The best advice I ever gave, I gave away by accident

For years, the best thing I made every week was something I never kept.

It didn't happen at a desk. It happened in calls. A mentoring session with a founder. A career conversation with someone weighing up two offers. A quick coaching chat that drifted into something genuinely useful. I'd say my piece, we'd both nod, and the moment was gone. By the afternoon I couldn't have told you what I'd said, let alone repeated it to anyone else.

I always assumed content was a separate job. A thing you sit down to do. Block the time, stare at the blank page, try to sound clever. No wonder I rarely did it. It felt like inventing insight from scratch, on demand, on top of an already full week.

Then I had a small idea that changed the whole equation. What if the content wasn't something I had to create, but something I was already creating and simply throwing away?

So I built a little pipeline for myself. My meeting notes live in Granola, still one of the best meeting-notes tools out there. Thanks to Oo Nwoye for putting me on to it. An AI agent reads those notes, picks out the moments where I actually gave a real framework or a sharp opinion, and turns them into posts in my own voice. (You could use Claude, or whatever agent system you're building.) I won't bore you with the rest of the machinery. The point is what it revealed, not how it works.

The first week was genuinely humbling. The system kept handing me back things I'd said and forgotten within hours: how to structure a first engineering hire, when a decision is a call and not a vote, why most migrations fail at the switch. I give this stuff away constantly, for free, without noticing. The expertise was never the bottleneck. Capturing it was.

There's a real reason this is so common. Founders and operators are experts in their field, not trained writers, and content stalls in the gap between what you know and how to say it. Even the ones who push through tend to burn out within about six months of forced consistency. Willpower was never going to fix that. A better source of raw material might.

Now I have the opposite problem, and it's a far nicer one. I'm sitting on over 260 finished, ready-to-post drafts, and I've published only a handful. I generate more every single day than I can realistically ship. My pipeline runs miles ahead of me. The constraint moved from what do I say to which of these do I post first. I’m thinking about making this a simple product for those who aren’t very techie or AI savvy, would you use it?

If you spend your days giving advice and none of your evenings writing it up, you already have everything you need. You're just letting it evaporate, the way I did for years.

What's the last good thing you said in a call or conversation that you never wrote down?