Max Levchin's Leadership Secrets: Building Affirm to $18.7B

The PayPal co-founder reveals his unconventional approach to managing brilliant people, calculated risks, and scaling culture at Affirm.

Key Takeaways

• Brilliant people often have extreme personalities, but they're worth the quirks if you manage them right
• Trust is the ultimate red line - once broken with integrity issues, there's no coming back
• You can't only hire A-players at scale, and that's actually fine if you understand the dynamics
• Failed experiments teach more than successes because analyzing wins feels boring and surface-level
• Direct feedback works when you separate the person from their work and maintain genuine humility
• Calculated risks require actual calculating - too many "strategic" decisions skip the math entirely
• Writing culture beats meeting culture for preserving institutional knowledge and making better decisions
• Being present during layoffs as a leader is painful but necessary for maintaining team trust
• Public company leadership gives you access to brilliant investors but adds regulatory complexity you never expected
• Speed matters enormously in startups, but knowing when to slow down for quality separates great companies from good ones

Managing Extreme Personalities Without Losing Your Mind

Here's the thing about brilliant people - they're often a pain in the ass to work with. Max Levchin doesn't sugarcoat this reality. "Brilliant people have extreme personalities more often than not," he says, and as a leader, you're constantly asking yourself "how extreme is extreme and what will it cost me if this person blows up?"

But here's what most managers get wrong. They either avoid hiring these challenging personalities entirely, or they hire them and then spend all their energy trying to tone them down. Levchin takes a different approach. He believes "most of the time truly brilliant people are worth the quirks," and the secret is learning how to channel that intensity productively.

• The pinnacle of teamwork isn't avoiding conflict - it's having people with "really strong opinions" who don't hold back but also don't destroy relationships in the process
• You want someone who can walk into a room and say "that is a pile of garbage and I love all of you, you're amazing people, but what we've built here is not going to ship because it's terrible"
• The key is separating judgment of work from judgment of the person - "this isn't your best work but I know what you can do" hits differently than "you screwed up"
• When someone consistently can't do better despite circumstances, that's when you have the harder conversation about team fit

The moment extreme personalities become toxic is when narcissism creeps in. "The second it becomes about you being better than everyone else, everyone else will start hating you," Levchin warns. If you're the boss, "they're going to hate you quietly, and that's the worst possible situation."

What's interesting is how he handles the irredeemable cases. While he believes in second chances for most behavioral issues, there's one absolute red line: "Once I can't trust you, I can't ever trust you." Integrity violations don't get do-overs in his world.

The A-Player Myth That's Destroying Your Hiring

Every startup founder thinks they're going to hire only A-players. Levchin calls this "absolutely absurd." Here's why that thinking is flawed and what actually works when you're scaling past the first hundred people.

"A-players by definition are exceptions," he points out. When you're the size of Affirm with thousands of employees, "it's impossible to only have A-players." But here's the part most people miss - that's actually fine if you understand the dynamics.

• Modern interviewing is broken because candidates are "coached and trained and fine-tuned to be exceptional interviewees" who've studied exactly what you want to hear
• You can filter out C and D players pretty effectively, but "differentiate between B and A is what happens once you brought them in"
• The real question becomes "what do you do to get the absolute best work from this person" regardless of their letter grade
• Some B-players "grind their way to being A-players" over time, making them excellent investments

The famous Steve Jobs quote about A-players hiring A-players while B-players hire C-players is true, but Levchin adds crucial context about why. It all comes down to fear. A-players "love working with other A-players because they know the demands of the job and the team dynamics just make them better." They want to be challenged.

B-players split into two categories. The good ones say "I'm a solid B and I know I'm not an A-player and that's okay, I am just going to grind out good code." These people are valuable because they'll tackle work that A-players find unchallenging.

The dangerous B-players are the ones driven by fear: "If I have too many B-players around, let alone A-players, I will look terrible, so the way to fix this is for me to become a manager and hire a bunch of people who aren't very good at all." These are the people who actively make your organization worse.

Why Most "Calculated" Risks Aren't Actually Calculated

Levchin's advice to "take calculated risks, do the calculating" sounds obvious until you realize how rarely people actually do the math. He shares a perfect example of when Affirm didn't follow this principle and paid the price.

Years ago, they expected a massive product launch and decided to scale their infrastructure preemptively. "We were prepping at breakneck pace and just scaled up every part of our backend infrastructure and then our frontend infrastructure." The launch day arrived and "it just fell apart massively because we basically overloaded our backend with our frontend."

• Someone should have calculated that scaling the front tier faster than the back tier would create "an over-subscription of frontend connections to the backend"
• The backend would start choking because "even if frontend doesn't do anything, just holding these connections is expensive"
• With a 10x mismatch, "you're going to bring down the database, and that is what happened"
• This was completely foreseeable if someone had done the actual calculating instead of assuming bigger is always better

Interestingly, Levchin's founding decision for Affirm was actually uncalculated. When he told bankers he wanted to "eliminate fine print so consumers do not have to read fine print to find out what the business model is," several well-known people "basically laughed at me and said you're insane."

His response? "I don't think it's really worth running a company that fits neatly into the mold of 'screw the customer, make some money.' I want to run the one that says 'never screw the customer, let's see if we can make some money.'" Sometimes the biggest bets are the ones you can't calculate but feel compelled to make anyway.

Building a Postmortem Culture That Actually Works

Most companies say they learn from failures, but few do it systematically. Levchin has built what he calls a "postmortem culture" at Affirm, and it's more structured than you might expect.

"For everything, create a document space," he explains. Every time there's a system blip, process failure, or model breakdown, "a separate conversation is carved off on Slack, a separate conversation is documented in a Google Doc." The key is segregation - mixing data from different incidents makes analysis impossible.

• You want "as much data as possible segregated into a separate pipeline of raw content as soon as possible"
• Someone needs to be specifically responsible for "writing it up and distilling it from pages and pages of raw data into essentially a white paper"
• The entire team comments on it, but you have to "completely separate any form of apologizing, responsibility taking" from the analysis
• People who screwed up need to "feel very free to comment on what happened" without worrying about making themselves look better

The timeline matters. "It's a week, maybe two weeks - that's the most you're allowed to take to generate a real postmortem." Then there's a weekly review where senior people evaluate whether the analysis is deep enough or if you're just describing what happened without explaining why.

Here's what most people get wrong about learning from failure versus success. "Analyzing success is boring," Levchin admits. "Well, we did well, it's good, I'm tired, I'm gonna go to sleep." When everything goes right, "there's a real kind of blur between the first five steps" in your funnel analysis.

Even with successful initiatives, he pushes teams to ask "could it have been 96% instead of 95%? Why did we do better?" You're still asking about failure - just potential failure rather than actual failure. "The question has to be like 'what should we have done better' even if it's the most successful thing ever."

The Remote Work Reality Check Nobody Wants to Hear

Levchin's take on remote work isn't what you'd expect from a tech CEO. He's neither fully pro-remote nor demanding everyone return to the office. Instead, he's focused on what actually drives productivity and team effectiveness.

"I care a lot more about output and productivity than I care about where you are physically," he says, but there's a crucial caveat: "There has to be enough together time." The idea that you can be fully remote and never meet your teammates in person is, in his words, "a stupid argument."

• Some people are genuinely more productive when they can "cut out all distractions and hide themselves in their spare bedroom and just write code for 12 hours"
• But "you form relationships, you understand each other's decision-making style, you break bread together, form camaraderie - that has to happen"
• If your team is global and most meetings are on Zoom anyway, it's hard to justify forcing people into offices just to sit in video calls
• The key is being "completely non-compromising about" having sufficient in-person interaction, not arbitrary office requirements

What's particularly smart about his approach is recognizing that different geographies have different needs. Their New York office "will have standing room only" while San Francisco might be "half full, maybe a little bit more." Rather than imposing a one-size-fits-all policy, they adapt to what works for each location and team.

Public Company Leadership: The Good, Bad, and Regulatory

Going public changes everything about being a CEO, and Levchin is refreshingly honest about both the benefits and the bureaucratic nightmare. "Some days yes, other days no," he says about enjoying public company leadership.

The good days are when being public "gives you access to more capital" and "exposes you to really smart investors." There are brilliant people who "will only invest in public equities" and "they are brilliant people who've seen it all before." As a public CEO, "they might own your stock and then they want to talk to you because they want to assess you as much as they want to help you."

The bad days? "I now have to go read this filing that we have to make because it's part of my responsibility to make sure I understand everything we're putting out there in the public." These regulatory requirements "were not what you thought you'd be getting into when you were getting a computer science degree."

• The temptation is to tie your mental health to market cap, especially "when you think you're doing a fantastic job" but the market disagrees
• "You're doing great, you know you're doing great, and the market thinks you're not" creates a painful temporal mismatch
• His solution is practical: he's configured his desktop so he "cannot see my stock ticker" to avoid accidentally discovering if it's "a red day or a green day"
• The mantra is Warren Buffett's mentor's wisdom: "In the short term the market is a voting machine, in the long term it's a weighing machine"

When it comes to employee morale during stock price fluctuations, Levchin takes a different approach than most CEOs. He rarely talks about stock price with employees, and when he does, it's to set realistic expectations rather than celebrate highs.

The Layoff Lessons Every Leader Needs to Know

"Morale was lowest" during Affirm's layoff about two years ago, and Levchin learned crucial lessons about leadership during crisis. The first time he had to do layoffs at a previous company, he was "terrified of owning the responsibility that I screwed up" and just wanted to "run and hide."

A good friend who was an executive pulled him aside with advice that changed his approach: "This sucks for all of us, you are the leader. You have to be out there helping people pack their boxes. You can play this from the comfort of your office or in the middle of the floor that's crying - go be with the people."

• Being present during layoffs is "between cathartic and therapeutic" for both the leader and the affected employees
• "You cannot hide from the grief that a layoff is" - trying to avoid the emotional impact makes it worse for everyone
• If you've built a great culture, "the blow is much softer than you think it'll be because people understand that you tried with every possible strategic or tactical idea"
• People are "more understanding and less hard on you than you are on yourself" when they trust your leadership

The Affirm layoff taught him something remarkable about company culture. "Every single person I talked to had said 'hey I get it, this happened, we grew too fast, we over-hired, I understand, I even understand why me and not the next person.'" Many even said they hoped for an opportunity to return because they loved the company.

Over-hiring during good fundraising periods is itself a calculated risk that sometimes fails. "You decide that you want every incremental engineer to build more features while the fundraising is good, and then suddenly it's not, and what do you do?" The analysis involves "how fast something will grow, how well you can have the market protect your cash flow if it doesn't, how profitable something will be."

Why Writing Culture Beats Meeting Culture Every Time

Most companies default to meetings for everything. Levchin has built something different at Affirm: a writing culture that "favors short, pithy one-pagers to novels or live rants." While he admits to "being born with a microphone in my hand," he's learned that documentation creates lasting value in ways verbal communication can't.

"If you have a depository - and we literally have a postmortem depot where people can go and review every major or minor thing we did going back to the beginning of Affirm's time - it's really valuable," he explains. You end up with "a library of case studies that are internal to the company" where people can learn from past decisions and mistakes.

• Written records preserve institutional knowledge that would otherwise become "oral tradition" and get lost
• "The most important thing about technical writing and business writing is simplicity" - if you're reading 12 pages about a feature, "you're probably going to miss details"
• For a feature that fits in "25 by 25 pixels, the description should be two pages, maybe three" rather than extensive prose about world impact
• Live rants work for motivation and aligning teams around strategy, but people typically "remember two" things out of the "12 things" you say

The practical reality is that people have limited attention spans. "We have distractions, notifications are popping up" and when faced with a 12-page document, most people think "oh man" and struggle to focus. "Majority of people are less focused than they'd like to be and less focused than they think they are."

This philosophy extends to his shareholder letters, which have an unusual quirk - every single one contains a reference to "The Big Lebowski." When asked about this, Levchin reveals it stems from discovering that many of his management team were fans of the film during interviews, and they made it their business to work quotes into official communications.

The Competitive Analysis Trap That Kills Focus

Here's something most CEOs won't admit: they spend way too much time obsessing over competitors. Levchin takes a contrarian approach that's worth understanding, especially as Affirm competes directly with companies like Klarna in the buy-now-pay-later space.

When asked about Klarna's growth and U.S. expansion strategy, Levchin deflects: "That's a postmortem they will have to write for themselves. I'm not sure I'm qualified, I don't have all the information." More importantly, he believes "it's very healthy for CEOs of startups to spend very very little time on other people's strategy."

• Competitive analysis is valuable "at the product level and sales strategy level" - understanding what competitors do that you could do better
• But asking "what is their strategy" is like asking "what weather will be 12 weeks from now" - largely unknowable and not actionable
• "We tear down all competitor products, all their sales efforts, everything we can get our hands on" because mistakes are "free" learning opportunities
• The focus should be on your own execution rather than trying to reverse-engineer competitor decision-making

This mindset has served Affirm well. They've been "taking share in the market against all competitors" and growing at 35% while offering "more services" than most competitors who "specialize in one very specific" area. Their decision "from day one to provide every possible service from the 45-day super short-term loan all the way out to four years" positions them to "take disproportionate share."

The Speed vs. Quality Balance That Actually Matters

Every startup talks about moving fast, but Levchin has a nuanced view on when speed matters and when quality can't be compromised. "Speed is one of the most important" determinants of startup success, but knowing when to slow down separates great companies from good ones.

He references Steve Jobs' practice of signing the inside of computer motherboards: "If you build something beautiful and amazing, you don't mind - in fact you take pride in signing the insides because if anyone ever looks, they will be blown away by just how beautifully the whole thing is designed."

• There's "natural and very healthy tension between shipping fast and building things beautifully"
• Building beautifully "requires design and time and consideration of what the next version would look like so you don't have to rebuild from scratch"
• Every startup that's been around for a decade has "someone undoubtedly who's refactoring some part of your code because it's grown enough moss that you need to deal with it"
• For crucial systems like "machine learning models that deal with credit underwriting," you "cannot let them be anything but the absolute best you can possibly produce"

The practical application means making exceptions for delays when they involve foundational systems. "When somebody tells me we're doing X and it's going to be a month late, my natural reaction is to jump up from my chair like 'how can it possibly, how dare you be so late.'" But for systems that will be "crucial for months and years to come," delays become acceptable if they ensure quality.

Affirm has a "prototyping team that is trusted/empowered to ship stuff that is just not really beautiful or production quality" to get customer feedback quickly. But here's the key: "Once you get data and you think the data tells you that this is great, you don't just say 'cool, let's leave it out there.' We basically pull down the feature and say 'okay, now we're going to do it right.'"

The AI Story Every Public CEO Needs (Or Doesn't)

As a public company CEO, Levchin gets constant pressure to have an "AI story" for investors. His response cuts through the hype: "I generally really dislike stories that have no substance." Instead of crafting narratives for the market, he focuses on a simple question: "Are we taking full advantage of the best available tools?"

"The most interesting available tools right now are AI," he acknowledges, but warns against backwards thinking: "If you sort of look through your closet and say 'we're not really using this stuff, I got to have a story for the market,' you're doing two things wrong - you're telling a story that isn't true and you're not using the most interesting available tools."

• If you're actually using the best tools effectively, "your investors might be happy if you told the story, but if you didn't, I don't think they should really care that much"
• Public company success comes down to "are you growing at the right scale, are you taking appropriate risks, are you managing to profitability"
• The "storytelling hour" of highlighting shiny new features "feeds the voting machine" but he spends "a lot more time feeding the weighing machine"
• On LLM efficiency limits, he agrees "we're getting to the upper limit" but believes "we're not going to stay with LLMs forever"

His prediction for AI evolution involves "multiple competing LLMs basically debating each other for the right answer" and other iterative processes beyond pure language models. The key insight: "You can amplify intelligence of humans by having three smart people in a room instead of just one person."

Levchin's consistent theme emerges: substance over narrative, execution over explanation, and long-term fundamentals over short-term market positioning. Whether it's AI adoption, competitive strategy, or public company communications, the focus remains on doing the work rather than talking about the work.

The conversation reveals someone who's learned that scaling a company to $18.7 billion requires constant balancing - between speed and quality, individual brilliance and team harmony, calculated risks and intuitive leaps. Most importantly, it requires the kind of institutional honesty that comes from documenting failures as carefully as successes, because that's where the real learning happens.