Skip to content
a16za16z

Building Cluely: The Viral AI Startup that raised $15M in 10 Weeks w/ Roy Lee

What if virality wasn’t a tactic — but the entire product? In this episode, a16z General Partners Erik Torenberg and Bryan Kim sit down with Roy Lee, cofounder and CEO of Cluely, one of the most talked-about consumer AI startups of 2025. Cluely didn’t raise a mega round or drop a feature suite to get traction - it broke through by turning distribution into design: launching viral short-form videos, pushing polarizing product drops, and building in public with speed and spectacle. We cover: – Why virality is Cluely’s moat – Building a brand-native AI interface – The Gen Z founder mindset – What most startups get wrong about attention – Why creators are the new product managers – Cluely’s long-term vision for ambient AI Cluely is a glimpse at the next generation of startups, where the line between product and performance is disappearing. Timecodes: 00:00 Introduction 00:50 Roy's Rise and Controversies 03:33 A Journey of Provocation and Ambition 07:52 Mastering the Algorithm 12:24 Building a Viral Tech Company 14:07 The Power of Distribution 15:49 The Partnership with Brian 20:09 Momentum as a Moat 20:57 The Evolution of Product Retention in the AI Era 21:58 The Importance of Speed and Innovation in AI 24:46 The Rise of Creator-Driven Software Businesses 25:56 Cluey's Journey: From Interview Coder to Viral Success 27:02 The Power of Distribution and User Data 28:54 The Translucent Overlay: A Game-Changer in AI UX 32:41 Building Hype and Authenticity in Product Launches 34:22 The Future of Professionalism and Corporate Culture 40:12 Concluding Thoughts and Vision for the Future Resources: Find Roy on X: https://x.com/im_roy_lee Find Bryan on X: https://x.com/kirbyman01 Learn more about Cluely: http://cluely.com/ Stay Updated: Let us know what you think: https://ratethispodcast.com/a16z Find a16z on Twitter: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Subscribe on your favorite podcast app: https://a16z.simplecast.com/ Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures.

Roy LeeguestBryan KimhostErik Torenberghost
Jun 25, 202541mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 3:53

    From dorm room to “center of the tech universe”: sudden scale and backlash

    Roy and the hosts set the scene: Cluely went from zero to massive attention extremely fast, drawing both admiration and intense criticism. They frame the conversation around how virality, controversy, and distribution are becoming core startup advantages.

    • Cluely’s rapid timeline: first code ~10 weeks ago; sudden enterprise revenue traction
    • The announcement sparked love/hate and widespread meta-commentary
    • Roy’s confidence that his virality assumptions are playing out in real time
    • Distribution and algorithm knowledge as the hidden edge
  2. 3:53 – 4:23

    Provocation as a personality trait: early life, camps of fans/haters

    Roy traces his “attention-grabbing and provocative” identity back to childhood and school. He explains how this trait consistently produced polarized reactions—and became foundational to his later approach to content and company-building.

    • Always outspoken and “no filter,” creating strong supporters and detractors
    • Provocation as a consistent through-line rather than a recent persona
    • Polarization framed as a feature, not a bug
    • Setting up why controversy maps naturally to modern algorithms
  3. 4:23 – 6:00

    Harvard rescinded, a year alone, and the decision to swing big

    Roy recounts getting accepted to Harvard early, then being rescinded after an incident that triggered a pile-on. The ensuing year at home became a psychological crucible that pushed him toward going all-in on building companies.

    • Rescinded from Harvard after suspension and “haters” escalating reports
    • Family context: parents run an admissions consulting business (high stakes irony)
    • A year of isolation amplified conviction and risk tolerance
    • Turning point: choosing a maximally “interesting life” and entrepreneurship
  4. 6:00 – 7:51

    Community college to Columbia: appeasing parents, finding a cofounder

    Roy describes the compromise path—community college in California, then Columbia—to satisfy family expectations while staying focused on startups. At Columbia, he quickly meets his cofounder and begins hacking toward what becomes Cluely.

    • California/community college as a pragmatic middle ground for family + building
    • Columbia as a checkpoint—Roy focused on cofounder search immediately
    • Meeting Neil early and iterating on projects until one worked
    • Earliest Cluely version emerges from rapid experimentation
  5. 7:51 – 12:34

    Why X/LinkedIn feels “behind”: short-form algorithms, volume, and controversy

    Roy argues that YouTube democratized distribution, TikTok shifted the game to volume, and short-form algorithms now reward controversy and digestibility. He claims tech Twitter/LinkedIn underestimates how aggressive these dynamics are compared to Instagram/TikTok.

    • From paid ads to quality content (YouTube), then to content volume (TikTok era)
    • Most X/LinkedIn posts optimize for intellectual status, not viral comprehension
    • Algorithms amplify controversy; X/LinkedIn supply of controversy is low
    • Roy’s content is “not controversial enough” for IG/TikTok—but explodes on X/LinkedIn
  6. 12:34 – 14:17

    Mastering the algorithm through stunts: Interview Coder, launch videos, “50 interns”

    Roy explains how repeated viral hits taught him virality is repeatable, not luck. He uses the Interview Coder controversy as the origin of Cluely’s attention engine, then stacks subsequent viral concepts to dominate timelines.

    • Interview Coder: cheating a technical interview, publicizing it, and triggering backlash
    • Virality as a repeatable craft after successive wins (launch video, “50 interns”)
    • Perceived “alpha”: X/LinkedIn hasn’t fully adapted to short-form logic
    • Using narrative provocation as a distribution wedge for a tech product
  7. 14:17 – 15:49

    A new marketing org chart: engineers + creators (and the contractor content factory)

    Roy details Cluely’s internal model: the company is built around building product and producing viral content at scale. He argues that creator-following is now a qualification for marketing, and that paid creator output can outperform traditional ad spend.

    • Only two roles: world-class engineers or world-class influencers
    • Hiring bar: 100k+ followers for full-time marketing/creator roles
    • “50 interns” concept maps to a modern content-production internship
    • 60+ contractors paid per video; cheap views relative to legacy advertising
    • Social video as the primary conversion channel (IG/TikTok)
  8. 15:49 – 20:09

    How a16z got involved: spotting conversion + ‘momentum as a moat’

    Bryan recounts meeting Roy, observing unusually high-quality inbound interest, and seeing proof that attention was converting to real revenue. He connects this to his thesis that in AI, speed across product and distribution creates defensibility.

    • Initial outreach, hesitation, then relationship-building before fundraising talk
    • On-the-ground signal: random engineers showing up because of Roy’s content aura
    • Turning point: proof of real revenue conversion from distribution
    • “Momentum as a moat”: sustained speed is hard to replicate in noisy AI markets
    • Fast diligence and fast deal-making once metrics clicked
  9. 20:09 – 25:56

    Why retention frameworks break in AI: gingerbread strategy, rapid iteration, and noise

    Bryan (with Erik) explains how traditional moats—slow-crafted products, network effects, retention-first thinking—are pressured by rapid model/platform change. The new requirement is to keep shipping, iterating, and distributing faster than incumbents can absorb features.

    • Old worldview: artisan products + retention as primary signal
    • AI reality: platform capability shifts weekly; features get copied/absorbed quickly
    • Snap analogy: continual invention as survival strategy
    • Early AI phase: “anything goes,” so momentum and iteration speed dominate
    • Founders must unify product speed + distribution speed
  10. 25:56 – 28:55

    Distribution-led product strategy: ship the concept, harvest usage data, then converge

    Roy lays out Cluely’s sequencing: push a compelling UX idea to a massive audience, then let data reveal the strongest use cases. He argues strong distribution reduces ‘shooting blind’ and allows fast validation via engagement metrics and revenue signals.

    • Cluely began as Interview Coder; expanded to a general ‘invisible overlay’ concept
    • Launch before fully ready; use virality to accelerate market feedback
    • Usage data (not just user interviews) guides product direction
    • Testing use cases via content (e.g., sales calls) led to ~$1M+ enterprise revenue
    • Viral fit can be measured faster than traditional market fit
  11. 28:55 – 32:41

    The translucent overlay: the AI UX bet and the land grab against incumbents

    Roy and Bryan zoom into the core UX insight: a semi-translucent overlay that integrates AI into your workflow rather than a separate chat window. They discuss defensibility as a race to own the ‘default overlay’ through mindshare, distribution, and first-mover momentum.

    • Overlay form factor as the breakthrough: AI should feel integrated, not separate
    • Iterated through ~20–30 versions while optimizing invisibility for interviews
    • Generalizing from coding interviews to ‘everything’ once the UX clicked
    • Acknowledgment incumbents will copy; framing as a land grab for default UX
    • Bet that Cluely can out-distribute even major AI platforms for this category
  12. 32:41 – 41:35

    Hype, authenticity, and the end of professionalism: a new corporate media era

    Roy argues modern audiences crave radical transparency and authentic founder-led storytelling, not corporate polish. He shares practical boundaries (e.g., don’t punch down), and predicts a broader cultural shift where companies succeed by being entertaining, controversial, and real.

    • Pre-launch hype as an asset: non-product videos build anticipation for the ‘real’ launch
    • Authenticity as an algorithmic advantage; controversy as anti-fragile attention
    • Guideline: never punch down; balance provocations with genuine respect
    • Declining professionalism as a long-term trend driven by democratized content
    • Vision: if Cluely wins, corporate culture shifts toward radical transparency

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.