The Price of a Stunt

Two things happened within a couple of weeks of each other. A founder posted screenshots claiming a YC partner had asked for $250,000 in exchange for a spot in the next batch. The post did over four hundred thousand views before anyone got around to checking the fonts, which did not match, and the spacing, which was off. YC called it engagement bait. No proof ever surfaced. The founder doubled down without naming anyone.

Then a truck started driving around San Francisco with a QR code and the Y Combinator logo on the side, promising instant acceptance. Scan it and you got a joke and fifty dollars off a flight on some app nobody had heard of. It did over four million views. It also got a lawyer-shaped email from YC within a day: take the logo off the truck, today, confirm in writing.

Both worked, if working means attention. Both are being copied right now by someone who saw the view counts and not the aftermath.

Here is the thing about attention as a metric: it is the only one of these that is instant. The bill comes later, and it does not arrive as a number on a dashboard. It arrives as an investor passing without telling you why. As a candidate who was on the fence and then read the thread. As a partner who was going to make an intro and quietly did not. You never see the deals you were removed from. That is what makes the trade feel free.

The people running these stunts are not stupid. They have done a rough calculation and decided that a few million impressions is worth more than the opinion of a few hundred people who matter. In a business where distribution is the hard part, that math can even look right for a week.

It is wrong for a boring reason. The audience you reach with a stunt is not the audience you need. Four million people saw the truck. Almost none of them were going to buy the product, and the fraction who were will not remember the name by Thursday. Meanwhile the small number of people who could actually change your trajectory - who fund you, join you, buy from you at scale - all saw the same thing, and they filed it under a heading you cannot unfile yourself from. Not "clever." Something closer to "will do this again, to us, when it suits them."

Reputation is not a score. It is a set of assumptions people make about how you will behave when it is inconvenient to behave well. A stunt is a very efficient way of publishing that information about yourself. You are telling everyone, in public, that you will borrow someone else's credibility if the borrowing is cheap and the enforcement is slow.

There is a version of loud that is fine. Loud about something true. A demo that is genuinely absurd because the product is genuinely good. A claim that sounds like a stretch and then turns out to be measured. That kind of attention compounds, because when people go look, the thing holds up. The stunt version does not compound. It spikes and then it is a liability with a permanent URL.

The test I would use is simple. Imagine the stunt works exactly as planned, and then imagine the most important person you want to work with sees it, in full context, with the replies attached. If that thought makes you wince, you already know the answer. The wince is the price, and you are just choosing whether to pay it now or later.

Most of the time it is not worth it. Not because it is immoral, though sometimes it is. Because it is a bad trade. You are selling something scarce and slow to rebuild for something abundant and gone by the weekend.