Summary

Clay was already taking off before the LLM wave; three assumptions shaped its 1-to-100 growth in two years. Kareem Amin’s claim: set the company up so “when the wave comes, you go with it” — Clay was already taking off before ChatGPT because it allowed any integration and was “a coding-like product just designed with go-to-market primitives.” The three assumptions — go-to-market people are creative so give them the most powerful tool, sell it to RevOps under the job title “go-to-market engineering,” and charge for usage instead of per seat because productivity shrinks headcount — cascade so cleanly that “almost every other decision in the company falls immediately out of those.”
The counter-position: rivals build simple, “coin-operated” tools that hand salespeople the answer; Clay bets sales is a creative task and the point is finding “go-to-market alpha” — differentiation, “otherwise it’s just noise” — which needs an open-ended experimentation tool. His product metaphor: “We’re not building a microwave… we’re building a guitar” — commoditized utility versus an instrument you spend a lifetime mastering.
“I think capitalism rewards risk more than anything else” — not hard work (“a lot of people working really hard and not making a lot of money”), not skill (sword-swallowers earn less than YouTubers). Real risk requires genuinely not knowing the outcome plus “a high potential for shame.” His test for founders faking it: one told him “we serve sales and we serve recruiting” — “You have to pick one.”
Creating “post-lack” is presented as an operating edge, not therapy-speak. Against the VC chip-on-the-shoulder trope, Amin decided “I’m not going to create from a place of lack anymore” — and found it’s “great negotiation leverage cuz there’s nothing you can give me.” Wholeness, he argues, produces more risk-taking and less organizational politics; to the post-scarcity objection that no one would work: “Have you been to Burning Man? People are still doing stuff.”
Talent strategy runs opposite the “hire fast, fire fast” consensus: Clay over-invests where others under-invest (recruiting, brand, content, community — an early content hire was paid “like a PM”), and sticks with floundering-but-excellent people past nine months; 50% become superstars. The verdict is situational, never global: “Were you good in the moment that we needed you to be?”
Vision is optional; honesty about it isn’t. “If you don’t have a vision, then you don’t have a vision” — most medical discoveries and many other breakthroughs, including the cosmic microwave background, are coincidences, and retold-clean founding myths “confuse the people who are on the path of figuring it out.” Clay’s current vision, held clearly: helping every company “find who their best customer is and find more of them and accelerate that.”
The provocation investors should sit with: a “death doula for companies.” Amin questions whether all businesses should scale (“this restaurant was really good, now it sucks”), citing Microsoft achieving an initial mission (“put a computer on every desk”) and now having “devices and services” as its mission. If a company has achieved its mission, “should we help it pass? Maybe it should have children” — better than incentivizing “zombies or… worse versions of their former selves.” He also asked five people how to scale from 100 to 300 and did the opposite of their advice; many predicted problems scaling 50→300, and “none of [them] happened.”

Deep dive

1. Clay began as “give the power of programming to more people” — years before LLMs

The founding ambition, with co-founder Nikolai: in the 21st century “you have computers that can do things for you and if you can’t tell them what to do, then you can’t capture any of the value that’s being created.” The inspirations were Ableton Live — extracting an idea from mind to machine — and the 1960s “mother of all demos”: fluidity and low latency between idea and execution is what leads you to the next stage of thinking.
His material metaphor: “if in the past it was oil… or steel… today it’s computation” — and spreading it is egalitarian. They considered reinventing the terminal but rejected it (“that’s just going to speed up engineers”), chose business users over consumers in 2017, and landed on sales and marketing: creative people who can’t execute their ideas.
The example as told: a customer sells to businesses with lots of garbage, so they use Google satellite view to find areas with garbage accumulation and target them — “an idea that you might have, but it might be super hard to execute at scale. And now something like Clay can actually translate that for you.” Hence the name: it’s go-to-market, “but now you’re an engineer.”

2. The counter-position: sales is a creative task, not a coin-operated one

Competitors’ pitch — people have “literally used the word coin-operated” — is a simple tool that tells you who to sell to and what to say. Clay’s bet: sales is about “go-to-market alpha… how are you different than everybody else? Otherwise, it’s just noise” — which demands a powerful, open-ended tool for fast experimentation. “We built that instead, which was very counterintuitive.”
Chapter two wasn’t a ChatGPT story: the company “was already taking off before ChatGPT came out,” and was positioned for LLMs by two prior choices — any-integration architecture and being “a coding-like product just designed with go-to-market primitives.” “As soon as we had LLMs, it just boosted everything.”

3. Three assumptions from which every other decision falls out

The 2022 decision stack was deliberately non-obvious: target go-to-market teams, specifically outbound, specifically users pulling data from many providers — and agencies, not startups — while supporting them in public and building community. “The main thing that happened was really the courage and commitment… to making those decisions.”
The three core assumptions: go-to-market people are creative, so give them the most powerful tool rather than the simplest; a powerful tool needs the right user, so sell to RevOps and frame the job as “go-to-market engineering”; and charge for usage instead of per seat, because Clay drives productivity and shrinking headcount shouldn’t be an anti-incentive.
The management payoff: “almost every other decision in the company falls immediately out of those. Anyone can come up with what we should do next if you believe these 3 things” — which he calls the best way to run a company going 1 to 100 in two years: communicate the assumptions, let everyone check decisions against them, and point the ship “so that when the wave comes, you go with it and you’re not fighting it.”

4. Statues off the elevator: courage first, because “capitalism rewards risk”

Amin had a dream of a statue-filled modern city and noticed old republics built statues to concepts — courage, integrity, justice. “I feel like people don’t talk enough about things that actually matter in building a business.”
His first statue: courage, because “I think capitalism rewards risk more than anything else.” Not hard work (“there are a lot of people working really hard and not making a lot of money”), not skill — sword-swallowers and jugglers earn less than YouTubers. Hedged as observation, not endorsement: “I don’t have a strong comment on whether that’s good or bad. I’m just trying to figure out what’s actually happening.”
His definition of real risk: you genuinely don’t know what will happen, coupled with “a high potential for shame” — that’s when you’re about to discover something new. The specimen anecdote: a YC founder told him “we serve sales and we serve recruiting.” His reply — “You have to pick one” — because the value is in committing to serve a specific group in a specific way.

5. Justice, long-term greed, and self-respect as the only judge

Justice earns its statue because it’s the precondition for stability: “there’s no stable way for some group of people to dominate others… the other person is a tremendous nuisance if they feel like they’re treated unfairly” — true for companies and countries alike. In practice: respect and fairness “in every situation, whether you’re hiring them or firing them or you’re disagreeing.”
The personal root: even fouling opponents his own team could get away with in soccer felt off — “I’m a long-term greedy person,” and “the only way to be long-term greedy is to stay in integrity.”
On hearing Brian Chesky’s hundred-billion-dollar remark, he related from a smaller scale: at Clay’s Series A he felt that adulation, prestige, and money are “an empty thing that passes.” “The only judge of anything is yourself and your own self-respect” — and an ambitious person’s standards are higher than anyone else’s “cuz you know everything that you’re thinking.” Patrick’s endorsement: after their last meeting, “optimizing for self-respect over everything else” was the one note that survived his hour-long memory filter.

6. Post-lack: creating from wholeness instead of the chip on the shoulder

Against the VC trope of investing in founders with “a chip on their shoulder” — and the creative-work version, fearing you’ll lose your productive anxiety — Amin’s turn came from friends saying “I don’t care if you succeed or fail. I love you.” Realizing he could have that now, he went around “probably pretty annoyingly” telling people, “I have everything I need. I’m good.” Post-lack is also “great negotiation leverage cuz there’s nothing you can give me. I feel great.”
To the objection that a post-scarcity world would stop working: “Have you been to Burning Man? People are still doing stuff.” Wholeness, he argues, makes you less destructive — the reason companies chant “customer first” is that unwhole employees think about careers and internal politics instead. “The more whole you can be… the more risks you can take.”
The retreat that crystallized it: ten days silent at a Hirudaya retreat (non-dual “direct path,” not Vipassana — meditating on the space between breaths, asking “Who am I? A very frustrating question. I would not recommend it”). Day five: feeling connected to all things, “no better than and no worse than anyone else.” Day seven: catching himself rehearsing telling friends about it — anticipating the future is “Freud’s death drive” — then snapping back into “such delicious boredom.”
The tradeable insight from one pea exploding in his head: he has an accumulation tendency — food scarcity at the retreat, and at the company “where’s the next ARR going to come from?” “Now I’m aware of when I have enough. And then I can take more risk.”

7. Unusual operations: over-invest where others under-invest, commit longer to people

Clay over-invests in recruiting, brand, content, and community — hiring overqualified people into roles doing them for the first time and paying accordingly: early employee Mishti “would have never taken a content role… we compensated her like a PM and we’re like, go do it.”
Against “hiring fast and firing fast”: when someone excellent is floundering nine months in, the questions are “Is it the role? Is it the context? Is it our decision-making that’s limiting them?” Outcome: superstars 50% of the time; the other 50%, “we just can’t create the conditions for them.” No global verdicts — “Were you good in the moment that we needed you to be?” — and people he knew were incredible have failed at Clay and thrived elsewhere.
The weirder inputs are deliberate: clowning (“edgy improv” — physical, emotional range, the jester “telling truth to power,” and the trickster, who tells “spiritual truths through jokes”) and magicians from 69 Atlantic, whose real craft is storytelling that sparks curiosity. Clay’s coming project: “re-enchanting the world” — in a world with AI, “this is magic,” and remembered wonder makes you more curious about customers. The brand takes more risk than B2B norms allow because “a lot of times people are authentic, but hedging — not really saying what they want to say because they’re afraid of the consequences.”

8. Music as management theory: endings, dissonance, and the guitar

Music’s lesson is that time and silence are materials: “where the room isn’t is where we are.” Companies obsess over beginnings and never think about endings — yet running one means “a microcosm of humanity… a mini society” — and songs model the options: fade out, end suddenly, end “with a whimper or intensity.”
From an essay called “the emancipation of dissonance”: music history as accepting ever more sounds as music — the fifth, then the third (chords, initially resisted), then the seventh (“you get jazz, and people were like, that’s unacceptable”), on to microtones and electronic anything. His product principle: “what are all these other ways of being that you think are off that actually could be productive if you put them in the right context?”
Picking up Patrick’s point that when anything can be created, pathos becomes important on top of logos: “we’re not building a microwave… we’re building a guitar” — six strings, twelve frets, a lifetime of mastery. And the idealized form of go-to-market: “It’s raining, I bring an umbrella. You were like, ‘Great. Thank you.’” — finding the person who needs the thing without manufacturing the desire.

9. Vision is optional — honesty about it isn’t

His answer to whether leaders must have a vision routes through the truth statue: “If you don’t have a vision, then you don’t have a vision. Don’t say something that actually isn’t happening.” Most medical discoveries are coincidences — “most of medical discoveries… the cosmic microwave background. You couldn’t have gone out and said, I want to discover the cosmic microwave background” — and dishonest retellings “confuse the people who are on the path of figuring it out.”
Patrick’s compression — “optimize for discovery rather than destination” — gets a refinement: “leave the space for the opportunities to discover,” take the next step, and let the vision come (he invokes vision quests: sometimes you act first and receive the vision after). Visions can also expire — companies started five or six years ago “are kind of confused now… it’s time to change the vision,” and it’s fine to say “we don’t actually know what we’re doing right now.”
Right now he does have one, held clearly: “helping every company in the world find who their best customer is and find more of them and accelerate that.”

10. Radical clarity in hard conversations — and doing the opposite of the playbook

Not all problems are communication problems, but often they remain unresolved because “we’re making assumptions about each other… doing game theory things.” His strategy: clarity about what he wants and what you want — “we might still disagree, but at least we are disagreeing from a place of clarity.” Firing example: “I’ve lost faith in the ability for you to bridge the gap. And I’m the CEO… I’m not saying that this is rational even… I’m taking responsibility… this is my call” — plus telling the departing person exactly what he’ll say on reference calls, including “the gaps that we have in our organization that didn’t support them.”
On scaling folklore: he asked 5 people how to scale from 100 to 300 and did things completely opposite to their advice. Separately, many predicted weird problems scaling from 50 to 300 people; none of them happened — following rules of thumb just imports their authors’ problems.
The closing provocation, from Patrick’s oldest note on him — a “death doula for companies”: should every business scale? “This restaurant was really good, now it sucks.” Microsoft achieved “put a computer on every desk” and now has “devices and services” as its mission; he said the idea of a company may date to the 1500s, perhaps the East India Company, with companies convening for a purpose. “If your company has achieved its mission, should we help it pass? Maybe it should have children” — rather than incentivizing “zombies or worse versions of their former selves.” His closing stance on Clay, after achieving wealth (“there’s no amount of wealth that fills” a lack — what it buys is choosing what to do with your time): “I’m not attached because I don’t need anything from it. I want it to evolve the way it wants to evolve.”

He's Building an AI That Can't Lie | Dan Klein, Scaled Cognition Cerebras Early Investor 周楠 on the Nvidia Challenge and Baidu's US Lab