The World's First Bullshit

I opened Twitter this morning and three different startups were announcing "the world's first" something. An AI CMO, an autonomous AI marketer, and a design agent "with taste," which is a phrase that made me close my laptop for about ten minutes.

None of them are the world's first anything. I'd bet money there are 40 AI marketing tools already shipping, maybe more, and the category lines are so blurry that "first" really comes down to how specific you're willing to get with your label. I could build an AI that writes haikus about SaaS pricing pages. I could call it the world's first AI SaaS Pricing Haiku Engine. Nobody could argue, because nobody else would have tried. That's what "first" means now. You've found an unclaimed sliver of territory so narrow that being the only person there is trivially easy.

But OK, forget that the claims are fake. What bothers me more is that "world's first" is the wrong thing to want in the first place. It's the wrong trophy entirely.

Thomas Newcomen bolted together the first commercially successful steam engine in 1712 and the thing was, by every measure, awful. Maybe 1% thermal efficiency. It ate coal like a bonfire eats kindling and it leaked through every cycle. James Watt showed up 57 years later, added a separate condenser, and built a version you could run a factory with. Newcomen got a Wikipedia footnote. Watt got a unit of measurement named after him. Google wasn't the first search engine. Facebook wasn't the first social network (Friendster was, and if you remember Friendster, congratulations, you're old). The iPhone showed up years after the Blackberry and the Palm Treo. Who ended up mattering? The ones people actually liked. Every single first mover on that list became a piece of bar trivia.

Founders keep doing this, I think, for two overlapping reasons and one of them is almost sympathetic. Silicon Valley has a mythology, basically a religion at this point, where the inventor is the saint and the timestamp is the sacred relic. The Xerox PARC researchers who built the graphical user interface were visionaries; Steve Jobs was the guy who walked through their lab, took notes, and shipped something your mom could buy at a mall. The mythology says PARC mattered more. In practice, Jobs built the product, and products are what people use.

But I think the bigger driver is more cynical than that, and it has to do with Twitter specifically. "World's first" is a cheat code for the algorithm. You can't verify it while you're scrolling, it sounds historic, and it carries a weight that "we also built an AI marketing tool, we think ours is pretty good" never will. The most extreme claim gets the most retweets, and retweets are what your investors see before they decide whether to write a check. I get why people do it. I'd probably be tempted too.

The problem is who it attracts. Novelty-chasers. People who'll try your product once, talk about it at a dinner party, and never open it again. The people who actually make a product successful long-term are the ones who care that your thing works well, and those people could not care less whether you were first or forty-first.

The second version of something is almost always better than the first, because the second version watched the first one break. Every good product is a correction of somebody else's bad product. That's how engineering works in practice: you watch someone else's bridge come down and you build yours with thicker cables.

What would it look like if founders were honest about this? "We looked at the 14 AI marketing tools that already exist and we think we've solved the 3 problems they all share." A less exciting tweet. But it's certainly more useful. It tells potential customers that you've done your homework and you're competing on substance rather than on who filed their launch post on Tuesday instead of Wednesday. Most founders would rather sound historic than sound competent, and their customers can tell.

The AI startup space right now feels like the early days of a gold rush, when the most important thing is to stake your claim loudly before anyone else reaches the same patch of dirt. But gold rushes end.

The people who built lasting wealth in the California Gold Rush were largely the ones selling picks and denim jeans to miners, the ones who understood that being best at serving a need that wasn't going away beat being first to a plot of land every time.

Levi Strauss arrived in San Francisco in 1853, 3 years after California became a state, and he wasn't first to anything, but he's still around.

Every founder tweeting "world's first" today should ask one question about their product: would anyone still care about this if 10 other people had built the same thing? If yes, you might have something, and if no, if the only interesting thing about the product is the claim of novelty, what you're looking at is marketing copy where the product should be.

You can announce "world's first" on launch day, before anyone has used your product, before anyone has even seen a demo. You can never announce "world's best." Other people have to say that about you, and they'll only say it if you've given them a reason to.