Twelve Months, And The Argument Is Finished

I was in the building at Lehman Brothers for the 2003 blackout. Five years later the firm was gone and the same name was still on the door. Nobody inside either event called it a turning point while it was happening. Month one of the plan is one sentence long. The last eighty years are not the baseline. The probability that the next eighty resemble them is closer to 10% than to 90%, and anyone pricing that at zero is preparing for nothing.

In January I handed a machine one of the briefs I had been paid to write for twenty years. It came back in four seconds. Not a rough draft with a placeholder in the middle. A finished product, better than the version I would have charged for. The numbers were already public. ChatGPT reached 100 million users in two months. Legal analysis was matched at the 90th percentile on the bar exam. Financial analysis that takes a consulting team a fortnight came back in eleven minutes. I am a translator between domains. That is what the last twenty-five years of work has been, and translation is the first thing the machines took. Every exit I could name had a model standing in front of it. The window to reposition is measured in months, not decades.

There are 338 ready-made tools an agent can reach through a single protocol. In my own room there are 14 machines running roughly 50 agents between them. They work while I sleep, and the electricity bill is the staffing cost. The industrial revolution took sixty years to restructure the economy. This one is doing it in six. One person directing a hundred specialised agents can produce the output of a company, which means the barrier between an idea and a company has collapsed to an afternoon. By the time this is obvious, all the positions are taken. Staring at four screens watching machines work is not laziness. It is the job.

Month four is build something, anything, now. The tooling is free, the distribution is free, and everyone waiting for permission is watching other people ship. Month five is the correction nobody enjoys: you are around 60% as irreplaceable as you believe, and the same test that says so also tells you what to build. Month six is the body. Sleep is when the brain cleans itself, and no strategy survives a system that cannot physically climb. Month seven is the 5pm to 1am window, the eight hours a day where most futures are decided by scrolling rather than building. Month eight is arithmetic, and it is the month that pays for all the others. I signed a EUR 20 million personal guarantee and watched the company fold on Christmas Day 2008. I lost savings in the 2013 Cyprus bail-in. Before either, I restructured a EUR 75 million industrial group and the only thing that saved it was refusing to forecast the quarter everything stopped. Do not predict. Calculate. The person with the lowest fixed costs holds the most optionality, and optionality is the only asset that pays out in every version of the next twelve months.

A plan is not a prediction. It is a position taken before the evidence arrives.

I have spent eight years learning this one slowly. The five people you speak to most decide whether you are still standing after the shock lands. Nobody rescues a stranger, and no amount of capital replaces a phone that gets answered at 2am. I climbed the conveyor belt. I bought the car in Sardinia, sat at the long table, and looked around the room at people who did not like each other. I was 65 on paper and empty in the chest, and the only thing that filled it was work that was useful to somebody else. Months nine to eleven are the ones the first eight months pretend are optional. They are not. Survival without a reason is hoarding, and hoarding ends the moment the currency changes.

Twelve months is not a calendar. It is an order of operations, and the order is the argument. Accept the rupture, price the machines, direct them, build while the window is open, harden the body, cut the fixed costs, hold the people, and name the thing you are surviving for. I have deployed 210 energy systems across Africa and Asia, built companies across twelve countries, and taken one through a Frankfurt IPO. None of that was decided by certainty. It came from moving before the evidence arrived, in every single case. The probability that the next decade is easier than the last one is not zero, but it is nowhere near what people are paying for it. What you do with that number in the next twelve months is the whole of it. Make it count.