re a pre-seed / seed stage company, consider applying to a16z speedrun! Applications close this Sunday May 11th; if you’re reading this now you can list me as a referral (Robin Guo, Substack).

Republished with edits

Here are the three pillars:

1. Team

2. Metrics

3. Market (and product)

Ideally, you excel in all three: a killer team with strong founder-market fit pursuing a massive opportunity, with early validation that your product resonates. Short of that trifecta, you should be remarkably spiky in at least one or two dimensions. Underlying these three principles is something equally crucial: storytelling ability. At its core, fundraising means convincing investors to believe in you, your vision, and your creation. The more compelling your narrative, the smoother your fundraise will go.

Let’s dive in.

Team

Everyone says team is important, but what does that actually mean? Investors are hunting for statistical outliers—three or four sigma humans with the clear potential to create outlier outcomes. Your job is to demonstrate that you're one of these rare, spiky people. This can manifest in many forms, but here are some signals that consistently stand out:

1. You build constantly, even without permission. Your GitHub is littered with side projects. Your weekends disappear into hackathons. You've repeatedly gone from 0:1 on projects, not because someone assigned them, but because you couldn't help yourself. Getting from 0:1 is fundamentally different from scaling 1:100, and I need to see you can do the former before worrying about the latter.

2. You possess a specific, unfair advantage. You've deliberately focused your skill points in areas that give you a unique edge. Maybe it's technical knowledge others don't have, customer insights from deep industry experience, or access to a network nobody else can tap. Maybe it’s IOI or IMO medals, or groundbreaking research papers. The question that should have a compelling answer: Why and how will you win when others start competing against you?

3. You have obsessive depth in a weird niche. Maybe you built a TikTok following explaining AI, went all-in on an obscure crypto protocol in 2014, reached grandmaster in competitive Starcraft, or spend weekends building custom drones. This matters for two reasons: First, you've demonstrated the ability to go unreasonably deep, essentially getting a "PhD" in something you care about. Second, these seemingly random niches often become tomorrow's innovation frontier. For instance, GPUs were "just for gamers" until they powered the entire AI revolution.

4. You're a talent magnet. Great people instinctively want to work with you. Can you convince exceptional people to join your mission? Can you recruit folks with more experience or credentials than you have? If you naturally attract top talent, you'll likely attract capital too, these abilities are strongly correlated and at the basis of both is convincing other humans to believe in you and your vision.

   Thanks for reading The Bird's Nest! Subscribe for free to receive new posts and support my work.

5. You have skin in the game. The best founders have made genuine sacrifices to build their company, quitting cushy jobs, moving cities, or surviving on fumes when the going gets tough. VCs notice when you've got nothing to fall back on because it signals true conviction. Those who've hedged their bets by keeping side gigs or consultant roles often lack the necessary focus and desperation to drive breakthroughs.

6. You’re fast. A team that ships weekly will outperform a team with perfect credentials who ships quarterly. When evaluating founding teams, I'm constantly looking for evidence of extraordinary speed: how quickly you built your MVP, how rapidly you iterate, and how decisively you make decisions.

7. You’re intellectually honest. Ideally, I’d like to see that you can publicly admit when you're wrong, acknowledge uncertainty, and update your thinking with new evidence. You should be confident of course, but also incorporate new data from the world to make better decisions.

Boiling it down, this is really about talent + idea + agency. You've got a unique skillset, a killer idea, and the drive to execute relentlessly. The traits I've outlined above are indicators of these three fundamentals. But let's get tactical about what your founding team should actually look like:

• Solo founders need to be full-stack builders. If you're going it alone, you can't afford to be blocked on anything. You need to build, sell, recruit, and fundraise, all simultaneously. The main advantage, however, is that you've got more equity to dangle in front of key early hires who are missionary about your goals.

• Co-founding teams with complementary skills. Look at your collective capabilities like a puzzle. Are there any glaringly obvious pieces missing? Whether it's technical chops, design sensibility, distribution know-how, or hiring skills, make sure your core team covers as many of the critical bases as possible.

• Proven working relationships matter more than you think. Co-founder breakups kill more startups than anything else. When things get tough, and they will, you need absolute, unshakeable trust in your partners. Previous battles fought together are the best predictor of surviving future ones.

• Your references need to be locked and loaded. I'll spend hours getting to know you, but the best folks to chat with are those who’ve worked with you in the trenches. Good VCs will do both front-door and back-channel reference checks. Prep your references thoroughly and make sure they understand what VCs are evaluating and can speak to your specific strengths.

Final thought on team: the pitches that always get me excited are the ones where my information gain per minute is extremely high. Convey what unique insight you bring that explains why you – specifically you – will win in the market you’re going after.

Metrics

I’m not going to rehash all benchmarks and goals for B2C and B2B companies since my colleagues have written about it extensively (B2B, social, marketplaces, growth). Instead, let's focus on getting early signal before you have booked revenue or a large cohort of DAUs. And by the way, the best thing you can do if VCs say no is to keep building and improve the business. Marc’s article two decades ago rings true today too.

So pre-revenue what do you look for:

• User engagement. Are your family and friend testers actually using this thing? Is your Discord with a hundred users constantly buzzing with product feedback? Even without traditional retention metrics, you need to get a read on how users react to your product. Look at concrete indicators: are they sharing it unprompted? How emotional do they get when discussing features? How many messages are they sending about it?

  • For B2B products, this typically means having a few design partners who genuinely love what you're building and are ready to convert to paying customers once it's production-ready. User love matters regardless of whether you're selling to enterprises or consumers.

  • If you're building with LLMs, run sentiment analysis on your Discord, Reddit threads, or sales calls. I personally dig through review pages and early community forums to see the raw, unfiltered reactions.

  • In particular, when users encounter bugs or downtime, do they get frustrated or just shrug? The best products create emotional dependence. If users are genuinely upset when features break or respond with "when will this be fixed?!" that’s gold.

• Marketing/waitlist metrics. Maybe your product isn't fully baked, but you can make the idea catch fire. This alone can be valuable since it demonstrates two things: potential market interest and your ability to cut through the noise with distribution.

Until you have hard metrics, you're sifting through fuzzy signals trying to find truth. Most people default to being nice, so users won't tell you how they actually feel until you detect it in their actions or unconscious behavior.

Market (and Product)

I've watched countless startups pivot multiple times over the course of their journey, and smart founders usually have good instincts about which markets can support big businesses and rotate the business toward that opportunity. While having UX mocks or an early demo makes things tangible, what matters more is your depth of thought about what you're building and why.

Some market considerations that should be in your pitch:

• Market definition and scope. Frame it at the largest defensible size. Uber wasn't just black cars - it was revolutionizing transportation globally. Airbnb wasn’t just blowup mattresses in a room - it was revolutionizing the travel industry.

• The "Why Now?" question. This breaks into two scenarios:

  • Some ideas are obviously good but fiercely competitive. Look at the mobile/web vibecoding space: Lovable, Bolt, Replit, and dozens more are all fighting for position. What's your specific edge?

  • Or if you're tackling an older market, what's changed? What technological unlock or timing advantage lets you succeed where others failed? If you're building another marketplace for trading cards or pet services, you need a compelling reason why this time is different.

• Path to monopoly. How will you dominate long-term? What moats are you building that keep competitors at bay? Network effects? Workflow lock-ins? Data advantages? Something needs to create compounding value that makes your product more compelling as you scale.

• Category creation vs. disruption. Are you creating a new category or disrupting an existing one? Each approach has different requirements and challenges. If you're creating a category, focus on the "from → to" transition you enable. If disrupting, show why your 10x better approach matters.

• Initial go-to-market. Who are your day one customers and what's your core ICP? This needs to be razor-sharp, with deep understanding of their behaviors, pain points, and willingness to pay.

• Map of the decision-maker, buyer, and user ecosystem. Especially for B2B, understand who makes decisions, who controls budget, and who actually uses your product. Markets where these align (same person) have faster adoption cycles than those with complex procurement.

For the product itself, I’m really evaluating it as a reflection of your team. How quickly did you build it? (evidence of velocity) Even if it's rough around the edges, does it show taste? Does the product embody your unique market insight (evidence of thought)? Later on, product will matter more but at the earlier stages it’s more of a reflection of the team.

Conclusion

At a16z speedrun, we back founders at the earliest stages who embody many of these principles: who are spiky, fast-moving, and taking a swing at big ambitious ideas. The application deadline is approaching this Sunday, but whether you apply with us or pitch elsewhere, I hope you found this essay helpful. Application here! Feel free to add me as a referral (Robin Guo, substack).

And if you need more help, I’ve written other essays such as 5 tips for fundraising in a bear market. Hope this was useful!

From the outside my path might look quite straightforward: Columbia → McKinsey → Riot → a16z. But underneath the linear trajectory are a lot of zig zags and false starts: getting rejected hundreds of times, pivoting from bio to consulting to games to tech, moving cross-country to LA with no friends, and dealing with some serious contractions in both the games and crypto markets. My thread here chronicles some of this journey:

The things I cared most about in college were career, dating, and friends, and I doubt that’s changed over the last decade. So while I’ll focus mostly on the career piece in this essay, I’ll interlace some frameworks on the latter two that are arguably even more important long term.

College students, this is for you. Parents, this is for you. Everyone else, feel free to swipe left on this essay, I won’t be offended.

1. Status is a tool, not a trophy

College is full of social status games. What fraternity / sorority you belong to, what party you’re going to, what company you’re interning at, and what grade you got. It can be easy to get caught up in these games, and they can be very useful at giving you external motivations for achievement. However, you must remember that direction matters more than speed.

You can run really quickly at a goal that is a mimetic desire, socially constructed. Think about it. Do you really want to be a banker/consultant because you deeply enjoy the job, or because it pays well, is prestigious, and there are a lot of other smart people you respect at those networking events? Do you want to join the social group because you inherently click with the people, or because that social club/group is the top business/social/whatever club?

This isn’t to say that grades / jobs / social groups are not important, they most certainly are. But I wish I had spent more time asking myself why I wanted what I wanted, and recognizing that getting a prestigious internship is not an end in and of itself. I pivoted several times already, from finance to research to consulting to games and now to tech. It’s ok to pivot, and it’s ok to go in the wrong direction for a while. But remember, if you’re too caught up in the race itself, you won’t re-evaluate whether you’re in the right race altogether.

So, Robin, how the hell do I pick a race? Well, first you need to run into the woods, embrace exploration over choosing a pre-defined path too quickly.

2. Lost is a feature, not a bug

To understand yourself, you must fail and get lost. Up until now, getting into a good college has probably been the one career goal that’s been assigned to you. You probably haven’t failed that much, and you probably haven’t been very lost at all. Now the paths get broader and you will have to make tangible decisions that affect your future self with limited data on whether you’re optimizing correctly.

That’s why you need to get lost. People will sell you optionality because it’s a very attractive alternative and certainty compared to not knowing what you want to do. Just take this next step, and postpone the decision. Well, it turns out that at some point you need to decide what you want your life to look like, and you might as well figure this out earlier rather than later.

My process has always been as follows: get lost, try things, develop a hypothesis, go deeper, and re-evaluate whether I need to get lost again or double down. This applies not just to jobs, but also to friends and to significant others too. I completely changed friend groups after freshman year. I went through multiple breakups in college too. I got rejected from hundreds of jobs and internships. It is inevitable that you’ll feel lost, but it’s just part of the process. College is a unique time when you have the safety net to try and to fail. A bad grade now won’t matter as much as later on if you made a strategic mistake for a company that lost them millions of dollars, you might get fired, but you have a family, kids, and mortgage to support.

Think about it this way, the true expectations and responsibilities on your shoulders right now are relatively low. You’re not (yet) responsible for hundreds of people, budgets in the hundreds of millions, or landmark shows/games/scripts/books/etc.

The only thing that will be hurt by failure is your ego.

If you make an app and no one uses it, so what? You’ve learned that product-market fit is hard. If you start a TikTok / YouTube and no one views it, that also means no one has seen you fail, only yourself. You need to remember that everyone starts as a novice, and really internalize that it’s only by trying will you get better and succeed. When you were a kid, did you worry about being bad at basketball / soccer / dance / swimming? No, you just did it. So especially now, you shouldn’t mind being bad at writing/coding/selling/building/drawing when you start. Develop shamelessness. Everyone has imposter syndrome.

Once you’ve found a thing you like, I generally find a stack ranking exercise useful to figure out whether to double down. Here are the parameters that matter to me when evaluating a job choice:

• Industry: Do I like the overall mission of the company? Do I believe that this market will do well?

• Function: Do I like my day to day role and activities?

• Compensation: Am I compensated fairly for my work? Does my pay scale well into the future?

• Hours: How much do I care about work-life balance?

• Location: Does my proximity to friends / family matter? How do I like the overall city structure and people?

• Risk: How much inherent risk am I taking with this job? What’s the worst downside scenario?

• Culture: How do I like the people I work with? Do I learn a lot from them and do they inspire me to be better?

You should figure out what matters the most to you. For me, I care a lot about the people I work with and the content of what I do, and less about the pay / hours / location. But for some people that’s really important! Once you have your stack rank, you should be able to find ikigai—your vocation, the thing only you’re uniquely suited to pursue. For instance, gaming and investing works for me well because I grew up in and around finance, and I’ve put tens of thousands of hours into games cumulatively. I have an inherent edge on anyone else who hasn’t put in the hours and isn’t as interested in the subject matter.

3. Know your game

As you explore the college sandbox, you need to figure out what the rules of the game are, and what strategies you can employ to win (if you want to win, which I hope you do).

For instance, if you’re an aspiring doctor, the game is grades. How do you min / max your amount of knowledge and studying to pass with flying colors. For a PhD, the game is research. Who are the best professors doing interesting research in the field you like, and how can you work for them and build a relationship with them. For investing, the game is preparation and practice. What theses on the market do you have, how have you generated alpha, and have you mastered all the basics of finance and business.

Some games are way more multi-dimensional. Dating is one of them. Startups are another. For dating of course looks and status play a role, but a lot of it comes down to timing, luck, and vibes too. There’s no cheat code for dating—just trial, error, and serendipity. Same with startups. You know you’ll need an excellent team, the ability to code a great product, and an interesting idea, but after that the environment is too multi-faceted and non-deterministic for you to have any one dominant strategy. You could be making an AI app that OpenAI then comes out with a feature for, or have a co-founder dispute on vision, or run up against the age old question of retention.

Either way, you should seek to understand the system you’re playing in, and what is under your control and what is not. Very practically, that means reading everything there is on the internet about your industry. Leverage the power of AI search and reasoning tools to your advantage. Cold email (or, even better, warm email!) as many industry professionals that you respect as you can, and then email more. Be prepared for those conversations to make them high throughput. Do the job, before you even have the job. Build, trade, code, create.

Every game has a different win condition. I used the McKinsey brand to cold email all the games investors in the industry. And I sent them startups I found interesting to provide value. I worked at Riot to build my network and credibility, and broke into investing when the timing was right. Now, my game is finding the most talented founders out there. That’s why I write, scrape Twitter & LinkedIn, host events, and grow my network with high quality people. Once you know yourself, you can know the game and how to win.

4. The summit worth climbing has no trail

After inevitably getting yourself lost, I encourage you to choose the riskier path. My one regret with college is not being more risk-on: with investments, relationships, and careers. You probably got here because you were risk averse. You got good grades, studied well, showed leadership, and did good extra curricular activities. Now it’s time to embrace risk.

I talked about how college is a safety net, but it’s also one great learning experience. And you learn faster by taking risks and making decisions. From the small minute things like whether to take a hard class or join a particular club, to the larger ones like what to invest in and what career to pick, I’d encourage you to choose the harder path each time. It is often not as treacherous as it seems and will be much more rewarding.

“Ships are safe in harbor, but that’s not what ships are for.”

Think about it this way. Dropout startup founders are, at the beginning, fundamentally unqualified to lead a large company. They haven’t risen through the ranks and learned fundamental management skills. However, because they were the ones that created something from nothing, because they are the owners of their capital, they are able to gain a position that will enable much faster learning than any other regular job out of college.

Or you can think about it purely financially too. This is probably one of the lowest points in terms of your overall net worth, and so you will overvalue a hundred dollars now versus later. Remember that your income and net worth will grow exponentially in the coming ten years if you do well.

Say your net worth is $100 today. Maybe you take $20 to invest in the stock market. But the net present value of your net worth is actually $1000, based on future earnings in the near term. By that logic you should actually be investing way more into the market. Take more risk, find the asymmetric bet of the next generation, and go all in. Do not lose all your chips, but seek the areas where you have asymmetric upside. Of course, this only applies if you are not financially burdened by debt or family responsibilities, in which case those will take priority.

5. Your time is fucking valuable

A very common college mindset is to undervalue your time. You’ll go to an event you don’t care about for boba, scroll TikTok mindlessly, or just procrastinate by doing a whole lot of nothing. Don’t think this way. A good principle from Naval Ravikant (founder of Angel List) is to put a dollar value on your time (like $200) that seems way too high, and then double it. Now you have a litmus test for how to make the most of your time.

College students are becoming content creators, doing fundamental research, creating startups, building open-source projects, and making hundreds of thousands of dollars from investing. You should build, you should network, you should learn. Make the most of your time and value it appropriately. Success may look easy from the outside, but the messy middle is a lot of repetitive tasks and discipline and failure.

If I were to think about how I spent my time in college, I would prioritize fitness, friends, projects, work and relationships, the classic things. Cut the extrinsic stuff that gives you low value: going out too often, clubs you don’t care about, classes that aren’t important, manual labor that trades your time for a low wage. Make a list of things you want to get done each week, each day, and get them done. Keep yourself accountable with a task list. It’s important to build internal accountability now; you need to discipline yourself to have intrinsic motivation.

By the way, this might seem at odds with the second principle around getting lost. It isn’t. Getting lost means that you were playing the game and trying and it didn’t work out. You got lost because you were at least venturing somewhere. But you should never stay still.

Conclusion

College is your sandbox—experiment boldly, fail fast, and remember: the path to meaning is rarely a straight line. In Part 2, I’ll focus more on the personal side of college, the importance of hard work, cultivating real relationships, and finding the tribe you belong to.

Thanks as always for reading :)

You already know this. So let’s cut to the chase and talk about a few ideas for raising money in a down market. This essay is for any startup that’s raising in a presently contrarian market. And if you want more game studio specific advice, you can read my earlier essay here.

Why take my word for it? We’ve advised hundreds of founders through SPEEDRUN, and have a unique vantage point on the market having seen so many deals go down at the seed, Series A, and Series B stage. What I’m writing about are some of the principles that we teach during the program, so we practice what we preach. We also can take lessons from adjacent markets in tech and crypto, whose cycles tend to move faster than in gaming.

1. Price the Market

Work with your current investors and founder friends to understand where the market is at currently. Another good way to get data is pull deals from Pitchbook in the last year in your sector, or ask investors when the last deal they did was. That’ll give you a sense of money velocity in the system. And remember that when deals are announced, these are lagging indicators. Don’t point to a deal that happened 1.5 years ago, announced five months ago, and say “Oh well I’m better than X so I should be able to raise Y too.” The most common mistake I see is if founders overvalue their startup and team and go out to raise too much. Because private markets are illiquid, valuations can be based off a few bidders and there are no clear metrics early on to benchmark revenue / EBITDA / FCF to. As such, investors are looking for signal in lieu of financials: team, story, opportunity, product, user metrics, etc. You need to know where your value is in market, and price at or below that point. Why?

It’s way easier to raise the valuation / round size once you have some term sheets and actual heat on the round. That gives you leverage. And leverage allows you to command the actual terms that you want. If you go out for too high of a valuation, you could turn off VCs that may have otherwise been interested because it’s just out of their range. For instance, a $12M round would be a Series A that only multi-stage funds can afford, but a $7M round could be considered a larger seed round that some early stage funds could stretch for. A bigger pool of investors increases the odds that you can get that first TS. I like to say there’s two games in fundraising: the game to get the first TS (sourcing and closing) and the game after the first TS (negotiating). Make sure you optimize your sourcing and closing game as much as possible.

Don’t raise based off ego, raise based off reality.

2. Evergreen Fundraising

In bear markets, investors will be more discerning about teams and traction, and you’ll likely have longer lead times for the round unless you have some crazy traction numbers (e.g., your D30 retention from early tests is close to 25% for a mobile game). That means you’ll have to spend more time doing relationship building, and giving the investor more conviction in your founding team and understand the business better.

There’s a few tactics in order to do this well. The first is investor updates. A good investor update will be structured and succinct. Investors get a lot of inbound, and so you want every update to highlight tangible changes in the company, such as an important hire, a new enterprise contract, a new build milestone, or an inflection in usage / engagement. Do not write overly long or overly frequent investor updates or they’ll just skim and tl;dr it. You can use open rate as a rough proxy for how effective your investor updates are. 60% should be a good benchmark to shoot for.

The second is very basic. Meet in person whenever you can. It’s a great litmus test on both sides. First you can see whether the investor is serious about your startup, since IRL meetings tend to take longer and are higher friction compared to zoom. And second you can get a read in person, and notice all the subtle cues that might be lost digitally. After all, it’s a two way street. If you take the investor’s money you’ll be working with years or even decades so you better get along.

The third is to test the waters early but not often. One tactic you can use is to schedule quarterly catch-ups with investors you respect and have a good relationship with, ostensibly for advice on company building, product strategy, hiring, etc. This will allow you to build a more durable relationship with the investor, and give you a soft check on how excited they are about you. If they schedule more meetings, if the meeting goes long, if they really dig into your metrics, these are all good signs that it could be time to raise. You can also ask these trusted investors specific metrics / milestones they’re looking for at different stages. Good investors will give you a direct answer: “I’m looking for X retention metrics or Y ROAS curves or Z demonstrated ability to ship from the team.”

However, do not formally raise too frequently. You can only come to the well so many times before investors start having doubts about the company health and downstream financing opportunities. What they’ll be thinking is, “If everyone else passed what did they see that I’m not?” I’ve seen some startups that are always raising for an entire year, and that can paradoxically make the raise harder. Make the hard decisions, cut burn, get back to shipping, and come to the market again when you can show some tangible progress. Marc wrote about this situation - “When VCs say no” - almost two decades ago but the advice still rings very true today.

Finally, you may want to go to a broader set of investors. In bull markets you have the luxury of choosing who you’d like to work with, but in bear markets you need to treat capital as more of a commodity. Tap the long tail of investors: smaller VCs, family offices, high net worth angels, VCs from other geographies. You’ll need to expand the radius, often to 60 or even 100 investors, to increase the chances you get a bite. The key is to get enough data on the long tail of investors that you can improve your signal to noise ratio and figure out who is most likely to be interested in what you’re building. You can do this by looking through Pitchbook or Crunchbase for similar companies and analyze their cap tables or look at what the investor track record was, or backchannel by asking founder friends for intel. The broader set of investors will allow you to more effectively leverage the next tactic I’ll discuss below.

3. Rolling SAFE Strategy

One strategy you can use if the leads aren’t biting is what we call the “Rolling SAFE strategy.” Essentially you turn a round into smaller chunks at progressively higher valuations that creates more urgency and allows smaller investors to effectively “lead” a small tranche.

Say you want to raise a $5M seed at $25M post-money valuation (equivalent to 20% dilution not accounting for the option pool). Turn that into a $1M on $15M tranche (6.6%), a $2M on $20M tranche (10%), and a $2M on $30M tranche (6.6%). These are dummy numbers, but you get the idea. You should break the round into achievable goals (the first two tranches) and stretch goals (the last tranche) by identifying what is the minimum amount of capital you need to get back to building and achieve your next milestone.

There are several tradeoffs for this strategy:

• Pro: You allow smaller investors to lead a tranche. You may be in the situation right now where a lot of follow-on investors say “I’m in for $XM, but only if you find a lead.” Very frustrating. With this method, you want to go to your biggest follow-on investors that have the most conviction in you and say “Hey, would you be interested in leading a capped SAFE at a lower valuation so that we can get back to building?” Your give is the lower valuation, and your get is that you start closing investors out, right now. Get two $500K follow-on investors to agree and your first tranche is closed.

• Pro: You create momentum. This is the most important part of the fundraise. Now you can go to other investors and say you’ve closed your first tranche and have $1M committed and wiring, and ask whether they would be interested in the next tranche. Investors, even at the early stage, can be momentum investors so you want to create a feeling of urgency and time pressure to get them to make a decision.

• Pro: You can more effectively price discriminate. Different investors may have different ownership and valuation requirements based off the size of their fund. The old adage is true: your fund size is your strategy. Rolling SAFEs allow you to group investors into different buckets so that you can try to get the ones you want in each tranche, and ideally assemble the team that you like.

• Pro: You can turn the round into a priced round if a lead materializes. One important nuance to this strategy is that lead investors will often have ownership requirements that are larger than your tranches. No problem, if they get really interested then you can also turn this into a back-to-back priced round, where you raised your first tranche as a SAFE and then the next one as a priced round shortly after. However, you should be careful about staging conversations, so that the lead investor does not come in too late when too much of the round is already committed and you don’t actually have space for them anymore.

• Con: You’ll take more dilution. You must create a give for your early believers who are willing to roll the first SAFE, and that often will be a lower valuation than you’d ideally like. You will likely take more dilution in this scenario net-net compared to a proper priced round. The situation you must avoid at all costs, is that stacking SAFEs will create a very punitive situation if the SAFEs do not realize at their cap. Often the SAFE cap is interpreted as the price of the round by future investors, but not always. If the market shifts or the company does not make sufficient progress, you don’t want to be left with a lot of SAFEs converting at a lower valuation and diluting you too heavily. In that scenario you will likely have to do some cap table management with your existing investors to figure out what’s fair.

• Con: You’ll have more constituents on your cap table. Because you won’t have one lead investor taking up most of the round, you’ll have to manage a group of investors who could be different from traditional VCs. In order to effectively raise in a down market, you likely have to go very wide, to family offices, to high net worth individuals, to different geographies, and these investors could have very different expectations for how you run the company. Be thoughtful about setting expectations early so you don’t have time-wasting disputes down the line.

4. Bridge Rounds

I won’t sugarcoat it; bridge rounds are generally a tough conversation with your current investors about the state of the business and what the future looks like. Usually you’ll have the insiders lead and have some external parties match in order to inject more capital into the company. The best advice is preemptive, you should already have a strong relationship with your current investors and they should be brought along for the ride with frequent updates. That way you increase the odds that they consider you one of the best teams in their portfolio, even if you haven’t quite yet nailed PMF yet for one reason or another.

Generally a bridge round will be flat or down-round in order to create an incentive for investors, and you’ll need to paint a story about where the bridge goes to. This $XM bridge will give us 12 months of runway in order to ship product Y and get metrics ABC. Or we have early signals of PMF through our retention / engagement / playtests / etc. and we need the extra capital in order to scale. The bridge should always bridge towards a tangible outcome, and you may have to stomach more dilution in order to get the capital you need. Ideally you have some interested new external parties that are interested but need a lead investor, and one of your insiders will step up to price that bridge round.

5. Opportunity and Runway Planning

Being too early to a market is the same thing as being wrong. Be realistic about the opportunity you’re going after, and whether you have to refactor the company and pivot to go after a new opportunity. On one hand startups are contrarian by design and you need strong conviction in your idea. On the other hand you don’t want to be blindly stubborn to the realities of the market. You should decide the balance. For what it’s worth, many great companies such as Discord or Slack famously came out of pivots where they realized the current direction wasn’t working, and went into cockroach mode in order to build the new product that they believed in.

The other piece of advice is to be more conservative when you do get a fresh injection of new capital since you may need to wait out the bear market. The typical advice of a VC is to plan for 18-24 months of runway, but they’re not always clear about how or when. You may want to consider planning for 36 months instead of 24 months. There are other options to extend runway through outsourcing, contracting, or hiring internationally. I’ve had some startups have success with looking for 10x engineer college grads based off their OSS projects or Github commits. Get creative and be thoughtful about your capital. Of course, if you see PMF you should double down on that opportunity, and in general your burn should match the stage/traction of the company. Stay lean while you’re exploring, and expand aggressively when you’ve struck gold.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z. A list of investments made by funds managed by Andreessen Horowitz is available at https://a16z.com/investment-list/.

During the Christmas holidays, people usually do a year-in-review or Spotify wrapped. But my journey with a16z is less year-by-year but instead two arcs: pre-SPEEDRUN and post-SPEEDRUN. So I thought this would be a good time to tell the story of how it all came about and the ups and downs in the journey so far: We’ve gotten more applications every class, with thousands of applications and a <1% acceptance rate. We’ve gotten talented founders who have created and scaled notable companies. And we’ve gone from a primarily gaming focused institution to one that supports many types of founders: from AI agents to gamified consumer apps to defense simulations to 3D modeling software to crypto Telegram games to sports betting.

This is the story of how one little idea spawned into a real institution with over a hundred companies and a hundred million dollars deployed in just a year and a half. In this essay I’ll cover the origin story, lessons and learnings along the way, and most importantly for aspiring founders, how we pick our class. And if you’re building a new pre-seed/seed startup please don’t hesitate to reach out.

So how’s it all start? Let’s rewind to the spring of 2023.

Small team, big dreams

It all began with a simple idea: there are many accelerators out there but none for the gaming industry. YC is the gold standard for accelerators, but it’s mostly focused on enterprise SaaS: in the last S24 class, only 16 are consumer of 254 companies. Gaming in particular is also a highly fragmented industry: infra vs. studios, LA vs. SF, mobile vs. PC, casual vs. hardcore, etc. There’s no program that unites all these disparate communities, and games are often related more to Hollywood than to Silicon Valley.

We wanted to change that. We wanted to unite the tribes, and we wanted to prove that the gaming industry reaches much further than just video games alone. We wanted to show that much of the cutting edge of tech - from graphics cards to AI agents/NPCs to social networks - comes from the gaming industry. We knew that just doing a few deals a year wasn’t going to make enough of a dent to prove our thesis, and so SPEEDRUN was the one way to do this at scale and build a real network effect of founders that would become close friends, help each other through their own journeys, give feedback on each other’s products, be the first customers, refer new founders, and pay it back when they’re successful through advice and angel investing.

We also knew that we were one of the few teams that could pull something like this off. This doesn’t come from a place of arrogance; we knew that there were a few core things when building an accelerator that played to our strengths: a brand that people desire, a scaled marketing operation to drive top of funnel, a strong network to tap referrals, and a broader team to handle the sheer amount of work that comes from planning several “weddings” (demo days) a year and partnering with so many companies in detail.

We broke it down into a few core tenants that I believe still hold true today:

1. Attract the best founders. If there was only one principle, it would be this. The quality of the founders in the program dictates the quality of the deals and the quality of the community. Everything flows from here. If you get the top 0.1% of people in the program, it will become desirable for others of high caliber to attend. If you attract B or C-tier folks, then you create adverse selection and you’ll never be able to get the S and A-tier folks to come.

   1. In particular, SPEEDRUN has been useful for breaking into net new networks as we’re able to take more risk outside of our usual mandate: Stanford/MIT/Berkeley/etc. undergrads, Turkish mobile studios, East / Southeast Asia as a few examples. We’re always looking for new talent dense networks to tap!

2. Help those founders win. The more wins that come out of the program, the more it helps reinforce tenant #1. Founders will speak highly of our team and the program, and their wins will become aspirational for founders that are earlier on in their journey. Of course, creating those wins is easier said than done, which is why we have a kickass investing and operating team (25+ strong just on games, and many more in broader a16z) to go tackle these challenges. We’re happy to report that our NPS is always 80+ and often closer to 90. As a comparison point, top brands like Coke and Mercedes Benz often poll in the 60s to 70s.

3. Maximize the Demo Day / fundraising experience. Ultimately, SPEEDRUN will culminate in a Demo Day and we want to create the best environment for our founders and the Demo Day attendees. This is often a very field-oriented, hand-to-hand, account management style motion where we try to talk to some of the top attendees to understand their ideas, concerns, and approach. The most important thing is finding them a few companies that really pique their interest – oftentimes this is a company that’s games-adjacent! - and match them with those founders.

4. Have flawless operations. SPEEDRUN is hosting a wedding, every week, for 12 weeks straight, with one massive culmination at demo day at the end. We would not have been able to achieve this without Samira, Marigold, Karishma, and the rest of the marketing + operations team spearheading this entire effort. It truly takes a village, and the attention to detail is critically important (how large the demo stations are, what the lighting looks like, sound for videos, speaker timing / coordination, website flow, the list goes on and on).

So with those four principles as our guiding light, let me tell you how we got to where we are.

The Beginning

We started with a prototype in SF and LA. Andrew Lee was brought on originally to set up this crazy idea, and I joined him soon after as we realized this was a mountain of work that needed more bodies. We worked together to put forward the first six week program, and it was definitely… a startup. We ran the first application system out of Typeform + Airtable, we had over a thousand applications, put together a quick branding for the class and program, drafted the week-by-week schedule / content, and brought together a very solid first class. At that point, we were still just trying to find the best founders from the games industry, and the content was more tailored towards starting games-centric startups (studios, infra, mobile apps, etc.).

A few things became immediately apparent to me as we were putting this together. First was simply how operationally intense something like this is. There’s a thousand little decisions that most people take for granted that must be made by someone, and are often taken for granted. For instance, someone has to think about the venue, AV equipment, timing/logistics, run of show, food and beverage… even how many chairs are in the room, how the chairs are arranged, how many breaks we have, the list goes on and on. But it’s important to get all of these right since the culmination of these decisions lead to a great experience and getting any one thing wrong can cause problems: too many breaks and you lose people mid-way, AV issues and you’ll waste time and look unprofessional.

The second is how powerful the community and network effects are. Founding a company is often a lonely journey, and you’re so heads down in things that it’s easy to just focus on your immediate next milestone or your competitors or the customer. Sometimes you need some friends and companions along the way, and those can often be hard to come by as the signal to noise ratio can be high at events, and it’s often difficult to set aside the time to create real trust and intimacy amongst founders.

The sheer act of us selecting and pulling a bunch of great humans in one room that are working together towards shared but different goals is very powerful. A lot of the startup journey is trying to get the data and intuition to make sense of a very non-deterministic world, and unlike corporate jobs or school it’s not laid in front of you in a structured fashion. Bringing people from different domains together is the best way to learn, especially if they have a high trust, high performance environment. There’s a lot of domain-specific knowledge that can only be gained through conversation. That’s where the alpha is.

As we iterated on the formula for SPEEDRUN we made a few important strategy changes to bring it to where it is today. Here are a few of the most important changes and learnings we had from SR001:

• We expanded the team. This project was way too large for just two people to run it, and we needed specialists in each field. We still structure it such that all the investing partners run their own cohort of founders that they’re directly responsible for. But for day-to-day operations, Josh came in to GM the program, having led large teams at both Meta and Blizzard before. Samira was hired to run marketing/ops, having an extensive career at 100T, Riot, and Zedd (yes that Zedd). Andrew focuses on the investor / demo day experience, and I focus on the investing side of the equation (applications, strategy, picking frameworks, etc.).

  • We’ve also expanded the broader games team as a result to handle all the new portfolio companies we support too. Remember that SPEEDRUN is literally 10x the number of deals we normally do a year, and so each vertical scaled up their support, with the games team now numbering 28 strong across talent, marketing, creators, content, people, and more.

  • But yes the core SPEEDRUN team is still run by… six people. Crazy huh?

• We focused on pre-seed / seed companies.

  • We realized that doing a program for companies at all different stages didn’t make sense. The help that a Series A/B company needs vs. a pre-seed company is very different: the former is concerned with scaling up product-market fit and getting to unit profitability, while the latter is still deep in the idea maze and hiring their first engineers.

  • Given the amount of effort we were putting in the program, it made sense for us to have a standardized deal that allowed us to buy meaningful ownership in each of the startups that we support, that way we’re incentive aligned to make sure they succeed.

  • We also wanted to separate our SPEEDRUN motion from our normal venture Series A motion, where we’re betting post-PMF on companies that we believe can become category leaders.

• We broadened our mandate.

  • We started this as a way to connect the game industry and teach a somewhat arcane art of making, evaluating, producing, and distributing games. That has become so much more as we now invest in the broader creative and entertainment industry. For instance, some of the companies I work with encompass everything from crypto games (Primodium, Astronomica, Matchday) to AI creative apps (Hedra, Manifest, Sekai) to AI agents (Co-human, Altera) to gamified consumer platforms (SweatPals, Favorited).

  • We even have companies that work in defense (wargaming, drone simulations), in sports betting and sweepstakes, in dating / sharing, in animation / vfx / creative tools, and more.

  • The reason for this is we believe that games are the forefront of innovation for the future, and encompass all things digitally interactive. Metaverse is a loaded term, but we already live digital-physical lives that involve interacting regularly through digital media like Zoom or Hinge or Discord. Twitter you can see as the global public forum, or you can see it as one massive PvP MMO where you do actions to gain status and recognition.

• We refined our processes.

  • There’s a million small decisions that add up to the overall experience from start to finish. Here’s just a few of the things that we had to think through and refine over each iteration:

    • Pre-sourcing: How do we build the SPEEDRUN brand? What does it stand for? What channels do we market on and how do we want to express ourselves? What does our off-cycle marketing beats look like? Are there particular interest areas / markets we want to go source more startups? What about talent networks?

    • Application: What fields do we need to cover? What do we want the completion rate to look like? How do we maximize signal and minimize drop out rates? What platform should we run the application on? What’s the turnaround time and how quickly can we get back to people? What proportion of startups do we want to interview and what are the most important signals we want to look for?

    • Program: How much of a time commitment do we want the program to be? How do we want to introduce founders to each other and give some structure to the relationships? What speakers do we want to bring in? What content do we want to cover? How do we want to lead the discussion for each section?

    • Demo day: How many attendees do we want? What kind of attendees do we want? How do we get the decision makers in the room? What’s the best venue to accomplish our goal? How long is each presentation? Do we have a break? What’s the food / beverage set up look like? Do we have a demo room / demo stations? How long do we want to run the social program for? How do we increase collisions and info gain?

    • Post-program: What does support look like after the program? How do we manage hundreds of companies while having a limited team? How do we manage the alumni program? How do we continue to engage the network and introduce new companies into it? What are the scalable ways that we can continue to help our companies?

• We built the infrastructure.

  • There’s quite a bit of digital and physical infrastructure we needed to build to handle the volume and scale of SPEEDRUN: a website, an application portal, a solid and repeatable AV team, a brand, a process, etc. These are the things that scale over time but are super manual to set up in the beginning

What we look for in SPEEDRUN companies

Many founders who are reading this post might be wondering how we select SPEEDRUN founders and how to better position themselves for success. I wanted to spend a bit of time going through our picking criteria because we also want every applicant to be as successful as they can and put their best foot forward.

Here are the things we’ll be evaluating, in rough order of importance:

1. Traction: As they say, traction is king. Early signs of product market fit are the best thing you can do to improve your startup’s odds of being accepted into SPEEDRUN. In particular, we care a lot about growth & retention. If you can focus on two things, it would be growing your user base (for apps) or your revenue (for tools) and improving your retention. Great traction will trump all other things, and it’s ok if the signs are super early (few hundred/thousand users for an app or design partners for a tool).

2. Team: Are you truly the best team to tackle this opportunity? And it’s better if you’re honest with yourself about where you’re at, both strengths and weaknesses. We look for several important traits and self-awareness is one of them. Others include ambition, leadership, insight, hustle, conviction, and intelligence. We’ll ask you a lot of different types of questions to get at your core motivations, why you believe something that many others don’t, why you’ll win, and how you’ll lead your team and build a company and a cult. As founders, you must shape the world into what you believe in. If I could boil it down to one formula it’s: Agency = Will x Capability.

3. Thesis: Following the last point, what is your unique view of the world that you alone hold? The best pitches I hear are the ones that really tech me something, and we’re able to have an intellectual debate about the market, product, and subject at hand. As VCs, we’re going to be three or four levels deep in each market we cover, but as the founder, you need to have the PhD level expertise in this area. After all, you’re spending 24/7 studying this particular market.

There are other variables that we consider such as market and product of course, but since a lot of these things are so early we trust that things will improve and change in the product. What we’re really looking for is proof points on whether we believe that this is on its way to becoming a generational company.

Subscribe now

And don’t forget! We get things wrong since there’s so much uncertainty at the earliest stages of the company building journey. We’ve taken several companies that have applied before and made more progress that gave us more conviction in the potential. Two points make a trend, and we want to be betting on high second derivative founders.

The future of SPEEDRUN

So you might be wondering at this point, where do we take it from here? What I love about our team is that we’re always pushing the boundaries, always innovating. I can’t share all of it, but what I can say is that we’re looking to scale, and become the institution that helps a16z support founders across their entire life cycle: from idea to product market fit to unit profitability to IPO and beyond. We want to become one of the best ways to work with a16z early, and hopefully the partnership continues through every phase too. There’s a few tactical things we’re implementing such as a visiting partner program for experienced entrepreneurs to pass on their lessons, a consolidated website that handles both application and investor flows, a revamped social/video strategy, and processes to continue to scale our help to hundreds and eventually thousands of companies.

And remember, if you’re a founder interested in SPEEDRUN, our doors are always open. And if you’re not a current founder but interested in speaking, dropping by, or working with us, feel free to reach out too!

I’ll paint a picture for what success looks like to me. If we can become one of the default go-to programs for any founder to work with a16z, if we can become the ubiquitous aspiration for every founder (“why not try speedrun?”), if we can build one of the best networks in games x technology, if we can provide real value back to the community, and if we can help build generational companies, that will be work I look fondly upon and feel deeply grateful for.

Onwards and upwards.

The views expressed here are those of the individual AH Capital Management, L.L.C. (“a16z”) personnel quoted and are not the views of a16z or its affiliates. This content is provided for informational purposes only, and should not be relied upon as legal, business, investment, or tax advice. You should consult your own advisers as to those matters. References to any securities or digital assets are for illustrative purposes only, and do not constitute an investment recommendation or offer to provide investment advisory services. Furthermore, this content is not directed at nor intended for use by any investors or prospective investors, and may not under any circumstances be relied upon when making a decision to invest in any fund managed by a16z. (An offering to invest in an a16z fund will be made only by the private placement memorandum, subscription agreement, and other relevant documentation of any such fund and should be read in their entirety.) Any investments or portfolio companies mentioned, referred to, or described are not representative of all investments in vehicles managed by a16z. A list of investments made by funds managed by Andreessen Horowitz is available at https://a16z.com/investment-list/.

Anime – what started as a niche genre of manga and animation has become a multi-billion dollar industry in its own right, with chart-topping games like Pokémon Go and Genshin Impact grossing billions and movies like Demon Slayer: Kimetsu no Yaiba becoming one of the highest grossing films of the year. Anime is driving pop culture today; just look at the fashion collaborations like Loewe x Studio Ghibli or famous lyrics from artists like Chance the Rapper and The Weeknd. We estimate the total market size of anime to be north of $50B, once accounting for anime, comics, and games (data.ai, Precedence Research). 

Yet, we believe that anime is just getting started – today, new technologies and business models are accelerating a once niche industry to even greater heights. Anime is where the future of creation and play is happening, whether with innovative games on our phones, user-created fanfictions and games, roleplaying as the characters themselves, or interactions with AI-versions of these fictional beings. We believe that anime will become one of the dominant formats through which we consume and engage with content. 

The usual lifecycle of an anime fan looks something like this: they are introduced to the format with great IP – the Attack on Titan anime or the One Piece live action show or one of the miHoYo games. They then deepen their fandom through new forms of social and creative engagement (UGC, VTubing, community formation, cosplay, etc.). What we’re most excited about is how anime engagement has evolved from passive to interactive over the last few years. This has happened along a few different axes: 

• AI companions, an evolution of classic visual novels, are the most popular for anime characters and IP. They are quickly becoming one of the dominant forms of roleplay. 

• UGC platforms (user-generated content) such as Roblox allow anime fans to create their own renditions of popular anime games much more easily than ever before. Fans can express their creativity with play that they define themselves. 

• Anime studios are adopting new AI technologies to create content faster and more cost effectively, but they are also iterating on new core loops with AI-native character interactions. 

• VTubing has transformed the way millions of anime fans interact with their favorite characters in new social and parasocial relationships by allowing any fan to roleplay as the characters themselves. 

Ready to dive in? Let’s jam.

A Quick History of Anime

Anime is a genre of animation that originated from Japan and is known for its distinctive art style: bright colorful graphics, strong outlines, large, exaggerated eyes and features, and brightly colored hair. It originated in the early 1900s with visual strips, but really started becoming more popular in the 60s and 70s with shows like Astro Boy, Lupin III, Akira, Mobile Suit Gundam, and of course, Hayao Miyazaki’s works at Studio Ghibli. Anime entered the mainstream in the 2000s with popular shounen anime like Naruto, One Piece, and now Attack on Titan. You might see football players referencing the Attack on Titan salute as a celebration, or pop culture memes all around the internet.

Nowadays anime is more popular than ever: Crunchyroll estimated that viewership doubled between 2020-2022 alone to over 300M fans worldwide. Netflix reported in 2022 that over half of its subscribers worldwide watched some anime content on its platform. The top 4 Discord servers worldwide are all anime-related. Anime is also a medium that caters to many different audience types, from mature seinen anime like Monster to classic action shounen like Demon Slayer or Naruto, from female-oriented shoujo like Ouran High School Club to contemplative slice of life stories like Mushi-shi. Because of this broad appeal, there’s a huge community of diverse fans that have engaged with the medium over many years and developed strong affinity toward IP and characters. This affinity has led to one of the most popular use cases of AI recently –  AI waifus and husbandos. 

The Rise of AI Waifus / Husbandos

First, to explain popular slang – what is a waifu or husbando? Waifus / husbandos are fictionalized girls and guys who viewers feel some level of emotional or romantic attachment to. This isn’t as rare of an occurrence as one might think; Polygon reported that 44% of surveyed anime viewers reported developing a crush on an anime character. “The other 56% are lying,” the report jokes. After all, anime fans tend to express their fandom deeply through merch, fanart, conventions, cosplay, and other means (66% make some form of fan art according to fandom). So it’s not surprising that they would also develop a deep attachment to the characters. Anime Expo in LA for instance, routinely attracts hundreds of thousands of fans each year, the a16z team included. 

Waifu / husbando culture derives from visual novels, and AI companions are the logical extension of these animated storybook games. Originally, visual novels were serialized books with anime-styled pictures in between. But now the term colloquially refers to the digital version in which the player has agency, with branching storylines where you roleplay as the main character, advancing through the plot and choosing to romance other male/female characters. This is a relatively niche market (~$150M according to some sources), but one that has an outsized impact on the industry. For example, Fate/Grand Order, one of the most successful gacha games with $5B+ cumulative revenue, is descended from a visual novel called Fate/Stay Night. 

The structure of the visual novel with branching dialogue trees is a precursor to modern AI companions today. While there are many practical use cases for AI-simulated human interactions – AI as therapist, as teacher, as assistant, etc. – one of the most popular realizations are as friends and romantic companions (more on the broader companion space from my colleagues here). For instance, character.ai’s top characters are all from Genshin Impact; Raiden, Yae Miko, and Hu Tao take some of the top spots at 390M, 202M, and 113M messages respectively as of the time of this blog, compared to Elon Musk at a mere 40M messages. Other top characters are Gojo Satoru (638M) from popular shounen Jujutsu Kaisen and Levi Ackerman (250M) from Attack on Titan.

Many of the top companion sites like janitorai, chub.ai, or spicychat.ai also feature anime companions – ranging from roleplay with school bullies to overprotective sisters to… furries? – as some of their top personalities. There are also vertically specific companion apps for anime: crushon.ai is an 18+ example, talkieai is relatively safe-for-work, while Hyper includes VTuber-esque live facial elements. There’s moemate which allows you to bring the companions cross-browser and play games like chess and checkers, and kajiwoto that allows you to more deeply customize your prompts and datasets and share them on server. 

AI is deepening the previously parasocial relationships we had with our favorite anime characters from passive linear media, into powerful new, interactive relationships. The reason why anime is so resonant as a genre for AI companions is because character affinity / romance is already built into many of the best anime stories. The majority of top anime games and visual novels are role playing games that feature a romance mechanic, and so it’s natural for fans to want to deepen their connection to their favorite IP and characters through active interactions. 

As AI advances, the interactions will become non-scripted and more multi-modal: we’ll interact with our companions through voice and text, their animations and reactions will be generated in real time, we’ll be able to play games with them, and they’ll have longer context windows to remember our shared history together. We’re already seeing startups pushing the boundaries for how we interact with content. Sekai is using AI to create short, bite-sized videos of you and anime characters in fanfiction-esque scenarios. Meanwhile, there are several stealth studios that are building otome games in the vein of Love and Deep Space ($200m+ annualized revenue, SensorTower), but where the side quests, dates, and interaction texts in the game are AI-generated. 

UGC Democratizes Creation for Anime Fans

Anime is the new playground for content creation. Fans often engage with anime IP by creating their own versions of art, novels, and games, and innovation is happening across the stack. There’s Nijijourney for art, NovelAI for fanfiction, pixivFANBOX for creator support, and UGC-platforms for games. For instance, several of the top Roblox games are based on IP like One Piece or One Punch Man, and for a platform with over 60M DAU and 300M MAU, this is a very broad player base of anime fans. In particular, Roblox is a good proxy for what’s trending with the younger generation given its demographic skews towards kids and teens. Similar to how AI companions give fans a new way to interact with IP, UGC allows them to reimagine and remix their favorite IP in wholly new contexts. 

Blox Fruits, the second most popular Roblox game by active users with 40B+ visits, is based off of One Piece, where players can choose to roleplay as a Marine or Pirate, fight enemies, improve their characters, and obtain special abilities through devil/blox fruits and Haki. The Strongest Battlegrounds, the fourth most popular Roblox game with 7B+ visits, is based on One Punch Man where you can assume the role of different heroes that have special abilities and combos to fight other players. And in Minecraft, many of the most popular servers that players frequent are based on Pokemon or Naruto. 

There are a few notable reasons for the popularity of these games. The first is that there’s clear player demand against a shortage of high quality anime IP games; one example is Palworld’s recent success as the “Pokemon with guns” game, selling over 25M copies in a month across Steam and Xbox Game Pass. The second reason is that the anime IP licensing landscape is notoriously difficult to navigate for developers, creating a potential undersupply of games. Besides the language and cultural barrier, IP may be owned by different parties for different regions and different formats (movie, game, manga, etc.). Large aggregators like Crunchyroll and Funimation may already own exclusive licenses for certain IPs altogether. Some startups like Kasagi Labo, Layer, and Story Protocol are tackling this issue to make IP more democratized and easier to access. 

We believe that as new AI tools emerge and existing UGC platforms embrace AI, we’ll see the barriers to creation drop and the release of even more fan-created transmedia content. Roblox has already incorporated AI into user workflows, and Google’s DeepMind team recently announced Genie which can convert prompts into 2D playable environments. The Tencent team also recently unveiled GameGen-O which pushes the boundaries for generated open world games like Genshin Impact. 

Beyond UGC platforms, AI models and tools are enabling first-time creators to make compelling anime content that previously would only have been possible with a team of professionals. For art, Nijijourney is an image diffusion model focused on generating high-quality anime images from text, a collaboration between Spellbrush and Midjourney. There are also various fine-tuned Stable Diffusion models (a few here hosted on citiv.ai) that capture different styles and poses, some trained on SD 1.5, others like Animagine XL V3 on SDXL 1.0. Yodayo integrates AI generated anime art and character chatbots together, and pixai.art contains models that mimic popular IP (Genshin, Pokemon, Blue Archive, etc.). 

For animations and video there are several options available for the aspiring creator. RunwayML, Luma Labs, Hedra, and PikaLabs have built foundation models for text-to-video generation. One can use domoAI for anime-specific polish, and MagnificAI and Topaz Labs to enhance images and image-to-video prompts. ElevenLabs is used for voiceovers and dubs, and Udio and Suno are used for music. For manga/comics, there are companies like Shortbread that are helping to create webcomics and manga more seamlessly, with context and consistency between scenes and shots. There’s still a way to go to improve model accuracy, speed, and cost, but it’s clear that as these tools evolve they will empower more creators that might previously have lacked technical or artistic skills to bring their anime ideas to life. 

Anime Game Studios Lead the Way 

On the other side of the spectrum, professional game studios are leading the charge for high production-value consumer experiences that build on or create new IP. Anime games are some of the highest grossing in the games industry, accounting for 20% of spend on the mobile app store despite only having usage penetration of <3%. For example, Monster Strike has out-earned Candy Crush ($7.6B cumulative revenue vs. $6.9B) on a much smaller player base of 10M MAU vs. Candy Crush’s 70M+ MAU (SensorTower, 2016-19). Meanwhile, miHoYo’s Genshin Impact is on track to be one of the highest earning games of all time at its current trajectory (see below), with nearly $4B earned since launching just over three years ago.

There are two ways that anime game studios broaden the horizon for players. First, they usually create the highest quality games of the most popular IPs like Dragon Ball, Pokemon, or Dragon Quest. Six of the top 10 highest grossing anime games are based on existing IP, allowing fans to interact directly with the characters themselves. Second, they often innovate on core game loops or technology: miHoYo’s Genshin Impact was one of the first true cross-platform anime games, GungHo’s Puzzles & Dragons mixed match-3 and RPG mechanics for the first time, and Riot’s Valorant was one of the most accessible entries into tactical first person shooters. There are also  several stealth startups experimenting with weaving LLMs into visual novel gameplay. 

So, what makes these games so compelling and why do they make so much money? We theorize that there are a few key components that contribute to their success: 

• Proven, better, new. A lot of the games that have become breakout hits like Pokemon Go, Genshin Impact, or Puzzle & Dragons combine familiar systems with some form of unique gameplay innovation. For Pokemon it was the unique location-based game system combined with being the first Pokemon game on the mobile platform and for Genshin it was combining the open world gameplay of Zelda with the RPG elements of a traditional gacha and true cross-platform play. However, for many games, the bulk of many systems will be borrowed from the existing status quo and best-in-class from other games: banner-based live ops and events, gacha monetization, power creep and player progression, character personality archetypes like tsunderes, etc. 

• RPG at its core. Almost all the highest grossing games this year were role-playing games (RPGs). In this genre, players have one, or multiple characters that they “role-play” as by leveling them up, fighting monsters, equipping items, changing builds, etc. This isn’t surprising given that RPGs make up the highest spend for all mobile games (31%) despite having some of the lowest downloads by genre (3%). They are simply very deep games that the most hardcore gamers tend to gravitate towards, and can monetize well through pay for power or gacha. 

• Strong / compelling IP. It’s not an accident that six of the top ten anime games license popular IP for their titles. In anime games, because character collection is a primary driver of gameplay and monetization, IP is very important to drive that core motivation. Story matters. 

• Gacha monetization. Gacha games have some of the highest average revenue per user (ARPU) in the industry and is a reason why Japan has some of the highest ARPU by country. Gacha traces its history back to gachapon, Japanese toy machines that dispense a random toy each time you put in a coin. The modern equivalent would be trading card packs such as in FIFA Ultimate Team. There is some controversy over this monetization practice, but more recent games have raised S-tier earn rates, added pity drops, and increased the drip of premium currency to balance out monetization and fun. 

For now, we’ve been covering mostly free-to-play (F2P) mobile games. However, there are several successful PC/console anime games as well: Doki Doki Literature Club, the Persona series, the Final Fantasy series, the Fire Emblem series, and Phoenix Wright, just to name a few. Many of these game franchises have been sold quite successfully through a premium business model. For instance, the Final Fantasy VII remake was one of the most anticipated in the series, and sold over 3.5M copies in the first 3 days. After three years, it’s now sold over 7 million copies total.

There are several startups that have taken up the mantle in the West to build anime games for a global audience, and we’re excited to see what kinds of new experiences they create for players. Odyssey Interactive is building competitive multiplayer PvP games like Omega Strikers. Pahdo Labs is building a UGC platform starting with the anime game Starlight Re:Volver. Believer is building an open world drawing on inspiration from Genshin Impact. Studio Sai has made a hybrid dating action game, while Starmi and Gaudium are both creating collectible hero gacha action games. These are just a few of the studios that are focused on realizing anime worlds for players. We believe that these studios will continue to push the boundaries of gameplay and technology, and we’re excited to see these new experiences come to fruition. 

VTubing Blurs Human-Digital Interactions

We’ve seen how AI companions, UGC, and games allow players different ways of interacting with popular anime IP. VTubing takes it to the next level where fans can become popular anime characters themselves. VTuber stands for “Virtual YouTuber” and refers to a content creator who overlays an anime facial / body rig that tracks their real time motion and projects a digital avatar onto the stream (e.g. Kizuna AI). This phenomenon likely spawned from virtual idols such as Hatsune Miku that exist as fully digital characters, similar to virtual influencers like Lil Miquela in the west. There are various things that VTubers will do on stream, usually a mix of singing, chatting, dancing, and gaming. It allows for the creator to adopt a different personality, and makes it easier to create content regardless of appearance or age. 

Today, VTubing is a relatively small but fast-growing community that has potential to ride the anime wave into a much larger segment of content viewership and creation. While VTubers only represent 0.4% of all streamers on Twitch, they represent 5.7% of total viewership. And in the first half of 2023, VTubing viewership rose by 28% despite overall Twitch views staying flat. Some of the top VTubers like Ironmouse have millions of viewers, and are professionally run by organizations like VShojo or Hololive. One day, VTubing could become just as ubiquitous as Memojis on Apple FaceTime or Snapchat filters.

Startups are tackling this market from a few different angles. Iron Vertex is the go to for professional rigs done by designers experienced in Live2D, but if you need a cheaper option, startups like Hyperonline or GoHibe are making it easier than ever to design your own avatar. Mtion.tv meanwhile helps users design the background and environment for their avatar with visual based programming. There’s also VTubers like Neuro-sama, which is an AI that represents a human VTubing streamer, and platforms that are native to the VTuber experience like Anilive. If anime companions represent the human-digital relationship and multiplayer games the human-to-human relationship, then VTubing sits in the middle as a unique parasocial relationship mixing digital and human interactions. 

Conclusion

Today, anime is more than an art style, it’s a content format and cultural aesthetic that has transcended gender, race, age, and geography to become one of the largest and most profitable creative industries. Anime is also leading the way for digital play, turning previously passive consumption of linear media into a new dynamic form of entertainment. We’re excited about how anime is shaping the future of interactivity through its embrace of new technologies and business models: from AI agents to UGC platforms, from artistry tools to novel game mechanics, from interactive VTubing to new anime platforms. It’s clear that the best is yet to come for current and future anime fans. 

You’re playing Duolingo.

This little green bird has used streaks, leaderboards, badges, XP, lives, characters, storytelling, progression, and more to ride its way to over a $7B market cap and $500M+ revenue. However, most people simply point to the features that can be slapped on top of an app rather than understanding the underlying psychology of why those features are compelling. Gamification design is important to achieve the goals you want.

For instance, former CPO of Duolingo Jorge Mazal chronicles in this essay on how their initial attempts at gamification (progress bar, move counter, referrals) all failed to move their engagement and growth metrics. Once they got the team identifying the right metrics (in their case, cumulative user retention rate) they could tackle the product features like streaks and leaderboards that really moved the needle and improved DAUs by 4.5x (!). Qualitatively, gamification can also improve knowledge retention amongst students too (longitudinal study of 617 students here). And of course, from a monetization standpoint games drive the lion-share of IAP-spend on the app store.

So it’s pretty clear why mobile app developers want to “gamify” their app. The problem is a lot of people talk about gamification from an optimization standpoint, trying to slap on features to improve engagement. But without understanding the core player motivations, it can be difficult to meaningfully move those target metrics. So I wanted to dive deeper into the role gamification plays and the psychology behind it, what player types each consumer falls into (killers, achievers, socializers, explorers), and muse about how games can be leveraged for several popular consumer categories: edtech, healthcare, fashion/e-commerce, etc.

This essay will be broken down into three sections:

1. The psychology of games

2. Types of gamification and their function

3. Gamified markets I’m excited about

I. The Psychology of Games

Games have been around since ancient times; the Royal Game of Ur is a two-player strategy racing game with some parallels to Backgammon that dates back to 2600 BCE while modern day chess traces back to the Indian Gupta empire of 600 CE. There are many definitions for what constitutes a game, but they all involve (1) players (2) interacting within (3) a defined set of rules to achieve (4) a formalized outcome. Often those interactions involve decision making and uncertainty; take poker as an example where the best players estimate the expected value of a bet based on the pot size and assigned probabilities of hands, but ultimately there is quite a lot of variance in the outcome (winning or losing money). There are two formal definitions of games that separate games into two categories: PvE (player vs. environment) and PvP (player vs. player):

1. PvE: A game has “ends and means”: an objective, an outcome, and a set of rules to get there. (David Parlett)

2. PvP: Games are a “system in which players engage in an artificial conflict, defined by rules, that results in a quantifiable outcome.” (Katie Salen and Eric Zimmerman)

Great, Robin, so how do PvE and PvP apply at all to consumer apps? They’re important because they tap into the core psychological motivations for why people engage in games all together: to compete, to solve problems, to socialize, to think, to be entertained, etc. Taking classes or trips together might be co-op PvE, while the entire culture of Wall-street bets and the leaderboards of Zwift / Peloton are certainly PvP. We can dive deeper into player psychology with Bartle’s Taxonomy of Player Types:

Bartle breaks down players into four different basic categories: killers, achievers, socializers, and explorers. The x-axis is the PvP or PvE part of the framework, whether you’re interacting with other players (PvP) or interacting with the world (PvE). The y-axis is the “vs” part of the paradigm, are you playing co-op with other players or are you playing competitively against them. Of course, modern day player game design has gotten more detailed in breaking down these psychographic profiles into specific sub-genres and audience types (e.g., the lone cowboy who explores the frontier and hunts down infamous outlaws for social gain falls into multiple categories), but this is the core root behind it all.

My theory is that all consumer apps fall somewhere on the spectrum of Bartle’s taxonomy, and that gamification can be used to expand breadth (reach a new audience/category) or depth (engage an existing audience more deeply).

Take social networks (Facebook, Instagram, Twitter) as an example. These primarily target socializers in both large (Twitter) and small (non-influencer Instagram) groups. However, they also target achievers with their system of likes and followers, and you have clearly defined players interacting socially to achieve a goal (increase your like / follower number) with a set of rules (social norms, text/image/video content of a certain size/length) and uncertainty in the outcome (some posts hit, some don’t). Of course, some may use social networks to PvP and engage in arguments with others, as that can be highly entertaining and viral, a meta-strategy for some top influencers. However, most people are just engaging in the social “read” part of the game, not in the “write” part, and even less in the “write against someone” strategy.

It’s a status game that we play.

Let’s look at some other categories:

• Travel (Airbnb, airlines, hotels) of course taps into the explorer in each of us, and most of the ads that you see are appealing to that primitive desire to explore. Airlines have long had literal status games that reward you for engagement, and it’s well known that people will go to extreme lengths to keep their premium seats and faster boarding.

• Finance apps generally use mechanisms to reward achievement: setting goals, saving more, investing better. Mint has rings to close, Venmo has semi-private social feeds, and Robinhood has, well, Wallstreet bets, the ultimate game that pits anonymous reddit users against hedge-fund veterans, or so they like to think.

• Dating apps target socializers as the primary component, but have added a layer of achievement on top with pay-to-win features (super-likes, roses, premium subscriptions for more swipes and algorithmic featuring). Tinder literally has a $500/month subscription for its highest tier members.

• Fitness apps are a combination of killers and achievers; you often have leaderboards so that the top competitors will go head to head against one another for glory (Zwift or WHOOP leaderboards), whereas those starting out on their journey might just be looking to get a streak going or lower their mile time to get that next badge (Apple Fitness rings).

Many of the common gamification “tools” that you can use to gamify your app will fall into one of these four categories too. The trick is figuring out which ones will actually enhance the experience for your users, because some could be detrimental to the core loop. For instance, if you designed a leaderboard in an educational app by proficiency instead of by time spent, that could discourage newcomers from learning because the goal will feel too far away to be achievable. Or, if you put a leaderboard in a stock-trading app by dollars earned or % returns, that could encourage more FOMO (fear of missing out) / jealousy / risk-taking behavior for people to aspire to those higher gains.

II. Types of gamification and their psychological function

1. Leaderboards (killers and achievers): These are commonly used to spur engagement through competition. The most important part is deciding what to reward in your leaderboard: skill, effort, time spent, output, etc. Traditionally competitive games like League of Legends reward pure skill (win/loss ratio), but not all consumer apps should use skill as the primary motivator. It works well for fitness (see Peloton example below), but social apps on the other hand might not benefit from a leaderboard since it would take away from the current use case for many people which is reading / socializing with their core group. For those that do want to play, there is a “mental leaderboard” of a theoretical number of subscribers / followers to aspire to.

2. Streaks (achievers, socializers): Streaks are a daily motivational behavior that create a sense of achievement from doing the habit every day, and strong loss aversion as you build up a significant streak. There are two types, singleplayer and multiplayer:

   1. The singleplayer streak is common in any daily activity - fitness, education, meditation, healthcare, etc - and can be used to reward a specific type of behavior. There are even apps like Conjure or Streaks that allow you to customize your own types of habit trackers. Games have advanced versions of streaks that serve as strong engagement and monetization vectors that other consumer apps can learn from:

      1. Battlepasses that give you rewards as you advance and will usually have a free and paid tier to increase % conversion

      2. Check-in rewards that give you different bonuses to encourage you to try out different parts of the game

      3. Come-back rewards where you get boosted rewards when you are a resurrected user that has stopped playing for a bit (increases resurrected user retention rate RURR)

      4. In general, games will have multiple “streak” type features that give a sense of achievement: user level, character level (often with multiple characters), battlepass progression, daily check-ins, beginner quests, daily challenges, and more. Just check out the home screen for Supercell’s new game Squad Busters. By the way, you can pay to maintain your “Top-5” streak if you miss a game in the Top-5; many games will do this to monetize loss aversion.

   

   1. Multiplayer streaks involve two or more players to work together to maintain a streak. This was primarily popularized by Snapchat in 2016, with the launch of their streaks and best friends feature. The streaks are socially reinforcing and require multiple people to participate, which was a genius design choice because if you have multiple streaks running then you have multiple people applying social pressure to sustain your engagement with the app. The best friends feature is essentially a semi-blind leaderboard, where you could tell if you’re on someone else’s list but not your exact position. This helps add some ambiguity and variance to the social game which keeps things exciting.

3. Characters and narratives (socializers, explorers): Compelling IP and stories can make players more emotionally attached to the core loop. This can be particularly engaging for socially motivated players that want to help others (including virtual companions!) or uncover mysteries.

   1. Many casual games like Merge Mansion or June’s Journey have innovated on the genre through adding IP and narrative. The core loop becomes that much more interesting when there’s a metagame on top.

   2. You may have also seen Royal Match’s “Save King Robert” interactive ads which invoke the player’s inner heroism.

   3. Duolingo’s Duo is probably the best example for a consumer app; the team has done a great job humanizing Duo and giving it a lot of personality. I mean, don’t you feel bad that you made Duo cry because you didn’t learn Spanish today?

      

4. Badges and XP (achievers): Badges and experience points (XP) are a staple in gaming used to reward players for completing specific tasks or reaching certain milestones, giving a sense of accomplishment or progression. They can be used to encourage desired behaviors or celebrate user achievements (more detail here). Headspace, for instance, grants milestone badges for minutes meditated.

5. Lives (achievers): This is a common game mechanic where players have a limited number of attempts or "lives" to complete a level or task before having to wait for a cooldown period or purchase more. This creates stakes and urgency, and mobile games also use it as a monetization technique, tuning levels to fit player skill levels and sometimes failing players right before they finish the level. Lives have to be balanced carefully for consumer apps, since they can also lead to reduced user engagement if you lose all your lives or psychologically don’t want the pressure. Duolingo employs a "hearts" system, where mistakes deplete the user's hearts, in order to discourage binge-learning and check information retention.

   1. One interesting vector for consumer apps to think about is whether they could create idle progression mechanics (AFK Arena) for monetization and retention. Banked steps in fitness games or time-based fetch quests in fashion games would be examples here.

6. Guilds / Social (socializers, achievers): Guilds, teams, or social features foster collaboration, competition, and community within an app. They appeal to both socializers who seek interaction and achievers who are motivated by group goals and recognition. Fitbit, for instance, allows users to join groups, participate in challenges, and compare stats with friends, leveraging the power of social accountability and support. Of course, these derive from the guilds of old MMORPGs like WoW, Runescape, Ultima Online, etc.

7. Likes / shares (socializers, achievers): Social validation mechanisms like likes, comments, and shares are powerful tools for driving engagement and virality, catering to those who seek validation for their contributions. Don’t think I need to give examples with this one :)

You’ll see that most of these features normally target achievers, because most apps are inherently singleplayer and thus use RPG-esque (roleplaying game) mechanics to enhance engagement. Managing your finances, booking a flight, learning a language or tracking your fitness start as singleplayer activities. That’s why a recent wave of innovation has been adding the social layer on top (e.g., Pinduoduo for fashion or Strava for fitness). What I’ll be interested to see is how apps change as they adopt deeper gamification practices and look increasingly like games themselves.

III. Markets I’m excited about

This essay got longer than I intended it to be, so I’ll cover a few areas here in brief that I’m excited to see gamified app experiences in. Each sector could be its own essay and market map!

• Healthcare: There are two areas where gamification could be useful, what I’ll define as strongform games and weakform games:

  • Strongform games: These are games that have specific clinical endpoints that they target in order to improve outcomes, which generally tend to be mental health / depression related. There are a lot of great startups I’ve met innovating here, from Hero Journey Club which conducts group therapy in Minecraft to Hedonia’s mood bloom for reducing depression to Mindforce Lab’s companion/pet game, among many more!

  • Weakform games: These take the form of gamified companion apps that work to improve patient adherence and engagement, one of the most common issues with the healthcare system. Are patients taking their medication / rehabilitation seriously? Examples are apps like MySugr or Endo Health for diabetes tracking and Noom or Plant Nanny for diet / nutrition.

• Fashion: Pinduoduo and Whatnot revolutionized the social shopping experience, making it communal and live, like a modern version of old auction houses. Meanwhile, Roblox and Fortnite are experimenting with what it means to bring fashion into a 3D digital space. While I don’t think we’ll be shopping in a 3D metaverse just yet, we already conduct a lot of our research, browsing, and shopping online (~30% of sales in Western markets). What I would be interested in is a modern take on Kim Kardashian Hollywood, Covet Fashion, Love Nikki, and other similar dress up games in a modern, multiplayer context.

• Music: Rhythm and dance games like Just Dance, Guitar Hero, and Dance Dance Revolution have been around for ages, just take a look at any asian arcade and you’ll see a wide variety of all the ways that people will mirror their actions to the music. But most music apps are still surprisingly singleplayer experiences: just read, no write, very little social. Tiktok is the biggest innovator here of course, but I’d love to see some more exploration for other game formats and music. What would be an inherently social experience that uses music as a primitive? Soundmap is a great example that combines the collection and trading of Pokemon with the familiarity of artists and songs.

• Fitness: Gamification has enormous potential in the fitness industry, as it can help users stay motivated, engaged, and consistent with their health goals (Johnson et al. conducted a systematic review of the existing literature here). Apps like Nike Run Club use a combination of challenges, badges, leaderboards, and social features to incentivize physical activity. Zombie Run! is another excellent example that overlays a narrative-driven game onto running, turning exercise into an immersive, story-driven experience. And there’s been a slew of great startups building interesting new experiences: Talofa Games, Genopets, and of course, Zwift.

• Education: Let’s end it how we began it. Researchers have found that gamification leads to statistically significant improvements in motivational, behavioral, and cognitive learning outcomes. While Duolingo has conquered the language learning market, but there are many other areas rife for disruption: mathematics, writing, science, etc. Kahoot! is another good example of gamifying learning through competitive quizzes.

And of course, there’s more than I haven’t mentioned here: productivity and finance being just a few of them. If you’re a founder building in any of these areas, let’s chat!

The current state of the games industry was looking bleak for quite a while last year. Several layoffs at Riot, Unity, Activision Blizzard, Discord, and more recently Take-Two have had chilling effects on the industry while games VC financing for the industry has returned to pre-pandemic levels (see chart from Pitchbook below). However, remember that many of the greatest tech companies were created during market downturns: Uber, Airbnb, Slack, Whatsapp, Square, Venmo. We believe the same will be the case here. Zig, when others zag. 

There is hope in the market now too; Konvoy just reported the first uptick in venture financing for games in 9 quarters. At a16z we just announced our $600M GAMES FUND TWO; Bitkraft has raised a $275M Fund III; PlayVentures has raised at least $78M for their third fund; and there are new funds like Laton Ventures that are coming on to the market too. That’s not to say the financing market for studios will be easy. Founders who raised in 2021-22 will be coming into a market that’s much different than the ZIRP (zero-interest rate phenomenon) environment of 2021 when money was more freely flowing. That’s because interest rates impact the opportunity cost of capital (among many other factors: venture debt, velocity of money, inflation, etc.). Investors during the ZIRP era were looking for more risk-on assets to deploy capital into because the threshold for returns is so low; investors have to take risks otherwise their cash is depreciating due to inflation. But when interest rates are high, investors tend to go risk-off because the opportunity cost is much higher, such as investing in stable T-bills returning 4-5% per annum. 

This dynamic particularly affects games. Because games take so long to bake and there’s not a ton of signal gained during that process, the capital at stake is considered quite risky and illiquid over a long period of time. Building a consumer mobile app might cost several hundred thousand dollars or less to full launch (a few good engineers hacking at it for a few months), but a game, to build just a MVP, could take at least a year of ten or so talented artists, engineers, and developers building the concept and often costing several million dollars. 

To illustrate this, the best example is the J-curve. The concept applies to many things, from careers to investments to global trade. For venture capital and startups, it’s the amount of capital that needs to go into the startup before you start seeing glimmers of hope and product-market fit. For a studio, not only is the time to market longer, but also the capital required and the time horizon. This also affects investor IRR (internal rate of return) potentially, since the time needed to return capital is longer. It’s also why, in my opinion, VCs might be more excited about crypto games, because they have faster time to liquidity (tokens) and more options for earlier monetization (NFTs) from a very practical standpoint. 

It’s similar to biotech or hard-tech investments, where the investment and time horizon is quite a bit higher, but the problem with games is that they don’t quite have the same moats as other products. With biotech you have patents; with hardware you have proprietary IP and physical capex (factories). There are also alternative financing and acquisition options with pharma snapping up promising preclinical drugs and governments granting large contracts to R&D products. With games, you’re competing for consumer attention that’s split among apps, videos, music, and more. That’s why investors are always looking for differentiators with games: faster GTM, cheaper UA, network effects, platform plays, UGC lock-ins. 

One thing I’ve been debating in particular is what the moat for a game could be. Player liquidity is a necessary but not sufficient condition for a multiplayer game that has differing skill expression, but it’s not a true network effect per se. With a ridesharing app like Uber, every additional driver reduces potential wait times for riders, incentivizing more riders to use the app which in turn draws in more drivers. But for games, after you hit a certain threshold of players then I’m not sure additional players are additive for enjoyment. There are a few moats that I’ve been thinking about:

• UGC is clearly a two-sided marketplace moat. More players incentivizes more creators which makes the platform more valuable for those players. But getting the high quality creators that create the best experiences is the most important thing. 

• UA could be a moat, if your LTVs are just better than other competitors, so you can bid and win better inventory and get the highest spending players. 

• Creator content and communities could be a moat. The more players that create content on YouTube, TikTok and other social platforms for a game, the more enjoyable it is to partake in that community vs. other games. But I’m not sure how defensible this is over time. That being said, strong social bonds and guilds can definitely be a moat (think of the corporations in EVE Online and how persistent they are). 

• Game specific-knowledge could be a moat. It’s harder to swap from League to DOTA because you have to relearn all the specific champions and mechanics, whereas it’s much easier to swap from CS to Valorant since a lot of the core movement and aim mechanics transfer over. 

Okay, so we talked about some of the dynamics of the market and game studio investing. In the following sections of this essay, I want to share my opinion on what investors are looking for, and propose some ideas for how to think about studio financing. 

The core questions that investors are thinking about at each milestone for a game are: 

• Seed: Is this the best team to tackle the opportunity and will they be able to get enough capital to get this game to market? Is the pitch differentiated enough and are they riding some sort of market tailwind that will allow them to become a venture-scale outcome? Have they thought through all the important company building parts: product, engineering, GTM, talent, and BD? 

• Series A: What’s your production velocity and how much have you shipped? What does the team still need to prove about the core game thesis? Can you get to at least an open beta with this tranche of capital, and is this playable the top 1% that we’ve seen this year? 

• Series B: This will likely be a post-launch round. There might be a non-dilutive publishing round after the Series A, but generally growth investors are investing against metrics rather than ideas. As such, they’ll be asking for the game metrics: D1 / 7 / 30 retention, DAU/MAU, LTV:CAC, CPIs, ARPU, ARPDAU. Is the game on the path to becoming a $B franchise and how will they scale the studio for Games 2, 3, and 4? 

If I were to really oversimplify some of these things, then I’d boil it down to this: 

A few common pitfalls to avoid if you’re going for VC-studio financing: 

• The game pitch isn’t ambitious enough. Remember the J-curve. The only way the J-curve works is if the payoff at the end is potentially massive to warrant the investment at the beginning. As much as I love games like Hades or Hollow Knight or Slay the Spire, VCs are looking for $B outcomes so there has to be a story that leads to that. You can take inspiration from indie games (they’re actually some of the most fun games out there) but there has to be a business model that aligns with how VCs work (generally multiplayer + live-ops, sometimes with a UGC angle too). 

  • Remember that when VCs are funding a studio, we’re not funding the game, we’re funding the company. This is an important distinction because the company has to evolve past a one-game studio into a multi-game publisher with an inherent distribution advantage through its own audience and platform. 

  • The reason why publishers will invest in smaller premium games is that they have options to de-risk their investment: production milestones for the studio to hit, revenue share to recoup their capital outlay, and more involvement with the governance and management of the studio itself. 

• The team hasn’t proven enough by the Series A. Series A’s are really murky in this environment. Compared to traditional consumer Series As where it’s a product-market fit bet (e.g., this social app has strong DAU/MAU, month-on-month growth, and D30 retention and we believe the unit economics will work at scale), games studios often haven’t gotten to traction yet at this point. So VCs are forced to go off of more qualitative things like team velocity, production quality, genre thesis, systems design, combat feel, etc. We have to predict that this product will be a large outcome without traction and “invest in our own genius,” which is always dangerous. It’s always easier to bet on something that’s working. 

  • So what can you do? In my personal opinion, the best thing that founders can do to increase their chances at a solid round are to be extremely disciplined about burn, have a strong thesis about the genre they’re going after, and execute against that vision as quickly as possible. Prove a lot, with a little. Build in public. Consistently test your game with players earlier on. 

• There isn't a unique insight. We’ve heard hundreds of pitches and so there’s a lot of genre mashups we’ve already seen, everything from Animal Crossing MMO to accessible Tarkov to Dark and Darker with guns. All of these are fine and dandy (and many could work!) but we want you to drill down to the nuts and bolts and tell us things we haven’t heard before. Find your edge!  

  • League succeeded because of the unique insight on DOTA/DOTA2 being very popular mods with terrible distribution, and capitalized on a new F2P business model and jump-starting from the existing community

  • Valorant had the unique insight of a strong tac-shooter community that hadn't seen innovations in years, and combined that with more recent hero-based shooter metagame. 

I have a few ideas for founders and investors that might increase the odds that the studio is more likely to ship the game while still retaining some of the necessary pressure of budget + production targets. I’d love any feedback on these as I’m just brainstorming, so feel free to email me if you have thoughts: 

• Larger seed rounds for more equity. One idea would be to raise larger seeds ($5M+) that give the startup more capital to get to a more fully formed product, but take more dilution early on (30%+, potentially co-led by two VCs). There are clear pros/cons to this one, since the founder will be taking on more dilution early but they would also have more money to work with in order to get the game to a solid second round. There’s an argument that the risk that early stage VCs take isn’t properly baked in for games, and so that should result in different deal structures. 

• Combo VC-publisher deals. The VC and publisher would work together to co-lead an equity + publishing deal at the Series A that will fully fund the game to launch with a marketing budget. This is similar to what the $31M Gardens deal had to fund their new cozy co-op game. The main advantage is to fully fund the game and de-risk downstream capital needs. The VC and publisher also serve to counterbalance each other, which generally will be better for the founder overall (better publishing terms, de-risked downstream capital). 

• Faster, leaner GTM. Lethal Company, Battlebit, Palworld, Among Us, and a slew of other viral hits from previously unknown developers who had unique insight into their genre show that there are other ways to create content besides the AAA studio model. I’d encourage founders to re-think how they might run prototyping, production, and GTM. Some ideas could be:

  • Outsource non-core development and find diamond in the rough talent internationally

  • Lean into emerging AI tools to quickly iterate on production (animation, SFX, VFX, voice, code, etc.)

  • Build communities earlier and share the development progress with them on Discord, Twitter, TikTok, and any social platform 

  • Playtest, playtest, playtest 

• Pulling up revenue generation. This is basically the Kickstarter idea, or selling early access / founder packs. The main risk here is that you don’t deliver on the product to your players and you aren’t faithful to your promise. However, it can be a useful way after you’ve built a community to continue to service a loyal audience and get them invested in your success. 

The games industry is at an inflection point. I believe that there will always be great gamemakers out there who are pushing the boundaries of play, but we need to make sure that the system is set up well to find these founders and ensure that they get a good shot at getting the capital they need to get the game to market. That’s why we built SPEEDRUN, to work with more founders building in games x tech and support them as they make their vision a reality. 

Subscribe now

1. A singleplayer mobile game with simple mechanics, midcore meta-progression, and strong monetization.

2. A co-op multiplayer rogue-lite that draws inspiration from roguelikes but is tuned for a native multiplayer experience.

I’m not a designer of course, but I’ve spent a lot of time in the genre and was inspired by recent poker-roguelike hit Balatro to put pen to paper on where this genre might go. Let’s dive in.

What’s a Roguelike Anyways?

For those that are less familiar with the genre, roguelikes are characterized by two core elements:

1. Randomly generated levels, including rooms, loot, enemies, and progression paths that are partially or fully generated by an algorithm.

2. Permadeath, where each run ends with the player losing their progress and starting anew.

These pillars make roguelikes compelling for several reasons. First, player skill is crucial, both in mastering the core combat mechanics and understanding how the interplay of abilities stack to create the best builds. For instance, in Hades - an action-roguelike where you play as the son of Hades trying to break out of underworld - part of the game is mastering the mechanics to dodge attacks, predict enemy movements, weave around rooms, and strike at the right time. And the other part is understanding how to min/max the right build given the randomized boons (powerups) you have in each run. What’s unique about roguelike games is that it forces you to make strategic tradeoffs that are different each run: Should I get an Aphrodite attack debuff? Or a Poseidon knockback? Perma-death means that each decision matters, and good players have many ways to show off their skill expression. However, it also means that the genre is fairly hardcore, which Hades solved well by layering in narrative progression with each run that meant you were always moving the story forward.

Second, each session is unique. Because the power-ups, enemies, and levels in each run are different, players are forced to adapt to each run differently. There are different synergies between all the powerups that allow for many types of strategies and win-conditions. It feels really good to finally, through luck and smart choices, get a “cracked” run that feels extremely overpowered, creating a “water cooler” moment that encourages players to share their newly designed metas. Take Slay the Spire as an example, which has a card-based combat system. Even though each character only has a limited set of cards, cards interact with each other uniquely, exponentially increasing the number of decks possible. In the picture below, you can easily see how the Silent might want to stack “Bouncing Flask” and “Crippling Cloud” that both apply poison with “Catalyst” that doubles the enemy’s poison. But if you don’t get those cards in your run, then you might have to run a Shiv-strategy or an energy-strategy, totally different from your initial intention.

Finally, roguelikes solve the content treadmill problem with player vs. environment (PvE) games. A lot of games, from casual games like Candy Crush to MMOs like World of Warcraft, have a set amount of content for players to grind through. The speed at which players will churn through content is often faster than the developer’s ability to create it. Roguelikes’ procedural generation and varied combinations allow for much higher replayability given a set amount of content. Of course, that also means they’re much harder to design, since you have to deeply think through how to properly balance all the components of the game.

The Mobile Singleplayer Roguelite

So why haven’t roguelikes gone mainstream yet? Well, actually, some of them already have. According to SensorTower, Survivor.io from Habby peaked at ~25M MAU after launch and still generates $120M+ in revenue annually. Habby took the proven, extremely accessible loop of Vampire Survivors - the only player input is movement through a virtual joystick - reskin it in a more accessible casual art style, and add modern meta-progression and monetization techniques (battlepasses, piggy banks, crafting, lootboxes, etc.). They also ran a smart TikTok strategy for UA with a vertical game format that was perfectly tuned for the platform.

The main challenge that these singleplayer mobile roguelikes face is that they need to tune the roguelike genre for a more casual audience. Permadeath is a pretty hardcore concept because it mandates loss as part of the core loop. Think about Monopoly Go as a comparison point, where even when you lose (i.e., your attack gets blocked), you still win cash in a colorful and satisfying explosion.

That’s why many of the successful mobile games will be rogue-lites; the player still dies but every run feels meaningful because they unlock coins / abilities that permanently power up their character. The challenge that roguelites face is that the metaprogression (more HP, more damage, new weapons, etc.) mean that the game is hardest at the start and gets easier over time, a monotonically decreasing difficulty curve. This is the classic depth vs. accessibility challenge.

The other important facet is making the game easy to intuit and learn. That’s why games like Vampire Survivors or Survivor.io do so well, because there’s literally one mechanic, movement. However, there’s more depth to the combat than you would think; learning the various skill trees, positioning attacks properly, and slipping through groups of enemies take time to learn, and better players will impose their own limits like no-hit highscore runs. .

That’s also why Balatro presents an interesting take on where roguelikes might go in the future. Most of the roguelikes we talked about up to this point are focused on combat. But Balatro takes the 52-card deck that we are all familiar with and gives it a unique roguelike twist. Add to it the simple satisfying loop of “number go up” and you get a smash hit. And with all roguelikes, there’s also depth: trying to quickly assess what your best hand can be, what cards are left in the deck, and making some risk-adjusted guesses on the next play.

I’m curious to see what other non-combat roguelikes might spin up because of Balatro. Maybe there will be interesting platformers, runners, or board games that people experiment with. Technically, even Flappy Birds is somewhat of a roguelike with its algorithmically generated pipes!

The one other important callout with mobile games is monetization and meta-progression. Currently many of the indie roguelike games have traditional premium-box monetization for $20-30. While this is certainly one of the best player-value deals out there, it won’t result in the billion dollar outcomes that VCs and publishers are looking for. For that, we’ll need to look to traditional F2P monetization techniques while balancing the game to not be too pay-to-win (P2W).

Yang at End Game is doing some interesting exploration here. One of the early game prototypes (see below) they’re developing combines the roguelike simplicity of survivor.io and the hero collection / monetization of Genshin Impact. Having a roster of heroes solves some of the pay-for-power issues by allowing for different meta team comps rather than one overpowered weapon that you can buy. Banners and live events can mitigate some of the power creep issues that roguelites encounter. It also introduces more depth of spend; Genshin Impact, one of the top grossing hero collectors, has some of the highest revenue per download (RPD) at $95 in Japan and $35 in China. Meanwhile, Azra Games is another company that’s using roguelike mechanics to enhance the replayability of its next-gen hero collectible RPG.

The Multiplayer Genre Mashup

The other direction that I believe will bear a lot of fruit is experimenting with what a multiplayer-first roguelike might be. Singleplayer roguelikes encounter a cursed design problem when you add another player. If one of the players die, do you continue the game onwards while the other person watches? Not very enjoyable for the weaker player. If you have a revive mechanic, how often can you use it? Use a revive too frequently and you remove permadeath and the tension of the game. Some of the games like Spelunky or Rampage Knights solve this by allowing the dead player to turn into a ghost that still has some core actions until they’re revived (similar to Among Us), but it’s still not a perfect solution.

Instead, it’s likely that a roguelike that’s designed to be co-op first will be more successful. However, there’s many different directions that this could lead. There’s No Man’s Sky, which generated near-infinite levels at the start for players to explore and colonize together. There’s Diablo, one of the original ARPGs, which took a lot of inspiration from roguelikes for its procedurally generated dungeons and random loot drops. There’s Don’t Starve Together, which has a co-op twist on the original survival crafting roguelike game with its large, procedurally generated map. And of course there’s Risk of Rain, which is the most traditional co-op roguelike that scales the difficulty, monsters, and loot up and down depending on how many people are playing.

The key you might be noticing with these games is that they’re more rogue-lite than rogue-like. Because of the hardcore nature of the most punishing, skill-based roguelikes, most of the multiplayer roguelites will likely borrow features from its parent genre to enhance its gameplay: procedurally generated levels, randomized loot / enemies, a variety of synergistic abilities to collect, and death / revivals as part of the loop. These multiplayer roguelites will likely have some sort of genre blending (FPS, ARPG, survival-crafting, etc.) and have a premium + liveops model in order to smooth the revenue curve over time. Helldivers 2 is a good example of what the future could look like here, with a galactic war map that changes and evolves based off the collective of player actions. There’s quite a few startups that are going after this thesis. A few are in stealth and so I won’t reveal them here, but Fuzzybot is one that’s public and combining roguelite ARPG combat with the crafting/collection loop of a life sim.

Conclusion

Roguelikes are one of my favorite genres of games, and I’m excited to see where they evolve from here. I must also mention the obvious potential that AI has here to change the kinds of games that can be created here. Levels could be generated with more detail and with more immersion, NPCs could react in more intelligent ways, and live ops could be personalized more effectively for the player. It could also change the kinds of games that we play. Suck Up! for instance, is a roleplaying game where you masquerade as a vampire and have to talk innocent townspeople (powered by LLMs) into letting you into their houses. There’s no perma-death with this one, but talking to LLMs is inherently random with many branching paths, similar to the level and enemy randomness of a traditional roguelike.

If you’re building in the genre and innovating on what’s new, feel free to reach out!

Subscribe now

Finding the Reason You Can Win

This is both on the personal and professional front. I’ll start with the professional first.

VCs are looking for alpha like any other good investor (excess risk-adjusted returns relative to a benchmark). However, that alpha can be much harder to find due to the sheer amount of variance in early-stage investing, and because markets are getting more efficient (e.g., the Benjamin Graham-style cigar-butt investing would likely not work in today’s day and age due to the efficiency of PE markets at pricing private companies). So we look in different areas: teams, genres, stories, metrics, insights, etc.

Those reasons can come in different shapes and forms, but there needs to be some spike on which the startup hinges its success. Perhaps it’s a clever go-to-market strategy like Content Warning, where they discounted their list price to $0 on April 1st to solve the cold start problem. Or it’s a strategy that leans into existing strengths like OTK launching Mad Mushroom, a new-form publisher that leverages this creator-led distribution meta. Loftia is another example of distribution strength, with a Kickstarter campaign that generated $1.2M in pre-sales and 500K+ followers across TikTok and Instagram. Palworld, legal troubles aside, was only able to go viral by latching onto a popular IP and giving it a spin that no one would have thought of: Pokemon with guns.

It can also be a spike in design. That’s a little bit harder to show, but qualitatively we evaluate whether you’ve really gone deep on the specific genre that you’re targeting. Have you thought through your audience and all the player motivations, have you played all the games in genre, have you identified the gaps, have you identified where you’ll innovate and why that’ll be compelling, have you figured out what’s going to really make this new game tick. Stephen Lim has always been my guru here, and I’ve learned a lot from him on this subject.

It can be a new market that you’re one of the leaders of right now. Primodium is led by two ex-YC brothers building in the frontier of fully on-chain games. Zaranova is creating an AI game where you must pretend that you’re an AI and trick others to believe you, while Altera is building some of the best agents in games starting with Minecraft. Trass Games just launched Yeeps, a new type of hide and seek game that leverages the unique tactile affordances of VR. Any new market is going to seem early and the games like toys, but this is where the next wave of innovation occurs.

It can of course be the team too. Which is what matters the most at the end of the day. There’s lots of great teams out there and they come from all types of backgrounds, but to provide a gross generalization they all have high IQ and high hustle at a minimum. On top of that, they’re often able to have a strong ability to manifest their vision into reality, whether that be the product, the capital, or the people necessary to get the job done. Just know that as VCs, it is our duty to do the diligence through reference checks to assess whether each team is really top 0.1% in their field, and it’s the founder’s job to prove that they are the best.

On a personal front, any job that’s worth doing is going to be a grind, whether that’s developing a game, founding a company, or investing in startups. If there’s one thing I’ve taken from my time here it’s that there’s a lot of really smart, talented people going after great ideas and working their butts off, and even then they might not achieve what they want. But if you don’t try, then it’ll never happen.

So at the end of the day, you have to have faith in yourself, and the people around you. People’s perception of you might change, they might disparage you on this or that, so what matters the most is executing against what you believe to be true.

Capital Requirements in a Shifting Market

You might be thinking, ok Robin, great but how do all these theoretical concepts apply to me? Here’s some very tactical advice for aspiring founders and game devs out there: find the right-sized idea for the right-sized founder.

So for instance, if you’re a new developer starting out and you want to make the next big MMORPG that’ll cost $50M-100M+ in budget, that’ll be a vision that’s really hard to execute on and get the capital for. You earn the right to raise big sums of capital, and only very few industry veterans can command the type of team and cache required to go after those opportunities. Instead, figure out a way to build up to that grand vision.

Look at the story of Mihoyo for instance. Started by college students hacking away at anime games, they only raised one round of capital to launch their first game, FlyMe2theMoon, a puzzle game based off of the ending of Neon Genesis Evangelion. They tuned their skills with Houkai Gakuen 1, 2, and 3, and used the 3D engine of the 3rd iteration to eventually develop Genshin Impact, their landmark title that generated over $5B in spend just on mobile alone.

The capital markets are also changing a lot in the gaming landscape. A decade ago, there were very few VCs that invested directly into content and platform-based publishers. Then Bitkraft (2015), Makers Fund (2017), and PlayVentures (2018) came onto the scene. Now there are mainstream tech VCs that have taken notice: Index investing in Backbone, General Catalyst investing in Triumph, Lightspeed investing in Gardens, etc.

However, the capital markets for content is still quite small compared to the hundreds of VCs that will fund the next enterprise SaaS or AI startup. Those markets are more developed, and there are sophisticated growth investors that will pick up the tab once you get to scale. As a studio, there are a decent chunk of options when you’re starting out (including our games x tech accelerator SPEEDRUN). However, due to the budgets that some games need downstream financing can always be a concern. There are only a select few VCs and publishers (often from Asia) that are willing to foot the $15-40M Series A/B checks that might be required to finish the game. So how do you get those?

Execution Is All That Matters

One of the founders I work with Christian says it best in his pitch here (PW: SR002). In games, there are fewer metrics to go off of if you play in PC/console land. Compare the Series A of a game studio to the Series A of a consumer app, and you’ll be comparing a vertical slice / beautiful corner to an app with hundreds of thousands of MAU and strong D180 retention. There are two paths here:

The first is just making the game really, really, and I mean really good. We play about a hundred games a year give or take, and a few of them stand out amongst the crowd. What’s important is to have a strong thesis about the genre that you’re going after, and really execute against the pillars of that thesis. If it’s a shooter then your gunplay should feel top notch. If it’s a survival-crafting game then the core building loop should be well fleshed out. A few studios that have impressed me in the last year: Red Rover, Ruckus Games, Gardens, Mountaintop, Mainframe.

The second is finding ways to GTM early and fast. Theorycraft’s done a great job here of hosting progressive playtests with players to get early retention / engagement data. Odyssey Interactive partnered with renowned animation studio TRIGGER for a killer trailer, and also launched their game PC first (and rolled cross-platform quickly after) with a clever Twitch streamer tournament. Web3 has also done some really clever things here with airdrops, token incentives, quests, early land (Pixels and Parallel are defining the early playbook here).

Remember, traction is king, so the more you can do to show that, the better. In a rough stack ranking of what VCs care about:

1. Users: DAUs // MAUs + Retention

   1. We care about this the most because it represents product-market fit

2. Monetization: CAC, ARPU, LTV, ROAS

3. Engagement: session times, # of sessions

4. Community: Size of Discord, # of Twitter followers, GMV

Concluding Thoughts

I often get asked what it’s like to be a games VC, and I tell them it’s the hardest job I’ve ever done - lots of uncertainty, long hours, very non-deterministic - but I wouldn’t want to do anything else. In what other job do you get paid to listen to founders tell you about their dreams, and help them achieve them in some small way. It’s really, really cool to just see people and companies grow and evolve over time.

I hope these essays will continue to be helpful to the community and give some more insights regardless where you are. I plan on writing weekly, on Mondays, on games, web3, consumer, anime, tech, UGC, VC, philosophy, and more. If you liked this please hit subscribe to help me grow!

Subscribe now