The machine at the front
A tunnel boring machine is a strange piece of engineering when you look at it properly. The part everyone pictures, the great rotating cutting head chewing through rock, is a fraction of the whole. Behind it stretches a hundred metres or more of train: the erectors placing concrete segments to line the tunnel, the conveyors hauling spoil back out, the crews keeping the whole assembly supplied, and above ground the trucks, the logistics, the people managing what comes out of the hole. The cutting head gets the photographs. The train is what makes the tunnel.
And here is the detail that matters. The machine cannot advance faster than the lining goes in behind it. That constraint isn’t bureaucracy or timidity. It’s the thing standing between a tunnel and a very long cave-in. The head breaks ground, the train makes the ground safe to have broken, and the speed of the whole operation is set by the slower of the two.
Entrepreneurial CEOs have always been tunnelling heads. The good ones know it. They want to break new ground, they’re impatient with anything that slows the cutting, and their instinct to push is usually the reason the business exists at all. The ones who succeed over time are the ones who understood that the pushing only works because of everything assembled behind them, and that the train imposes a real limit on how fast the head can turn.
What AI did to the cutting head
AI has made the cutting head dramatically faster, and it has done nothing comparable for the train. A CEO with these tools can now produce in a week what used to take a quarter. The strategy paper drafted in an afternoon. The new market analysis done overnight. The prototype standing up by Friday, the restructure modelled, the product idea turned into something you can click on. Ideas that once had a natural cooling-off period, imposed by the sheer effort of developing them, now arrive fully formed and keep arriving.
The rest of the business, though, still runs at the speed of people. The team still has to understand each initiative, staff it, fit it around the work they were already doing, and keep serving the customers who pay for all of this. Their capacity to absorb hasn’t changed. What’s changed is the rate at which absorption is being demanded of them, and the gap between the two is widening every month the CEO gets better with the tools.
Enough ideas to destroy the company
This problem is older than AI, which is worth knowing because it means the answer is too. Jeff Bezos tells a story about the early years of Amazon, when Jeff Wilke, who went on to run the company’s entire consumer business, pulled him aside with a warning: “Jeff, you have enough ideas to destroy Amazon.” Bezos, who says he can stand at a whiteboard and produce a hundred ideas in half an hour, found the thought shocking. Wilke’s point was a manufacturing one. Every idea released into the organisation is work dropped into a queue, and if the organisation can’t absorb it, the idea isn’t just worthless. It’s destructive. It creates backlog, work in process, distraction. You have to release the work, Wilke told him, at the rate the organisation can accept it.
What Bezos did next is the instructive part. He didn’t stop having ideas, and he didn’t apologise for the cutting head being fast. He started keeping a list, holding ideas back until the organisation was ready, and putting his energy into a different question: how do I build an organisation that can be ready for more? He worked on the train so the head could turn faster. If that discipline was necessary when the ideas came from one man at a whiteboard, consider what it’s worth now that the whiteboard talks back and drafts the implementation plan.
Are they with you, or are you digging alone
The failure, when it comes, doesn’t look dramatic from the front. The CEO is breaking ground daily and it feels like the business has never moved faster. But behind them, initiatives are stacking up half-finished. The team is three announcements behind and privately triaging which of the CEO’s priorities to quietly drop. Middle managers are absorbing the pace by shielding their people from it. Nothing is being lined. The tunnel is collapsing behind the machine at roughly the speed the machine is advancing, and the person at the head is the last to feel it, because at the head, all you can see is new ground.
So if AI has made you faster, and if you’re the kind of leader who reads pieces like this one, it probably has, the question worth sitting with isn’t how much faster you can go. It’s whether the pace you’re setting is true for the entire business. Is the team actually with you? Have you spent any of your new speed on them, on the tooling, the capability, the clarity about what to stop doing, that would let them absorb what you’re producing? Or are you a hundred metres out in front, in the dark, mistaking the sound of your own cutting head for progress? The ground you break only becomes a tunnel when someone lines it. Make sure the train is still attached.