Apr 14, 2025 · 1h 12m · 20vc
Benchmark GP, Victor Lazarte: The 3 Traits All the Best Founders Have · 20VC with Harry Stebbings
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of 20VC, host Harry Stebbings interviews Benchmark General Partner Victor Lazarte about his transition from bootstrapping in Brazil to venture capital investing, exploring his unique philosophies on AI labor replacement, the evolution of relationship-driven board leadership, and the critical founder traits required to build the next generation of trillion-dollar tech giants.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Harry holds 19.2% of the talking time here. How this is scored →
Harry as informed peer 3.3 Guest teaching 3.4 Guest disagreement 1.3 Harry pushing back 2.3
speaking balance: gold is Harry, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 39:40 Victor calls AI augmentation 'bullshit'
Victor aggressively rejects standard corporate PR narratives that AI augments workers, calling it bullshit and insisting it fully replaces labor.
Hardest push from Harry ▶ 53:00 Harry rejects VC 'first call' clichéHarry directly refutes Victor's standard VC line about being the founder's first call, arguing your mom should be your first call and asking for real honesty.
Biggest teaching moment ▶ 32:23 Victor explains RL evals and Mercor defensibilityVictor educates Harry on the evolution of LLMs from pretraining to pure RL with human evals, demonstrating why Mercor provides indispensable infrastructure for AI labs.
Harry holds his own ▶ 9:34 Harry drills on mental health crisis dataHarry cites stark public health statistics about young male suicide rates versus cancer to challenge Victor's tech optimism and question why mental health solutions remain small.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Harry as informed peer | Guest teaching | Guest disagreement | Harry pushing back | Why |
|---|---|---|---|---|---|---|
| Bootstrapping in Brazil: Victor's Origin Story | 4 | 2 | 1 | 2 | Harry demonstrates extensive preparation by citing specific conversations with Benchmark partners and founders, then pushes Victor on why he failed to raise capital initially. Victor clarifies that bootstrapping was due to low VC sophistication in Brazil rather than a strategic choice. | |
| Market Timing and the Monumental Rise of AI | 2 | 4 | 0 | 1 | Victor educates Harry on market timing heuristics, citing Bezos's 23x internet growth insight and drawing parallels to 100x LLM flop growth. Harry listens attentively and prompts Victor to expand on market adjacencies. | |
| The Rise of AI Companions and Agent Interfaces | 2 | 4 | 1 | 1 | Victor outlines how agent interfaces will replace traditional UI shifts, comparing Character.ai to Friendster. Harry listens as Victor details how businesses will adapt to conversational agent models. | |
| The Rise of AI Companions and Agent Interfaces | 3 | 3 | 1 | 3 | Harry interjects with stark mental health statistics about young men and questions why existing mental health apps like Calm have failed to reach massive scale. Victor agrees and builds on the premise. | |
| Redefining Human Happiness and AI Relationships | 3 | 5 | 2 | 3 | Victor makes a provocative argument that AI friends can fulfill relationship needs similarly to religious beliefs and God. Harry offers a sharp counter-thesis that CPG and consumer brands like Lululemon have replaced God, but Victor maintains his focus on deep AI-driven personal narratives. | |
| Unlocking Extreme Trajectories: Brex, Mercor, and Founder Obsessions | 3 | 4 | 1 | 1 | Victor shares details of his early friendship with Brex founder Pedro and draws parallels to Mercor's founder Brandon. He details Yuri Milner's heuristic on how founders spend free time as an indicator of obsessive capability. | |
| The Fiduciary Duty of Board Members and the Power of Context | 3 | 4 | 1 | 2 | Harry questions whether younger founders require a different style of coaching and board involvement. Victor explains the network effects of West Coast knowledge sharing and defines a board member's primary value as deep context and hypergrowth experience. | |
| Active Outbound: Sourcing the Best Venture Deals | 4 | 3 | 2 | 3 | Harry pushes back on the common myth that top VC firms like Benchmark receive all their best deals inbound. Victor agrees, confirming that 90% of his sourced deals are active outbound hunting. | |
| AI Revenue Quality, Defensibility, and the 3 Stages of LLMs | 5 | 6 | 2 | 4 | Harry challenges the durability of AI startup revenue, asking whether rules of scaling SaaS have broken and if Mercor's revenue counts as ARR. Victor provides a detailed breakdown of the 3 stages of LLM training (pretraining, RLHF, pure RL/evals) to prove Mercor's underlying defensibility. | |
| Labor Disruption and the $10 Trillion AI Opportunity | 4 | 4 | 2 | 3 | Harry questions if investors are getting paid for the risk they take at sky-high AI valuations. Victor argues that capturing market share in the $10T labor market justifies high entry prices for eventual trillion-dollar winners. | |
| Destabilization, Wealth Redistribution, and the China Factor | 5 | 5 | 4 | 5 | Victor bluntly dismisses tech PR lines that AI 'augments' workers as 'bullshit', asserting AI fully replaces people. Harry pushes back on UBI, arguing humans require purpose, while Victor counters that wealth concentration will trigger populist redistribution unless China acts as a unifying external threat. | |
| Rapid AI Adoption and the Future of Trust | 2 | 4 | 1 | 1 | Victor predicts people will soon blindly trust and follow daily commands from AI companion apps. Harry asks for timelines and practical examples, which Victor illustrates through fitness and accountability habits. | |
| Benchmark's Craft: Deep Context and Strategic Partnerships | 5 | 3 | 3 | 4 | Harry directly challenges the generic VC cliché of wanting to be a founder's 'first call'. Victor clarifies Benchmark's distinct model of high-conviction, low-volume investing to maintain intense contextual depth. | |
| The Modern Board Role: From Governance to Amplifying Ambition | 4 | 4 | 2 | 3 | Harry points out the legal tension between a board member's fiduciary duty to shareholders versus maximizing founder ambition. Victor explains that VC underwriting happens at check-writing, after which board work focuses entirely on founder enablement. | |
| The Decision-Making and Voting Process at Benchmark | 4 | 3 | 1 | 3 | Victor explains Benchmark's 1-10 voting system without 5s to force clarity, emphasizing that individual GPs retain final check-writing autonomy regardless of vote totals. Harry probes on how Peter Fenton briefed him to ask about preventing firm extinction. | |
| Fund Size Constraints and the HeyGen Investment | 4 | 2 | 1 | 2 | Victor reveals his first check at Benchmark was a $55M check into HeyGen from a $600M fund. Harry highlights the high concentration risk (over 10% of fund), while Victor maintains fund size is not a constraint for inception rounds. | |
| The Relationship-Driven Approach: Sierra and Brett Taylor | 2 | 3 | 0 | 1 | Victor illustrates how world-class founders like Bret Taylor choose Benchmark for strategic partnership over pure capital, having dinner with partners rather than pitching. | |
| Quick Fire Round: One Stock to Hold for Ten Years | 2 | 2 | 0 | 1 | In the quick-fire round, Victor picks Duolingo as his top 10-year stock holding due to its potential to build AI tutors. He also praises Green Oaks and Conviction as top funds. | |
| Concluding Thoughts and Farewell | 1 | 0 | 0 | 0 | Harry thanks Victor for a conversational and free-flowing interview, and Victor expresses his gratitude. |
Statements from this episode (44)
Lazarte: Claims that AI only augments workers are "bullshit"
“Big companies still talking about like, oh no, AI, AI is not replacing people. AI is actually augmenting people's abilities. Like this is bullshit. You know, like it's fully replacing people.”
Lazarte: A Trillion-Dollar Company Will Be Founded Within Three Years
“What I'm telling you is in the next three years, I'm sure someone will start a company that's going to be worth a trillion dollars.”
Lazarte: Wildlife Studios bootstrapped out of necessity, not choice
“So, so yeah, I bootstrapped a mobile gaming business, and it makes for a good story, but the reality is we tried really hard to raise money. Like, we didn't bootstrap by choice.”
Lazarte: Best early Wildlife Studios term sheet offered $50k for 50%
“The idea was, like, I loved mobile games, and I thought it was going to be a good business, but there wasn't, like, a very strong, well-thought-out thesis, and I guess that's why, like, the best term sheet that we got is someone offered us, like, 50,000 dollar…”
Lazarte: Early Brazilian VC deals included predatory 20x liquidation preferences
“So I had friends that they raised money in Brazil and they agreed to like a 20 times liquidation preference, which seems like oh, yeah, there's this clause in the contract, but I don't really know what it is. And yeah, like, people do that, and they end up los…”
Lazarte: Mobile app downloads grew 42x between 2008 and 2009
“And we saw that app downloads, like, this was in the market, and app downloads had grown 42 times. From 29, from 2008 to 2009, happy downloads, mobile app downloads had grown, like, 42 times.”
Lazarte: LLM usage measured in FLOPS grew 100x over two years
“LLM usage has grown a hundred x in the last two years. So if you measure like in the number of flops, it went from like one extra flop to a hundred extra flops now. And like, of course, this data is not very precise, but it's directionally correct.”
Lazarte: Now is the best time in over a decade to start a startup
“It's quite clear. It's the best time in over a decade to start a company. Because we have this insane phenomenon of adoption, like adoption of LLM inference.”
Lazarte: Character AI is merely the Friendster of AI companions
“In a few years, we're going to look back, and we're going to say that Character AI was Friendster. And what I mean by that is there's going to be a company that is going to be very important, and we'll look back and we'll say Character AI was the precursor of …”
Lazarte: AI agent UI shifts allow startups to rebuild incumbent businesses
“Whenever you have a shift, like, a UI shift of this magnitude, a lot of the businesses, like, they don't adapt, and so I think there's an opportunity to recreate a lot of the successful business of today in an agent format.”
Stebbings: Calm is the largest mental health company at a $1.5B valuation
“Really the largest mental health company is, is what? Calm? Respectfully, it's not a huge business. It's, you know, one and a half billion. Fantastic journey, but it's not a huge business.”
Stebbings: Consumer brands succeed by filling the void left by traditional religion
“I say, actually, that the death of religion is the thesis for CPG and consumer brand investors today. And you're like, what? Thesis? Because we've shifted cult-like worshipping from God, Buddha, and these deus figures to High Rocks and CrossFit and Taylor Swif…”
Lazarte: AI will understand most people better than anyone within 5 years
“I have this very strong belief that five years from now, for the majority of people, the person that will understand you the most is going to be an AI.”
Lazarte: I would not have invested in Mercor without Brex
“Frankly, I don't think I would have invested in Mercor if I hadn't invested in Brex.”
Lazarte: Brex co-founder Pedro Franceschi discovered early iPhone jailbreak
“That, like, back in the day, he was the first person to find one of the jail breaks for the iPhone, right?”
Lazarte: Mercor grew from $1M to $100M revenue in 11 months
“Brex went from like zero to a hundred million dollars in revenue in like in 18 months after launch, and Mercor went from like a million dollars in revenue when I invested to like over a hundred million, like in 11 months.”
Lazarte: Mercor founder Brandon is generating $100M ARR at age 21
“And then in, in the Brandon case, it's like, okay, he's a 21 year old, running a company, doing a hundred million dollars a year. Like, there's very few people that have been in that place, but turns out that Pedro has been in that place, you know, like, early…”
Lazarte: Board value depends on past hypergrowth exposure and context depth
“I think the way a board member adds, adds value is like, it's a function of two things. One is like, how exposed have you been to hypergrowth in the past? And then How much time do you put in to actually understand the context of the company so that they, that…”
Lazarte: Most of my Benchmark deals are sourced via active outbound
“No, so the majority, like the majority of what I do is like outbound. Like the majority of what I do is outbound.”
Lazarte: I have done four Benchmark deals, co-founding one of them
“Yeah, so I've done, so I mean, I've done four companies since I started at Benchmark. Three have been this way. And then the fourth one, I actually, I'm a co-founder. So I started the company with someone that, that I knew for a long time.”
Lazarte: Reaching $10M revenue no longer guarantees category victory
“SaaS was a moment in time where, like, rule-based investing worked really well. You know, like, you get to ten million dollars in revenue, like, you're a category winner. I think this is no longer true.”
Lazarte: AI replacing human workers is venture capital's top opportunity
“One thing that I think is super exciting right now is, like, just replacing people. It sounds really bad when you say it this way, but I actually think it's the most exciting opportunity and venture right now, and it's actually gonna be fantastic for humanity.”
Lazarte: AI models will surpass humans at job interviewing
“I think models are gonna be better at interviewing than people. And like the recruiting process now is just so inefficient that if you have like a model interviewing, like that's going to make it a lot better.”
Lazarte: AI market shift is much bigger than mobile
“I think the reality is, AI is a big shift. I think AI is much bigger than mobile. Maybe, like, all the other magnitude of the internet.”
Lazarte: US labor market spend is $10T versus $1T software market
“If software spending in the US is like a trillion dollars, but labor spend is like 10 trillion, and now, like, the AI companies are going after the labor spend, and so the outcomes are going to be, like, insanely larger than they were before.”
Lazarte: Only 1% of computer knowledge work will exist in 10 years
“What percentage of knowledge work as we know today will exist in 10 years? Like if you define work, knowledge work as work you do behind a computer, I think it's probably like one percent.”
Stebbings: UBI destroys purpose and drives human addiction and vice
“I don't buy UBI, by the way. I think that is fundamentally lacking. You don't? No, humanity requires purpose. If you remove, if you add UBI, you remove purpose. And people say, oh, we're going to fish and we're going to paint. I promise you that is not how hum…”
Lazarte: U.S. is winning the AI race against China by all measures
“I think by all, like, I think by all measures, like the U.S. Is winning. If you look at, like, all the AI products, like, ChatCPT is the most used, ah, AI assistant in the world, and sure, you can talk about, like, oh, how about DeepSeq and this and that, but …”
Lazarte: ChatGPT reached nearly one billion monthly users in three years
“Three years ago, like, we didn't have ChatGPT, and now there's, like, close to a billion people that, that use it every month”
Lazarte: Consumers will soon blindly follow instructions given by AI apps
“I think this thing that will happen quite soon is, like, we're gonna have an app, and like, you're gonna wake up, and you're gonna look at your app, and you're gonna read what it says, and you're just gonna do it. Like, the app will tell us, like, go do this. …”
Lazarte: Benchmark participated in a funding round for Cursor
“So we made, like we participated in one round of cursor.”
Lazarte: Writing small checks as one of many investors makes zero difference
“But the reality is, like, if you write a small check, and, like, you're one of many, then, like, you don't make any difference. It doesn't matter how smart you are.”
Lazarte: Benchmark's lack of process and small team are key advantages
“And I actually think that this is a big advantage for benchmarks, like, because we don't have any process, and we have a small team, it's easier for us to, like, change things, right?”
Lazarte: Board members should amplify founder ambition, not focus on governance
“It turns out that it's more important to, like, maximize the upside than it is to manage your losses, right? So, and the way this changes the role of a board member is, you're not that concerned, it's like, oh, like, some of your investments, like, are gonna l…”
Lazarte: VC fiduciary duty ends at check writing to fully support founders
“And my fiduciary duty comes in at the moment where I write the check. It's like, I'm underwriting that person. It's like, hey, do I think this guy will be a good CEO? And once you're there, it's like, I'm just, I'm here to enable this guy. And like, even when …”
Lazarte: Benchmark doubles board meetings for struggling startups instead of walking away
“A lot of venture capitalists would say, hey, like, just cut it loose, you know, like, forget about this thing. It's like, we should double the frequency of the board meetings. It's like, oh, okay. And it's because we're here for as long as the founder wants to…”
Lazarte: Benchmark partners can do deals regardless of partnership vote outcome
“Typically, like, you bring the company in, everyone talks to the company, and then people vote, but it doesn't matter what the vote is, if you really want to do it, you can do it.”
Lazarte: Benchmark bans partners from voting 5 on investment opportunities
“So like, we vote from one to 10, and like, you can't vote five.”
Lazarte: Ten trillion-dollar companies will be created in the next decade
“There's going to be 10 trillion dollar companies created in the next decade.”
Lazarte: Benchmark rounds are defined by founder quality, not check size
“And a benchmark round is when there's a founder that is very capable going after a very big market, and as long as you have that, like, that's a benchmark founder, it doesn't matter if you're raising, like, five million or if you're raising, like, two hundred …”
Lazarte: My first check at Benchmark was $55M into HeyGen
“My first check at benchmark was a fifty-five million dollar check into HeyGen.”
Lazarte: Benchmark's current venture fund size is $600 million
“It's actually a six hundred million dollar fund.”
Shorts cut from this episode
▶ How AI is changing the future for Grads 😱 · 20VC with Harry (@0:00)
▶ “You Should Still Study Computer Science” 🧑💻 · 20VC with (@49:48)
▶ How Benchmark Pick Their Founders 🔎 · 20VC with Harry Stebb (@1:01:25)
▶ Is Silicon Valley still the best place to start a company? � (@23:32)
▶ Duolingo’s AI teacher is genius 🦉 · 20VC with Harry Stebbin (@1:10:09)
▶ The Truth about AI? 🤖 · 20VC with Harry Stebbings (@0:00)