Writing

AI's first 'I'm done' is almost never its best work. Simulated accountability pressure — just asking 'are you sure?' — surfaces blind spots that self-review misses.

I ran the same question through 6 AI tools and scored them against peer-reviewed evidence. Every tool got something wrong that another got right.

Nobody practises arithmetic speed anymore. The same thing is happening to prose, research, and analysis — and it changes what humans should get good at.

Naval's famous line is easy to nod at. The hard part is actually identifying yours — and being honest about what isn't.

30 days of Oura data showed activity type doesn't predict stress. Meetings do.

From AppleScript's rigid English-like syntax to LLM tool-calling — what changes when the computer actually understands you.

Automation that collects without learning is just a cron job. The difference is a feedback signal — a number that goes up or down.

Karpathy's autoresearch and every useful AI tool share the same pattern: the code is trivial, the feedback loop is the product.

You need the tool to build the tool. The answer is: build the dumb version first, use it once, then have it build its replacement.

Every AI lab builds single-vendor orchestration. The cross-vendor layer is a gap — and it's a gap for a reason.

Multi-agent AI orchestration frameworks are commodity. The competitive advantage is knowing which agent to use when, what breaks, and how to recover.

When pattern-matching feels like wisdom, what are we actually experiencing?

Ethics isn't about knowing the answer — it's about feeling the tension

The question I can't fully answer for my son

When AI conversations feel insightful because the language model is good at producing insight-shaped text

The things that matter most in production are the things that get benchmarked least

When the paper becomes more interesting than the answer you set out to find

The missing layer in enterprise AI isn't smarter models — it's structured memory that humans can actually review.

Prompt instructions are suggestions. Hooks are constraints. One survives a model swap.

The 'ship and calibrate' loop works beautifully for reversible decisions. For the big ones, you're mostly guessing and then making the guess true.

AI commoditises execution. What remains is taste — the 'that's the one' reflex. And the only way to sharpen it is to ship and see what reality says back.

The bottleneck in a world of AI tools isn't crafting the output — it's knowing which output is worth crafting.

Most task systems are binary, but a third class exists — tasks triggered by external notifications — and managing them like a backlog item is the wrong move entirely.

AI agents that require you to go looking for their results aren't agents — they're automation with better UX. The loop closes when results arrive, not when you remember to check.

Adding more AI tools doesn't make you faster if you're still the junction between every agent step.

Having great AI delegation tools and not using them isn't a tool problem — it's a pattern recognition problem, and that distinction changes everything.

The execution layer of an AI system is only half the infrastructure — the reporting layer is what determines whether anyone acts on the results.

The gap between AI assistants and AI agents isn't about reasoning capability — it's about whether the thing can survive your laptop closing.

The gap between 'agentic AI' hype and what actually ships in production turns out to be a workflow — and that's a feature, not a failure.

An experiment with AI deliberation revealed something uncomfortable: accumulating confident opinions feels like convergence on truth, but isn't.