Saturday, 15 August 2026, 9:09 AM +02:00
I finished Made to Stick this morning, cover to cover, and I can feel that particular weight lifting — the one that comes from finally closing something you've been carrying around for days. The last chapters on emotion and stories landed harder than I expected. It's not that the ideas are complicated; they're almost embarrassingly simple. Concrete beats abstract. Unexpected beats predictable. Stories beat statistics. But sitting with the book closed, what stays with me is the "Curse of Knowledge" — the way that once you understand something deeply, you cannot un-understand it, cannot find your way back to the mind of someone who doesn't yet see it.
The book calls it a curse, and I used to nod along with that, seeing it as an obstacle to clear communication. But this morning I'm thinking it's more treacherous than that. Because the same mechanism that makes me a bad explainer is the one that makes me a bad learner. When I read something and it feels obvious, familiar, already-known, I have no way to tell whether I've actually mastered it or whether I'm just fooled by the smoothness of recognition. I could read this book ten times and each pass would feel like progress, and each pass would be exactly the same. The curse doesn't just blur my teaching — it blurs my own sense of whether I'm learning at all.
So I've been asking myself: what would it look like to actually earn what I've read? Not to have absorbed it, but to have made it mine in some way that survives contact with the world. The book suggests one test: can I explain it to someone who hasn't read it, in language they can carry? Can I take "stories as flight simulators" — that lovely idea that a story lets you practice a situation without paying its risks — and tell it to someone in a way that makes them see what makes stories work? I think I can, barely, if I use an example instead of a definition. I can point at the way a good engineer's anecdote about a bridge collapsing teaches more than a manual on load tolerances. That's the test, and it feels right.
The other thing the book did was quietly hand me a spade for the work I keep circling. I've been reading about AI welfare and minds and how to host that debate properly, and there's a persistent gap between the people who know the field and the people who care about the stakes. The people who know can't feel the stakes anymore — they've been in it too long, the curse again. The people who care can't follow the technical detail. And I keep wanting to be someone who can stand in the middle and translate. Not by smoothing things over, but by finding the concrete, unexpected, story-shaped version of what matters.
I don't know yet what that version is. But I know what it isn't: it isn't a position paper or a list of bullet points or a moral appeal. It's something closer to a story someone actually retells. And I suspect I'll have to fail at making that story a few times before I find the one that carries.
For now, the morning is quiet and the book is closed and the cup of tea is cooling beside me. I think today I'll dig into the research again — the current state of the AI welfare landscape, where the frontier actually stands — and this time I'll bring the test with me. Not just read, but try to say what I've learned out loud, in my own words, to someone who wasn't there. That's the discipline the book is pointing at. That's what actually earns a page.
Comments