Jun 25, 2025 · 41m · a16z
Building Cluely: The Viral AI Startup that raised $15M in 10 Weeks w/ Roy Lee
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 the a16z podcast, Cluely co-founder Roy Lee and a16z partner Brian Kim join host Erik Torenberg to discuss how Cluely raised $15 million in 10 weeks by leveraging viral distribution, creator networks, and momentum as a moat in the AI ecosystem.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The host holds 20% of the talking time here. How this is scored →
The host as informed peer 3.1 Guest teaching 3.4 Guest disagreement 1.8 The host pushing back 1.5
speaking balance: gold is the host, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 33:00 Roy rejects product timeline critics
Roy vigorously rejects the premise of internet critics asking where the product is, asserting that building massive pre-launch hype is the superior playbook.
Hardest push from the host ▶ 30:00 Host challenges defensibility against incumbentsThe host refuses to accept simple viral momentum as sufficient protection, pressing Roy on why OpenAI will not simply copy the overlay feature using their superior reach.
Biggest teaching moment ▶ 8:13 Roy explains algorithm arbitrageRoy re-frames how social media algorithms operate, educating the host on why short-form content feeds on TikTok and Instagram create dynamics that tech Twitter users fail to comprehend.
The host holds their own ▶ 23:53 Host synthesizes historical venture thesisThe host draws clear parallels between Paul Graham's early YC thesis on technical founders, Andrew Chen's channel exhaustion theories, and modern creator commerce to explain Cluely's positioning.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The host as informed peer | Guest teaching | Guest disagreement | The host pushing back | Why |
|---|---|---|---|---|---|---|
| a16z Disclosure and Legal Disclaimer | 2 | 1 | 1 | 1 | The host facilitates light intro banter regarding the viral announcement, making minor observations about over-analyzing social media threads. Roy and Bryan share the positive and negative reactions to the launch in an agreeable tone. | |
| Roy Lee's Origin Story and Getting Rescinded | 1 | 2 | 1 | 0 | Roy recounts his background, getting rescinded from Harvard, and spending a year at home. The host asks standard narrative questions without pushing back or offering deep analysis. | |
| Social Media Algorithms and Platform Arbitrage | 3 | 5 | 2 | 1 | Roy details his thesis on platform arbitrage, short-form algorithms, and how tech Twitter lags behind Instagram and TikTok. The hosts contribute peripheral examples like 4chan and meme supply chains while Roy educates on viral mechanics. | |
| Building Cluely and Deploying a Creator Workforce | 2 | 5 | 2 | 1 | Roy explains his operating philosophy of dividing startup roles strictly into world-class engineers and world-class influencers. The host asks clarifying questions about revenue conversion while Roy strongly asserts his distribution thesis. | |
| The Deal Story: Brian Kim's Pursuit of Cluely | 1 | 2 | 1 | 0 | Bryan Kim tells the story of pursuing Roy and closing the investment deal despite Roy initially turning down multi-stage VCs. The host purely listens to the narrative. | |
| Investment Thesis: Momentum as a Moat in AI | 6 | 4 | 1 | 2 | The host demonstrates strong domain knowledge by connecting Paul Graham's YC thesis, Andrew Chen's essay on marketing channel fatigue, and modern audience co-founders like MrBeast. Roy and Bryan elaborate on using distribution velocity to find product-market fit. | |
| UX Innovation: The Translucent AI Overlay | 5 | 4 | 3 | 4 | The host presses Roy on defensibility and the threat of OpenAI building translucent AI overlay features directly. Roy confidently claims Cluely can out-distribute incumbents and explains the translucent UI discovery. | |
| Staged Strategy and Shifting Corporate Culture | 5 | 4 | 3 | 3 | Roy aggressively dismisses online critics who ask where the finished product is, arguing pre-launch distribution guarantees PMF. The host provides synthesis by comparing Cluely to corporate-driven media trends and shifting professionalism standards. |
Statements from this episode (18)
Roy Lee: Cluely reached $1M enterprise revenue within 10 weeks
“We wrote our first lines of code 10 weeks ago. All of a sudden we have like over a million dollars of enterprise revenue coming in.”
Roy Lee: Replace marketing heads with fewer than 100k social followers
“If any company in the world has a marketing team and the head of marketing does not have a 100,000, at least a 100,000 followers. Like, you need to replace them. Like the game has changed.”
Roy Lee was a college student six months before raising $15M
“I think, like, literally six months ago, I was some random college kid in a dorm, and now I feel like I'm at the center of the tech universe.”
Roy Lee: Tech workers on X and LinkedIn misunderstand video algorithms
“I think it's growing increasingly clear that people on ex-LinkedIn are behind, and there's such a very, extremely small intersection of people who understand how developed the algorithm is on, like, Instagram, TikTok, and people on tech, like, Twitter, LinkedI…”
Roy Lee lost early Harvard acceptance after police escorted him back
“I got accepted to Harvard early, and then later that year, I was just always doing crazy shit, and I snuck out of a school field trip past curfew, like, the police had to come and escort me back, because we were out at, like, two a.m., we were all, like, 1617,…”
Roy Lee's parents run a Harvard admissions consulting business
“My parents literally run like a college admissions consulting company. So we literally teach kids how to get into Harvard.”
Roy Lee: Instagram requires more controversy for virality than X
“And this is something that I've said before, but I guarantee, like, my videos do not go as viral on Instagram, like, Instagram TikTok. And the sole reason is because on those platforms, they are not controversial enough.”
Roy Lee: Future startup founders will overwhelmingly be Gen Z creators
“You're not gonna get more millennial founders. You're only going to get more Gen Z founders. And I guarantee you, Their backgrounds and content are the exact same as mine, and they know exactly what I'm doing, and I'm sort of like the canary that's leading the…”
Roy Lee: Every full-time Cluely employee has over 100,000 social followers
“There are only two roles. You are either a world-class engineer who's building the product or you're a world-class influencer. And for a full time every single person has over a 100,000 followers on some social media platform.”
Roy Lee: $20,000 in short-form video equals Super Bowl ad reach
“Look, we pay very little money for the amount of views that we get, and different companies, like, they're paying literally millions of dollars for Super Bowl ads when you can get the same quality and quantity of views. 20,000 dollars.”
Kim: Momentum is the primary moat for early-stage AI startups
“Eventually we're turned into a product that need to be retained, that need to be used every day, but we're still in this early stage of AI where I think momentum is the moat.”
Lee: Predecessor tool Interview Coder reached 250 million impressions
“What we realized after we got about, like, two hundred fifty million impressions on the whole scenario is, like, wow, we just got so many eyeballs on this thing”
Lee: Cluely has surpassed 1 billion total content views
“And as a result, like now we're past like a billion views overall on clue.”
Lee: Strong distribution eliminates the need to interview users
“I feel like you don't, like if your distribution is strong enough, you don't need to like talk to your users. You just need to look at their data.”
Roy Lee: Translucent liquid glass overlays are the future AI interface
“I think everyone's going to inevitably get the translucent overlay. This is how an integrated AI should feel, and, like, Apple shows everyone that, like, liquid glass is the translucent overlay that, that, that, that will be the form factor of AI in the future…”
Roy Lee predicts Cluely will achieve better distribution than OpenAI
“If the question is about distribution, then I think there's actually, like, a pretty strong case for us to make that we will actually end up distributing better than OpenAI, and It's enough that you could probably bet on us at like, well, like a 30,000 X disco…”
Roy Lee: Cluely generates more revenue than the latest YC batch
“We wrote our first lines of code 10 weeks ago. We're, like, earlier than the latest YC batch of companies, yet we're, like, generating probably more revenue than every single one of them, and, like, the product is, like, it's literally two and a half months si…”