I played piano at a venue called the Palm Court. It was a proper piano in a proper room, which meant I was playing for an audience of people who were paying money to be there, and if I was bad, they would know it immediately.

For several years, I was reliable. Not brilliant. Reliable. I showed up, I played, I didn’t make spectacular mistakes. The management knew that whatever kind of evening I was having, the job would get done.

One night—one particular night—I showed up having already done the serious work on a bottle of wine and made a very good friend in a much stronger drink after that. I remember being at the piano. I remember it being a problem. I remember it being obvious that this was not my best work.

The next day, I expected the conversation. The professional one. The one where they explain that they’d need to find someone else. I had broken the covenant: they had paid me to be reliable, and I had been the opposite.

They rehired me the very next day. Not because they didn’t notice. They absolutely noticed. But because over years of regular work, I had built something. A reputation. A pattern. One night did not erase it.

I learned something from that, though not immediately. It took me years to understand what the venue had actually understood in real time: skill plus reliability is a currency. You can make a mess once and still keep your credit because you have deposited months or years of dependable work in the bank. But if you’ve never made a deposit, one mistake is bankruptcy.

In research, this is the difference between the people who ship and the people who don’t. Linus Torvalds didn’t make Linux perfect before sharing it. He made it regular. He showed up. He iterated. He shipped things that worked well enough, which meant people trusted him to ship the next thing. The person who waits for perfect is the person who ships never. The person who ships regularly, even imperfectly, builds optionality and reputation and the ability to recover from mistakes.

In open source, in vulnerability research, in any practice that depends on compounding trust: you are not measured on your best day. You are measured on your pattern. The people who know you, the communities that depend on you, they are betting on your trend, not your exception.

The Palm Court didn’t care that I had one bad night. They cared that I had a hundred good ones. The one bad night didn’t change the pattern; it was just noise in a signal they already understood.

I played at the Palm Court for years after that, and on most nights, I was sober and passable. On some nights, I was better than passable. On one night, I was unreliable, and it didn’t matter because I had already proven I was the kind of person who shows up.

The lesson isn’t “it’s okay to be drunk at work.” It isn’t. The lesson is that reliability is a reserve you build over time. It means that when you do mess up—and you will—you have somewhere to draw from. The people who know your pattern will believe you when you tell them it was an anomaly. The people who know your pattern will hire you back.

Ship consistently. Ship well enough. Ship enough times that one failure is a statistical outlier, not a trend. The people who do that build trust. The people who don’t, don’t.