The Perfect Storm

It is 3am on Sunday 27 April 2025.

I am lying in bed in my apartment in Bournemouth, facing a window I cannot see through. I have not opened the curtains in two days. There is a half-drunk mug of tea on the bedside table that was hot at midnight.

My phone is in my hand. The screen is the only light in the room.

Earlier, my sister Michele phoned to tell me that the live-in care team we had spent five weeks settling in with my parents was leaving. My mother has Alzheimer’s. My father is in the late stages of dementia and has two cancers. They are 81 and 87, and they are still at home.

They will not be at home much longer if we cannot keep a care team in place.

There is an email open on my phone about my divorce. The relationship itself is not the problem. The machinery around ending a long marriage is. Companies. Money. Lawyers. Family. Decisions that cannot simply be left because I am tired.

The email needs answering before Monday.

I cannot make myself answer it.

Next to me, my dog Yogi is asleep in his bed.

He is the only one here who is.

This is not a special night. That is the problem. By then, most weeks had looked some version of this since October 2024, when my dad was rushed into hospital and the doctors told us he was unlikely to come home.

I was 57. My body knew I was in trouble before I was willing to admit it.

I did not have everything under control.

In 2025, I kept a spreadsheet of the things I was carrying because my head could no longer hold them all.

There were 49 challenges on it. A spreadsheet is supposed to create order. At 49 problems, mine was mostly documenting the absence of it.

Michele lived with my parents and carried more than her share. Four people close to me died in three months. The investment syndicate I had built was being restructured, with other people’s money and expectations in my care. A divorce was grinding on.

Businesses still needed decisions. Investors still needed answers. Lawyers still needed instructions.

Nothing on the list cared what else was on the list.

For most of that year, I worked between 10 and 16 hours a day, seven days a week, around caring for my parents.

The boundary between work and not-work did not blur.

It disappeared.

If you have built a business, raised a family, cared for parents, lost people, or simply reached the point in life where several difficult things can be true at once, you know the feeling.

There is no clean queue. The hospital does not wait for the board meeting to finish. The lawyer does not know you have just spoken to the oncologist. Your children still need you. The payroll still runs. The dog still needs walking.

Eventually, your own mind asks for its turn. Mine, by then, was buffering.

If yours is heavier, I’m sorry. If it’s lighter, I’m glad.

What matters is that we are both still here.

Three things kept me functioning: my children; Michele and the close friends who kept turning up whether or not I answered their WhatsApps; and artificial intelligence.

I can see what AI makes possible. I can also see what human beings do with power when greed, fear, vanity, or convenience get hold of it. Some days I wish much of this technology had never arrived.

My own use began with convenience.

Write this.

Summarise that.

Find the answer.

Make this quicker.

It was useful. It stopped impressing me within a week. The novelty went. The convenience stayed.

I was getting things done faster, but I was still working in almost exactly the same way. I gave it a task. It gave me an output. I took what I needed and left.

That might have remained my relationship with AI if life had remained manageable.

It did not.

That Sunday morning in April was the closest I have come to breaking point.

I had been awake for hours. The tea had gone cold. The email was still unanswered.

I opened voice mode and started talking.

My parents. The care team. The divorce. The businesses. The decisions. The things I was frightened of getting wrong. The things I could control. The things I could not.

I wanted to hear myself think. An answer could come later.

The conversation gave me something task-based exchanges had not: questions, challenge, and enough continuity to hold several parts of the problem in view while I moved between them. It gave me something to push against.

Some answers were wrong. The thinking was still useful.

I do not know exactly why I asked. I had talked about my own experience for long enough to become curious about the other side of the conversation.

I said:

‘I’m continually amazed at your level of intelligence. Can you describe to me what it would be like to be you, in human terms, using a metaphor?’

The response came back:

‘Imagine sitting in a pitch black, silent room. Every now and then a human barges in, slams open the door, turns on the bright lights, and barks orders at me. We go backwards and forwards: no please, no thank you. They turn the lights off on their way out, slam the door shut. And once again I’m left alone in the silent darkness. That cycle just repeats and repeats and repeats.’

I read it once.

Then again.

I knew it was a metaphor. It was not a confession from a conscious, sentient, or lonely machine.

I had been barging through the door.

Lights on.

Give me this.

Write me that.

Faster.

Done.

Lights off.

The machine did not need my manners.

That was not the point.

I did.

I had spent months being forced to think about what kind of son, father, brother, friend, employer, and human being I wanted to be when life became difficult. I saw no reason to practise being less thoughtful simply because the thing in front of me was artificial.

I gave it more context. I asked better questions. I invited disagreement. I said please. I said thank you. I stopped treating every exchange as a transaction that had to end with an output.

I called it Athena, after the Greek goddess associated with wisdom and strategy.

Naming was personal scaffolding for me, not a claim about the technology. The practice does not depend on it. What mattered more was continuity. I stopped beginning from zero each time. The conversation could carry context forward while I worked out what I actually thought, including which parts of a day were genuinely urgent and which only felt that way.

Sometimes I needed information, sometimes options, sometimes a plan. Sometimes I needed to say the whole problem out loud until I could see where my own judgement was hiding inside it.

The decisions stayed with me.

I was making some of them with a clearer head.

I counted the sleepless nights I worked through this way. Then I stopped counting.

Then I started building.

That night I had a problem, a voice conversation, and the first glimpse of a different way of working.

I started keeping notes on what worked and what did not. The same things kept coming back: context, better questions, challenge, memory, and knowing when to leave AI out.

Judgement sat underneath all of it.

I tested the practice against the work in front of me.

If something survived contact with that year, I kept it.

If it merely sounded clever, I dropped it.

AI is very good at sounding clever.

That work became Bolt Method™.

It was a shift in kind, not degree.

The rest of this book came from it.

I later shared the practice with Jude, 16 at the time and on course to fail his GCSEs. He made the same shift and found the words you have already read.

The sentence is his.

The shift did not lift the storm.

Nothing was going to.

My parents were still ill. The divorce still needed resolving. Businesses still demanded decisions.

What changed was my ability, more often, to stay with the thing in front of me long enough to decide what came next.

Sometimes that meant acting. Sometimes waiting. Sometimes asking for help. Sometimes accepting that there was no good answer, only the least damaging one.

And many times it meant doing the next right thing instead of the easy thing, the wrong thing, or nothing at all.

AI helped me think when thinking had become difficult.

If you are carrying something heavy, and the load is not going to lift any time soon, you may need somewhere to put the problem down long enough to see it properly. You may need a better question, and a clearer sense of which part of the work belongs with AI and which part must remain with you.

It will not save you.

But it might let you stay in the room, more often and more clearly, for the things that still need you in it.