In the spring of 2023 I watched a friend raise three million dollars on a single sentence. The sentence was good — I will give it that. It named a real problem, gestured at a model that had only existed for a few weeks, and implied, without quite promising, that he was the person to stand between the two. He said it on a Tuesday. By the following Monday the round was oversubscribed and he was choosing which investors to disappoint.
He called me, elated and slightly unwell, the way you feel when something has gone better than you deserved. "I don't even have the thing yet," he said. He meant the product. He had the sentence, and the sentence had turned out to be the product, at least for the purposes of the wire transfer.
That was the texture of the whole year. The release of GPT-4 had done something strange to the market. It had compressed the distance between a plausible idea and a fundable one to almost nothing. For about eighteen months, a clear sentence about what AI would change was worth more than a year of revenue had been worth the cycle before. Money is, among other things, a device for measuring belief, and belief had gone vertical.
I am not going to pretend I was above it. I raised in that window too, and I raised more easily than I had ever raised before, and the ease was intoxicating in a way I want to be honest about. When capital is hard to get, the raise disciplines you. Every conversation is a small interrogation, and you leave each one knowing a little more about where your story is thin. When capital is easy, the interrogation stops. People say yes before you have finished being wrong out loud. You walk away with the money and without the lesson the money was supposed to come wrapped in.
So here is the thing I actually learned, which is not "have a good sentence." Everyone had a good sentence. The thing I learned is that how you raise teaches your company how to behave, and a raise that comes too easily teaches it the wrong things.
A round closed on a sentence tells your team that narrative is the unit of progress. It tells you that the next milestone is the next story, not the next shipped thing. I watched companies — good ones, run by serious people — quietly reorganize themselves around the production of fundable sentences, because that is what had worked. The deck got better every month. The product did not. They had trained themselves, with real money, on the wrong reward.
The founders who came out of 2023 in good shape, the ones I now watch from a little distance with something like envy, did something unglamorous in the middle of the mania. They raised on the sentence — you would have been a fool not to, the terms were absurd in your favor — and then they spent the money as if they had raised it the hard way. They kept interrogating themselves after the investors had stopped. They treated the easy round as a loan against a discipline they still owed, and they paid it.
My friend with the three million is still around, which in this vintage is its own kind of success. He spent the first six months building the thing the sentence had described, and discovered, as you always do, that the real problem was three layers beneath the one he had named. The sentence had gotten him the money. It had not gotten him the company. Those turned out to be very different acquisitions, and only one of them was for sale.
If you are raising into a hype cycle — and if you are reading this, you probably are, because there is always one — the question is not whether your sentence is good enough to get the check. In a hot market the check is the easy part, and the easy part is exactly the part that will mislead you. The question is what the ease is concealing, and whether you have the discipline to go looking for it after the money has already told you that you don't need to.