Blog

Fred Brooks published The Mythical Man-Month fifty years ago, and we still manage projects as if he never wrote a word. Men and months aren't interchangeable, there's no silver bullet, and adding people to a late project makes it later. In the age of AI agents — the newest silver bullet — the old book reads less like history and more like a warning.

AI made generating things almost free. It did nothing to the harder skill underneath: knowing which thing is worth making, spotting when the confident output is quietly wrong, and having the taste to tell good from plausible. That's not a soft skill anymore — it's the whole job.

When a machine can generate the code, the copy, and the first three designs for almost nothing, the value moves to the three things it can't do for you: verifying the output is right, having the taste to tell good from plausible, and the judgement to know what was worth building at all. And underneath all three sits the one asset AI made scarcer, not cheaper — context. Output evaporates; context is the gold.

Projects are transient; capabilities are what's left when the project ends. Before you plan the push, draw the capability map — it's the difference between navigating and just moving.

Being stuck isn't a lack of effort — it's usually a lack of a different angle. A handful of thinking tools I reach for when a problem won't move: borrowed, battle-tested, and deliberately oblique.

A project with no structure isn't free — it's just unaccountable. Here's a lightweight skeleton, from idea brief to go-live, that gives the work a shape without drowning it in ceremony.

Effort isn't a virtue you maximise. It's a resource you spend. The hardest discipline in creative work isn't pushing harder — it's knowing the moment to stop pushing and ship.

Hope is not a risk strategy. The teams that ship hard things aren't braver than everyone else — they've just made naming and citing risk a continuous, boring habit instead of a heroic one-off.

Projects rarely die from a lack of talent or effort. They die from fog — unclear language, hidden assumptions, and the quiet bias that makes us see what we expected to see. Clarity is a discipline, not a personality trait.

Most disagreements in a project aren't disagreements at all — they're two people guessing about something neither has checked. The cheapest tool in the manifesto is also the one we use least: get up and go find out.

Milestones are for someone else's chart. Waypoints are for the people doing the walking. The difference sounds small and changes everything about how a team checks in.

An idea you can't test isn't a plan — it's a wish. The Hypothesis principle asks for something uncomfortable: state what would prove you wrong before you spend a cent proving you're right.

We treat being wrong as a failure of character. In creative work it's the mechanism by which anything gets found. The trick isn't avoiding wrong — it's getting wrong cheaply, early, and out loud.

Every good kitchen runs on mise en place — everything in its place before the heat goes on. Most projects skip it and wonder why service falls apart. Here's what 'prepped' actually means before you start a push.

Strip away the frameworks and the jargon and all of programming is three things: loops, sequences, and decisions. Everything else is abstraction from the truth — useful, until you forget it's there.

Your project is not the centre of anyone's universe but your own. The teams that create lasting value work from that humility — building with the people around them, not performing process at them.

Why call it a 'push' and not a sprint, a story, or a task? Because the word carries the thing the others lost: a unit of work with a real shape — and most importantly, a real end.

After three decades in technology, I've seen methodologies rise and fall. Most projects fail, but all journeys complete. Push is a philosophy that prioritises waypoints over milestones, shared value over ceremony, and the creative journey over forced timelines.