When Mick stopped working for other people in 2020, after forty years in software engineering, he took a year to learn whatever he felt like learning. Online courses. His first hackathons. He hadn't written production code since the early 2000s, so he taught himself Python and got good enough to build with it again. No plan for any of it. That was the point.
Kevin stopped around the same time, after forty years in oil and gas accounting that ended with him as CFO of a public company — a field where staying current isn't optional and the rules move under you. Somewhere in business school a professor drilled his students until they could get through the Wall Street Journal in ten minutes. Not reading faster. Deciding faster what was worth reading at all. He still reads it most days.
We've been friends since we were thirteen, back to a best-of-25 chess match in ninth grade. Neither of us has gotten any less competitive since.
That's eighty years between us of having to keep learning after school stopped teaching us, and two different ways of going about it. Mick accumulated: trade magazines, then bookmarks, then blogs, then podcasts and newsletters, more sources every year and a growing conviction that the good stuff was out there if you could find it. Kevin triaged: a lot arrives, most of it isn't for you, and the skill is telling the difference fast.
Both of those turn out to be the same problem seen from opposite ends.
Neither of us wanted to be done working. We wanted to build something that mattered, and we had a list of ideas we'd been adding to for years.
Some got five minutes. Some got weeks. Two got far enough to build — a government data visualization tool, and a newsletter — before we set them down and went back to the list.
Then generative AI arrived and made something buildable that we already believed: that the hard part of learning on your own isn't the reading. It's deciding what deserves the reading.
Deep Digests came out of that. You choose the sources you want to follow, it reads across them every day and hands back summaries, and you decide what gets your full attention. It's the first thing we built that we didn't set down.
We built Deep Digests first and went looking at the domain it belonged to afterward.
Once the product existed, the question was what else learning on your own actually involves. We've landed on five parts, though not on how they fit together:
Deep Digests works on one of those. We're still thinking through the rest.
We could spend a long time getting the map right. We'd rather build for the parts we understand well enough to build for.
So that's mostly what we're doing. We write as we go too — what we're working out about AI, and what we're working out about learning itself.
That's our end of it. We build for it and write about it. You're the one doing the learning.