Insights

Advice, thoughts and opinions on software, AI, industry and the work behind it.

The most impressive tool in web development is wasted on games

Product configurators, virtual tours, live dashboards, training sims: the tech that makes them feel alive came out of games. It runs in the browser today, and almost nobody outside games is using it.

Anatomy of a software project that quietly failed

Software projects rarely die in a dramatic crash. They drift, the champion leaves, the vendor goes quiet, and one day everyone admits it's dead. Here's how to see it coming.

A field guide to weasel words in software pitches

Software sales language is built to sound like a promise while committing to nothing. Here's what the favourites actually mean, and the question that turns each one into a fact.

"Let's build an app" is usually the wrong instinct

We build apps for a living, and we'll still talk you out of one when the app is the answer to a question nobody asked. Here's how to tell what you actually need.

Your backup is a hope until you've restored it

Most businesses have backups they've never once restored. The day you find out whether they work is the worst possible day to find out. Test it on a good day instead.

Per-seat software pricing is a tax on hiring

Per-seat pricing charges you more the more people you employ, whether or not they use the thing. Here's how to run the numbers and when building starts to win.

The demo always works. Break it before you sign.

Every demo is choreographed to win. The trick is to make the vendor drive through your worst records, not their clean ones, while you watch.

Custom software versus off-the-shelf, and what it really costs over time

A small monthly fee can cost more than a one-off build. What matters is the bill over five years, and who controls it.

How to prepare Firebase for burst traffic without bill shock

Firebase is great for lean realtime apps. The bill risk starts when reads, writes, functions and storage traffic are left open-ended.

Before you connect two systems, decide which one is right

Systems integration fails when two tools disagree and nobody has decided which one wins.