one thin slice. hot dog stall, not hotel buffet

A founder showed me a deck last month: a layer-1 network (L1), a new authority primitive, enterprise clients, eight patents, raising 15 million. I lost him at slide two. What you have is a vision. If all these pieces come together, it works. That’s not how a pre-seed raise happens. A pre-seed happens on one problem, one use case, enough people who’ll pay for it.

What he presented was a hotel buffet. What gets funded is a hot dog stall. The best hot dogs in the city. Then next year you add pizza. In year three something else. Someday Domino’s sells burgers too. Not in year one.

Same thing with a solo engineer building borrow against tokenised stocks. Brilliant wedge. Then he showed me perps, hedging, strategy builder, backtesting. I told him to hide all of it. When I hear your company name, what do I think? It can’t be three products. In a very thin slice, if you’re number one or two, nobody can displace you. If you’re number five in four areas, nothing matters. Nobody bets on that.

Facebook’s first version had one feature. You could see who at Harvard was single. No photos, no wall, no comments. One campus, one use case, until people kept coming back. Then the next campus. You don’t have resources, and you won’t until you IPO, so you focus or you die.

The question that reveals it: “what if Claude does this” or “what if Google does this”. When a VC asks you that, the real message is that your ICP is so wide the product looks like anything, and anything looks like Claude. A founder told me his ICP was “creators and B2B”. That’s everyone, which means no one. Narrow it to: traders who’ve traded 100 times this month with 1 to 5 million in volume who already run a script and want to move to AI. That narrow. Now I can picture one person with the problem.

There’s a reason this is hard. Smart people want to do everything. In India it has a name. The Dhirubhai problem. I’ll own the oil, the plastic, the whole assembly line. That works once a century. Everyone else dies trying.

the call where i said it best

A Polymarket-style betting app wanted Monad to amplify its launch across all markets. I said pick three. The World Cup. Your favourite team plays, the bot bets 50 dollars for you, no buttons. Liquidity concentrated on one thing people actually want, instead of dispersed across everything. Get 30,000 users on that. Then the adjacent category opens up. Then the next one. That’s how everyone who’s won did it.

do this this week

Write your ICP as one sentence that describes a single person with a number in it. If the sentence needs “and”, cut.

the line to remember

just choose a very small battle but win the hell out of it. if you come number five in four areas, nothing matters.