All Newsletters

On Consistency, Availability, Partition Tolerance — and the constraint Eric Brewer never named.

The story everyone's missing while they argue about AI.

The biggest unsolved problem in AI product design isn't hallucination. It's miscalibrated trust.

AI didn't make technical skills irrelevant. It made the half-life of any specific skill shorter than it has ever been.

This isn't a compliance checkbox. It's a product architecture conversation.

We've spent a decade automating how we hire. We've barely started on what we do after.

Not a framework I read somewhere. The actual sequence, built from real engagements.

And almost nobody in tech is taking it seriously yet. Let me give you a number.

I've been in rooms in Seoul where the question was: "How do we enter India?" And I've been in rooms in Delhi where the question was: "Why isn't this product working here?" Same company. Different years.

And that acceleration is costing companies more than they're willing to admit. There's a conversation I've had at least a dozen times in the last two years.

The seduction of connectivity There is a moment in every platform build where someone opens a browser tab, finds an API documentation page, and says the most dangerous four words in enterprise technology: "We can just integrate." It feels like progress.

The question every professional is asking wrong Most people asking "how do I show AI skills on my resume?" are solving the wrong problem. They are thinking about keywords.

The question nobody asks before they ship Every AI conversation I sit in eventually arrives at the same moment. Someone shares the accuracy metrics.

For a long time, this was celebrated as the mantra of startup innovation. But as organisations grow, the very mindset that fuelled innovation starts creating fragility.

Most leaders talk about trust as if it were a personality trait. But in high-performing organisations, trust is rarely accidental — it is engineered.