Grit, commitment, dedication, consistency

Dostoevsky's Notes from the Underground was the first book that made me uncomfortable in a way I couldn't argue my way out of. Its narrator is brilliant, self-aware, and completely paralysed — proof that intelligence without action is just an elaborate way of standing still. I think about that whenever I catch myself researching instead of building.

Crime and Punishment is about what happens when you decide the rules apply to other people. Raskolnikov reasons his way, step by careful step, into something monstrous — and every step is logical. It's the most persuasive argument I know that clean reasoning from a bad premise gets you somewhere terrible faster than muddled thinking ever could.

Solzhenitsyn's The Gulag Archipelago is the heaviest book I've read and I don't think I'll ever be done with it. Its line about the line between good and evil running through every human heart, rather than between people, permanently ruined my ability to think of myself as the reliably good party in a disagreement.

Lanza's Biocentrism and Penrose's The Emperor's New Mind pull in opposite directions and I keep both. Biocentrism makes the case that consciousness isn't a footnote to physics; Penrose argues that mind isn't computation, that there's something in human understanding no algorithm reaches. I work in AI. Holding a serious argument for the limits of computation next to the thing I build with every day keeps me honest about what these systems are and aren't.

And the Bible, which I return to differently than the others — not as an argument to evaluate but as the ground I stand on when the arguments run out.

Jesus, first and above the rest — not as a historical figure to admire but as the person I'm trying to follow. Everything below is a distant second.

Tesla, for the discipline of building the whole thing in his head before touching material, and for the willingness to be right in public while everyone else was wrong. Da Vinci, for refusing the premise that you have to choose — anatomy and engineering and painting were one investigation to him, and I've never believed the specialisation argument since.

Michael Jackson, for a reason people find odd until they look at the footage: the rehearsal. The genius everyone talks about was preceded by an obsessiveness about repetition that's genuinely uncomfortable to watch. He'd run the same eight bars until it was involuntary. That's not talent, that's consistency wearing talent's clothes.

And Dostoevsky, for looking directly at the worst of what people are and refusing to either flinch or despair.

Own the whole stack. I want to know what the database is doing and what the pixel is doing, because the bugs that matter live between layers and nobody who owns only one layer can see them.

Ship it. A system that isn't in front of users is a hypothesis, not a product. Zoomaa, Zoomaa Partner, MigoMap and AuraGo are all in app stores because that's the only version of the work that counts.

Design for the failure, not the demo. No network, simultaneous writes, a model that's confident and wrong — these aren't edge cases, they're Tuesday.

Be honest about numbers. Every figure on this site names where it came from. Where a measurement has a weakness — a metric I labelled myself, a rubric that's inherently subjective — I'd rather say so than let it look sturdier than it is.