Sign Your Name to It

Peter Harris describes his job as getting coffee for the interns. It is a joke, and like most good jokes it is carrying a serious argument. Harris is a founding partner of University Growth Fund, a firm managing roughly USD $100,000,000 across three vehicles, with positions in Lyft, Spotify, Toast, Carta, OpenAI, Anthropic, and Databricks. The interns he claims to be fetching coffee for are the ones running it.

Two of the funds are early-stage venture and growth equity. The third, newer one is a private credit vehicle that will also do private equity investing, concentrated in business services, home services, and healthcare services. The firm co-invests alongside a wide range of other funds. None of that is what makes it unusual. What makes it unusual is that the analytical work, the founder meetings, the models, the memos, and the votes belong to students.

At any given moment the program runs somewhere between fifty and sixty students, drawn from schools around three cities where the firm keeps a physical presence: Salt Lake City, San Diego, and Atlanta. The offices exist so students can congregate and build relationships with each other while they work deals, which Harris treats as a load-bearing part of the design rather than a perk.

Asked whether he takes his hands off the steering wheel, Harris reached for the obvious comparison and then refined it. It is like learning to drive next to an instructor who has a brake pedal on the passenger side. The brake pedal is still there. There is adult supervision. But the student is driving. The day before the conversation, a group of them had walked the manufacturing floor of a major defense company the fund was evaluating, questioning a CEO who has raised tens of millions of dollars.

The tenure is deliberate. Most students join the summer after their sophomore year, though the firm recruits three times annually. Those who fit are promoted and typically stay eighteen months to two years. They spend their junior year on deals, go do another internship that the firm often helps arrange, and ideally land a full-time offer by the end of it. Their senior year, with the job already secured, they return as senior associates leading deal teams and meeting founders and investors directly, functioning something close to a partner. Harris is candid that the payoff shows up less in the first job out of college than in the second and third.

Ask Harris about artificial intelligence and the fund's internal tension surfaces immediately. He wants his students to be AI native, because he believes they have to be in order to succeed. He also believes the opposite thing at the same time, and considers both true: real-world experience is precisely what makes a person capable of leveraging AI well.

His own example was domestic and unglamorous. He spent an hour and a half one evening building a financial model to decide whether to add solar capacity and a battery to his house. The first cut was decent. It was also riddled with mistakes. He caught them only because he knows how to do finance, which meant he knew where to push back, where to correct, and where the answer was still wrong.

So the instruction to students has stayed the same as it was before any of these tools existed. You need reps. You need to own what goes down on the paper. You need to understand it. AI is there to accelerate, complement, and teach, not to absorb the work. The best students take that seriously and improve at remarkable speed. With the others, the firm spends some time helping them reassess their priorities.

Whitwell recognized the standard from an earlier era of the same problem. As a young analyst at a boutique M&A shop, he built models from scratch because there were no preset templates. It was punishing, and it taught more than any template could. The practice he remembers most, and which he thinks firms should revive, was that partners made analysts sign their names to the output. You put your name on the printed page. You own every number on it.

The reason that mattered was the meeting. A one-or-two-years-out analyst sits in a room with a partner and a Fortune 50 CEO, invisible to both, until the client puts a finger on a number and says that cannot be right. Every head turns. If you did the work, you can explain why the number looks counterintuitive and what is actually happening beneath it, and then you get to be happily invisible again. If you did not, there is nowhere to go.

The signature has more force now, not less. When the output is some blend of a person and a tool, ownership is the only thing that resolves cleanly. The formality of putting your name on something is what makes you stop and check it a third time. And after the fact there is no version of the excuse where the tool made the error and the person did not, because the person signed.

Whitwell has long described money as an amplifier. It brings out more of whoever a person already is, for better or worse. Give it to a Mother Teresa and more good gets done. Give it to someone whose horizon ends at the next Ferrari, and, with no disrespect to a beautiful car, that is roughly what you should expect. His question to Harris was whether AI works the same way, since serious effort seems to compound into serious output while lazy prompting produces exactly what it deserves.

Harris took the frame and extended it in both directions. A friend had recently posted a chart on LinkedIn asking whether money creates happiness, showing the effect rising until somewhere around USD $500,000. His reading was the same as Whitwell's. Money amplifies whoever you are. An unhappy person becomes more unhappy. A happy person gets more of whatever it is that makes them happy.

The version he presses on students is sharper, because their livelihoods depend on it. If you outsource all of your critical thinking and all of the hard work to a machine, what exactly is an employer hiring you for? They will not pay for it. They will buy more tokens instead, because you are not adding value. The amplifier only has something to multiply if you brought something.

What Whitwell has found is the inversion that most people miss. In the last year the tools have forced him to think more rather than less, in two directions at once. Creatively, because projects that were never worth starting are now trivially startable, and he has repeatedly found himself asking why he did not do something three months earlier. And critically, because the output is not reliable and he remains responsible for the quality of the analysis. Where might there be an error? Where is the weak link? What is missing? He has more room for those questions precisely because the mechanical work compressed.

Harris brought the same lens to the investing side, where the picture is less comfortable. Every time a major lab ships something new, he goes through what he cheerfully calls a small crisis, because a portion of what he thought he knew stops being true.

The clearest illustration came at a demo day a few months back. Every company presenting was performing: real customers, real revenue, real problems being solved. Talking with the other venture investors afterward, they reached a shared and slightly uneasy conclusion. Four years ago they would have fought tooth and nail to get into every one of those deals. That day, none of them was confident any of the companies would still have a business in five years, and even less confident it would be a venture-backable one.

The mechanism is the power law, which has always concentrated venture returns in a handful of names and is now concentrating them harder. The examples are unmissable: the scale of OpenAI, Anthropic, and SpaceX, and the volume of capital that has flowed into them. What that does to a fund manager is raise the bar on a specific question. Can this company still have five to ten years of runway ahead of it in five years' time, since the exit depends on selling it to someone who believes it has a future?

University Growth Fund's answer has been to move to both ends of the barbell and abandon the middle. At one end are companies with a genuine moat because they are foundational or infrastructural, which is the reasoning behind the OpenAI, Anthropic, and Databricks positions. At the other end are companies whose protection comes from outside the technology entirely, where AI can only help.

Healthcare is the cleanest example of the second category. It is heavily regulated, dense with licensure and statute, and hard to enter. Those external walls keep competitors out while the technology is deployed internally to improve margins and operations. The fund's investment in Nest Health illustrates the shape: care coordinators go physically into the homes of Medicaid families and assess the children, raising quality of care while lowering system cost. Nothing is replacing the person who has to see the environment and examine the child. But the analytics, the tracking, and the efficiency around that person can improve considerably.

Asked for an example from the vulnerable middle, Harris described a company without naming it. A solo founder had built it from zero to USD $1,000,000, hired a few people, and taken it to USD $9,000,000, all inside roughly eighteen to twenty months. Ferocious growth, in education, a sector the fund cares about. What he could not convince himself of was that fifty other competitors would not spin up a workable version and start competing.

Harris does not think AI destroys software businesses, and he is precise about why. The so-called SaaSpocalypse applies to venture-backed SaaS specifically. What he expects overall is a massive expansion of SaaS, because the reason those businesses needed venture capital has largely evaporated. You used to raise because you had to buy developers to build the code and sales reps to acquire the customers. That requirement is gone.

What replaces it is a large population of owner-operators running businesses at single-digit millions in revenue with respectable economics, perhaps twenty percent EBITDA margins, throwing off something on the order of USD $2,000,000 in cash. Harris was unambiguous that these are wonderful businesses to own and that he would happily own them. He also cannot invest in them, because they will never return a hundred times capital. That is the paradox he keeps turning over: the same shift that makes him a more creative and more productive person, building apps for himself and others and helping his mother modernize her law practice, is the one narrowing the aperture of his day job.

The reason the model works is not charity. It is a flywheel, and Harris can trace every revolution of it.

It starts with scarcity. Experience in venture capital is genuinely hard to come by, and what the fund offers is unusually rich. Students routinely come back and report that their later internships and even their first full-time jobs felt like a demotion, from a rung near the top of the ladder back down to the bottom, where nobody cares what they think.

That experience converts into outcomes. Students leave earning roughly one hundred twenty percent more than their peers at graduation, into jobs many of them would not have had access to otherwise. Because that impact is real and measurable, Harris can walk into a founder meeting and describe it without exaggeration, which turns out to matter. Founders, in his experience, are by and large people who want to help and want to make a positive dent. Asked to support students by accepting the fund's capital, they tend to say yes.

That access produces performance. The funds have made about a hundred investments. Roughly a fifth have gone public, a number more have been acquired, and Harris is careful to note it is still early innings. Investors get returns, which makes them willing to commit to the next fund, which funds more students.

Then the loop closes in the way that makes the whole thing compound. Alumni are now working at excellent funds, and they send deal flow back. Lyft, OpenAI, and Databricks all reached University Growth Fund because former students wanted the current students to have the experience they had, and made sure the fund got a look.

Harris draws a general principle out of it, and it is the least sentimental version of the idea. Business frequently defaults to an eye-for-an-eye posture, a mentality in which someone has to lose for someone else to win. What he has learned is that in most situations, though not all, sufficient creativity produces an arrangement where everyone comes out better and everyone wants to play again. Repetition is where the returns actually live, and not only the financial ones. Meaningful friendships, experiences, and accomplishments are built the same way.

Whitwell's observation was that there is a stubborn perception that doing good has to cost you something. Giving purely for its own sake is good and needs no defense. But there is a distinct kind of energy in a structure that does both, that exercises real capability and remains economically sustainable, because sustainability is what keeps the giving going. For capitalism to work in its best expression, the profit is not incidental to the impact. It is the fuel supply.

The most quietly radical thing Harris said had nothing to do with technology. When he joined the fund he took a substantial pay cut, to less than what the students he was training would earn at graduation. He was happy to do it, on the reasoning that life is short and he would rather spend as much of it as possible doing work he loves, for as long as he can, rather than deferring enjoyment until he is old.

Which is the argument against the most common sequencing error in the whole subject. A great many people plan to give back when they retire, or after the exit, or once the number is finally large enough. Harris did not wait, and neither did his students, who are making an impact as sophomores. The permission structure people are waiting for does not exist.

Pressed on what money has actually unlocked for him, Harris's answer was specific rather than abstract. He loves to travel with his family, and to give his children new experiences and exposure to other cultures. He has done substantial international work, particularly with the International Rescue Committee. For his fortieth birthday he threw a large Lego party for refugees in Salt Lake City through a local refugee organization.

The other thing money and technology bought him was the ability to make things he lacks the skill to make. He is not an artist and says so plainly, but he has always had ideas. Now he uses the tools to render what he pictures, most often as elaborately custom birthday cards for his family, which he admits he spends a decent amount of time getting exactly right. Whitwell recognized the pleasure in it, remembering an early graphics program on an Apple IIe that let you design a card, print it, and fold it, with a few hundred graphics to choose from. The constraint then was the library. The constraint now is the imagination.

Which returns the chapter to where it started, with a man who says his job is to get the coffee. Money is an amplifier. So is the technology. Neither of them originates anything. They take whatever a person actually brought, judgment, effort, generosity, or the absence of all three, and make more of it. The only durable question is what you are handing the amplifier, and whether you are still willing to sign your name to it.