A Career Guide for People in Their Early Twenties

Marc Andreessen has this idea that you cannot plan a career, and that the smartest thing a young person can do is stop trying. I'm stealing it because it's the single most useful thing I know. I know it sounds like a cop-out. It isn't. About a year ago I got rejected from an investment banking scholarship, and the funny thing is I wasn't even sure I wanted it. Banking was just the thing I had told myself I wanted, for so long that I had stopped checking whether it was still true. If I had kept forcing that story, it would have been one of the worst decisions I could have made.

So I did the opposite. I came out to the Bay with no plan and an open mind, fell in love with the startup world, and then got essentially every job I never imagined doing: venture, growth, talent investing, talent recruiting. Now I run my own fund. None of this was on a roadmap. There was no roadmap. There was just a willingness to follow what was interesting and work with people who were better than me.

Here is the frame that actually helps, and it's where Andreessen's point gets practical: treat your career like a portfolio. Every job is a position with its own risk and its own return, exactly like an investment. Some are safe and pay cash. Some are volatile and pay in learning, access, and optionality. The mistake most people make is loading up entirely on the safe positions when they're young, which is precisely when they should be doing the opposite.

The real reason this matters is compounding. The most important thing for building anything over a life is your rate of compounding, not your balance at any single moment. If you believe you can compound faster than the average person, and you believe people are going to live a lot longer than they used to, then the highest-return move you can make early is to increase that rate. You do that by learning as much as possible as fast as possible. There is almost no reason to optimize for short-term wealth, because a slightly higher compounding rate, run over a long enough life, dwarfs any early paycheck.

And in your early twenties your costs are almost nothing. Maybe some student debt, some rent, some food, a little travel. You are not paying a mortgage. You are not saving for a kid's tuition. That stuff comes in your mid to late thirties, and when it does your risk tolerance collapses whether you like it or not. Right now you hold the rarest asset there is: the ability to take a swing that doesn't pay in money. Spend it.

So what do you actually weight? Elad Gil wrote the cleanest version of this, and his ranking matches what I've seen recruiting. Optimize for network first. Not just who you work with day to day, but the whole web of people around the company: the founders, the early employees, the investors, the advisors. In this world the same people work together again and again. Fall in with the right crowd early and you get a decade of outsized opportunities, because those relationships compound the way capital does.

Right behind network: the market and its growth rate. Early in your career the trend you're riding matters more than almost anything you personally do. Join a company in a great market and suddenly everyone wants to hire you, and the openings multiply both inside your company and across every other company in that market. Pick the wrong market and the best case is you end up at the equivalent of some telecom equipment company in the nineties, technically still employed, going nowhere. The catch is that almost nobody is actually good at spotting which wave is real ahead of time. That's exactly why it pays so well when you get it right. Don't assume you can see it clearly. Choose the wave carefully and stay humble about how hard the call is.

And weight optionality. Are you doing the same job in the same industry on repeat, or does this role open doors to something new? People badly underrate this and just keep doing the one thing they already know. The job that lets you switch markets or switch functions is usually worth more than the one that pays slightly more to keep you in place.

What about prestige, the name on the resume? It matters exactly once. If you went to a good school and then put in real work at a known company, that reputation rubs off and people assume you're good. But Elad's point is that the second and third brand do almost nothing. Once you have one or two solid names, stop optimizing for logos and start optimizing for everything else: network, market, optionality, the people. Collecting brands past that is just vanity that costs you the things that actually compound.

Listen to everyone, and trust no single camp. The people giving you career advice are almost always biased by whatever happened to work in their own vertical. VCs will tell you to bet on startups. Quants will tell you the only rational move is to maximize expected value. Big company people will tell you to go collect the brand and the stability. They're all partly right and all partly selling you their own past. Alpha decays. The thing that minted the last generation in one corner of the world is often already priced in by the time it reaches you as advice. That includes my advice in this essay. I'm a talent investor. Of course I think networks and early-stage bets are everything.

So your job is not to pick a camp and absorb its gospel. It's to collect success stories that look nothing alike and hunt for what they share. The founder who quietly built for three years, the researcher who rode a market wave, the operator who compounded relationships across a decade. Find the common threads underneath the very different surfaces. That's where the real signal is. If you overindex on one experience or one group, you miss the forest for the trees, and you end up running last decade's playbook with full conviction.

Now money. I watch brilliant people pour into quant, and if your only goal is liquidity, fine, it works. But I don't think quant maximizes for the most ambitious version of a career. Run the rough math. Five years deep in quant and you're maybe capping out somewhere under thirty million. Five years at a genuinely good startup and yes, your base is lower, your guaranteed number might be a fraction of that. But the ceiling is gone. You can push toward nine figures, and more importantly you come out of it as a different person.

The equity trap is worth spelling out, and Elad nails it: owning one percent of a mediocre company is almost always worse than owning a tenth of a percent of a great one. The mediocre company's equity can be worth literally nothing. Cap tables get wiped far more often than people expect. The great company's stake can be worth a thousand times more, and on top of that your cash comp at a great company climbs over time anyway as the company wins. So don't chase the headline number. Chase the company that's actually going to matter.

A caveat I want to be honest about, because it's easy to overstate this. Big companies are full of genuinely brilliant people, and you can learn an enormous amount inside them. I'm not telling you to avoid them on principle. But if I'm being honest about my own bias, I think a few years at a great early-stage startup beats a few years at a great big company for most young people, and it's not close. At a startup you're forced to do everything, you see the whole machine, and you build with people who will scatter across the ecosystem and pull you along with them. The real risk with big companies is monoculture, spending so long inside one institution that you absorb only its way of seeing the world and lose the ability to think any other way. A few years at a great large company, going in deliberately to learn from specific people, can absolutely be worth it. I'd just default to the startup, and treat the big company as the deliberate exception rather than the safe starting point.

If you're in school and you genuinely don't know, go find out. My honest advice is to take at least half a semester, ideally a year, and go work at a startup. Not every school allows this and not everyone can afford it, between visas and money and family. But if you possibly can, do it. The whole point is that you cannot reason your way to the answer from a dorm room. You have to go out and feel whether you like it. Exploration is the work. And if you're going to do tech, get to Silicon Valley, or New York if you can't. The networks out here are simply denser than anywhere else, and proximity is half the game.

Say you're ready to jump. Where do you go?

First, what to avoid. There's a loud, glossy version of the startup world that exists mostly on the timeline. VCs and operators throw money at young kids, the kids post and post and post, and almost none of those companies survive. Never join a startup whose main output is content. If the project is real, the work speaks. If the loudest thing about it is the posting, run.

The founders I've watched go on to raise from the best investors are usually the opposite. They're low-key. You can barely find anything about them or their company online. They're eighteen, nineteen, twenty years old, quietly building something serious, and you have to be inside the right circle to even know they exist. So if you're talented, here's the move: post your actual project. Post something meaningful, not a daily performance. Build in a way that can be found, and the right people will find you. We do look.

And think hard about the obvious prestige seat. A frontier lab or a big growth-stage rocket looks sexy, and sometimes it genuinely is the right call, especially if the people there are the best in the world at what you want to learn. But ask who you're actually serving and who you're actually building relationships with. If you join a big lab and you're still deeply engaged with peers your own age, building real bonds, even better. The seat matters less than the people around it. This is also Elad's point about role: early on, don't over-optimize for the exact title you'll hold. Get into the high-growth company and the roles will come to you as it scales.

Here is the real argument for going early, and it's about compounding relationships, not equity.

Picture this. You're twenty. You join an early-stage company that raised from a Tier 1 fund. You stay two or three years. The company does fine. Not a Cursor-level outcome, but with a strong team and good backing it's hard to truly flop. You leave at twenty-three having built something real alongside the engineers, the founders, and the investors who watched you do it.

Now their Tier 1 backer hears that one of the founding employees just left. They've seen your work for three years. They like you. When you want to start something, that relationship is already there. Compare that to spending the same three years at a quant firm or a frontier lab, one of hundreds of researchers and engineers. You might do excellent work and learn a ton. But to a Tier 1 GP you're a number on a large org chart they never saw. Why would they back you over someone they watched build from the front? Neither path is wrong. They just compound into very different networks, and you should choose with your eyes open about which one you're building.

That's the whole game in your twenties. Split your attention roughly fifty-fifty between two things: optimize for network and optimize for impact. Work with the best people you can find and do work that actually matters. Stop optimizing for money. The money arrives on its own when you do good things alongside good people. It always has.