--- title: What are you willing to be wrong about? subtitle: Why cheap software made architecture and strategy the whole game status: draft type: paper created: 2026-08-25T00:00:00.000Z share_slug: willing-to-be-wrong tags: - thesis - strategy - architecture - ai-delivery - decision-making - uncertainty - pi-planning - real-options - leadership - paper - non-defensive-culture - directional-correctness - strategic-thinking - cost-dynamics - intentional-uncertainty - constraint-theory - resilience - competitive-advantage - organizational-learning - technology-economics - adaptive-planning related: - "[[editorial-voice-style-guide]]" description: Examines how AI-driven cost reduction in software development elevates architecture and strategy as primary competitive differentiators. mdb: index_tier: 2 version_comment: "Shared — public" --- # What are you willing to be wrong about? *Cheap software extends reach but brings strategic thinking centre stage.* ![](./figures/the-mapping-room-2.webp) You know this room. You've sat in it, on one side or the other. The MD and their business leads come in braced. The last few times this meeting wandered into detail that left them behind. Nobody enjoys nodding along to a language they don't speak. They hold the reins tight, steer for outcomes, and try not to get drawn into the plumbing. This is fair - an MD should know their business, not message queues or options for RAG stores. Across the table, the engineers are braced too. Barricades prepared in a dialect that's half explanation and half moat. And if the conversation turns to delivery slippage, because it often does, the dialect gets more dense. The reins tighten and trust erodes. Nobody's behaving unreasonably. The wariness on both sides is earned scar tissue, from real projects that really went wrong. Strategy too vague to align with, a platform with gaps and technical debt that makes change risky and difficult to estimate. Everyone's being sensible, and the sum of all that sensible behaviour can be a standoff – two parties defending their lane, braced against the other, and both right. Somewhere at the back of all this an entirely legitimate grievance - it's two-sided, fundamental and strategic. You cannot build on fog, and you cannot answer all the questions you need to make the fog clear. That room has existed in every company that ever built software. For most of that history it was survivable. Expensive in friction, tense once a quarter, but survivable. Things moved slowly enough that a defended standoff could sit for years without anyone quite paying for it. Things do not move slowly any more. ## The floor dropped Something changed under everyone's feet in the last eighteen months. The usual story is that AI raised the ceiling, that we can all reach higher now. That's not quite it. What actually happened is that the floor dropped, and the cost of *building* fell through it. A small company can now stand up a platform, wire in an integration, or spin up the kind of internal tooling in weeks that used to need its own line in the budget, and months or even years of delivery. This is genuinely good news. An SME can operate at a level that was simply out of reach a few years ago. But watch what that does to the room. When building was slow, the standoff was cushioned by pace. A fuzzy strategy could only misdirect so much work, because only so much work got done. A barricaded, over-cautious architecture could only ossify so fast. The defensiveness cost you, but gently, spread thin over years. Drop the floor and the cushion goes with it. Now a fuzzy strategy misdirects a quarter's worth of build in a fortnight. The architecture nobody agreed on hardens into a shape nobody chose, at five times the old speed. The business is funding velocity in a direction that was never actually declared, and the engineers are pouring that velocity into foundations without the strategy to pour straight. The same reasonable, defended standoff that was a manageable inefficiency for decades has quietly become the single most expensive thing in the business. The constraint used to be *can we build it*. That particular constraint is dissolving, only to be replaced by a new one, made of two things that a cheap, "vibey" fast build does not help with. ## You are going to be wrong Here are those two things. - One is whether you can keep moving without turning everything into a brittle, tangled mess you're frightened to touch. Call that **Architecture**. - The other is whether all that motion is pointed at growth, or sustainability. Call that **Business Strategy**. Both of them come down to decisions made in uncertainty, and where this caricature above highlights the problem. The instinct in that room, on both sides of the table, is *defensive*. The business hedges by not committing to a bet, engineering hedges by keeping everything loose, flexible, nothing nailed down. Both feel like prudence, in the slow world they nearly were - in the fast world they become a mess of built up debt. A business that will not commit to a bet builds a bit of everything and a lot of nothing. An architecture that commits to no boundaries is not flexible; it is a formless puddle, going wherever the ground tilts. The defensive posture that feels the safest is the one that guarantees you move fast in no particular direction. As an approach the phrase "luck not judgement" comes to mind. So here's the question. **You are going to be wrong.** ***What are you willing to be wrong about?*** Stop trying to be un-wrong. You can't be, not at this pace, and the trying is precisely what has both sides frozen in their crouch. The useful move isn't to eliminate wrongness. It's to *choose* it – to decide, on purpose, where you'll plant a stake and accept you might have planted it in the wrong spot, and where you'll deliberately keep your options open because the world hasn't told you enough yet. That's not recklessness, it's the opposite. It's the only form of care that survives contact with this much speed. ## Good news... Two disciplines particularly have matured their response to this uncertainty years ago. - The first from the way serious enterprise engineering plans. The core admission is blunt: **the plan will be wrong, don't describe the whole cathedral.** Understand your immediate needs. Have an architecture that you're aiming for and make your design directionally correct, but commit to the thing in front of you. Replan actively, on a cadence, as a scheduled and reassuring act not a crisis convention. - The second comes from finance, from the way risk is actually managed when the future genuinely cannot be known. The insight is interesting: **the decision you have not yet made has value.** Deferring a choice you don't have to make today isn't indecision; it's holding an option, and that's worth something precisely because the world is uncertain. It will teach you more if you wait - you decide what exactly you must decide now to stay pointed the right way. Both these disciplines have a thing in common, they're institutional ways of not being defensive about being wrong. One makes the wrong plan cheap and expected by revising it on a cadence, the other makes the unmade decision an asset rather than an embarrassment - but critically they both set a direction based on strategic thinking. So... turning all of this from some good words into something you can do on Monday. To be willing to be wrong about something *on purpose*, you have to have said what it is. A bet you are making and a bet you are declining to make, a boundary you're committing to and one you're leaving open. The first discipline is simply to **declare** it - not because declarations are sacred, but because a declaration is the only thing two people can stand over and point at, argue over and work towards. The declaration is where the MD's bets and the engineer's boundaries turn out to be one list, seen from two chairs. Then you **measure** it, so you can tell a bet that's wrong from one that's merely early, and **evolve** it out loud when the evidence lands, a dated decision rather than a silent drift, because a strategy that changes quietly was never really being held. ## The room, again Go back to the room - the same people, around the same table. What was missing was never talent, and rarely money – a business can buy talent and at these new prices, do an astonishing amount of engineering with very little money. What's missing is the conversation that has suddenly become the highest-leverage thing in the business and remains the rarest: the honest, undefended, regularly-held conversation between the person who owns the bet and the person who owns what the architecture must protect. You are going to be wrong. That was always true. The only thing that changed is the long term cost of pretending otherwise. So. What are you willing to be wrong about? ***@William**, September, 2026*