What an ottoman taught me about working hours
An ottoman put me in the hospital for 61 days. It was the best career advisor I've ever had.
An ottoman put me in the hospital for 61 days. It was the best career advisor I've ever had.
The last time I posted here, I said I had momentum I didn't want to waste. Then I disappeared for eight months. This is why.
In February I lifted that ottoman. Badly. All spine, no knees. The exact lift every manual-handling poster warns about. By the next morning I was in A&E with pain I still don't have adequate words for and legs that had stopped listening.
"For two months I couldn't walk. When a doctor asked whether the problem was pain or weakness, the most honest answer I could give was: "Both - I tell them to move, and they don't move.""
There was also a darkly comic subplot in which the MRI I urgently needed was blocked by an air-rifle pellet that has been lodged in my face since 2013. A prank gone wrong. Long story. Different post.
"Here is the thing about lying in a bed while your legs ignore instructions: you get a lot of time to audit your hours."
I used to work nights. Not occasionally. For months on end. All-nighters for clients, for deadlines, and for a part of me that has always needed to go above and beyond to prove something to someone I have never been able to name. Childhood, probably. The invoice never says who it's for.
I turned 40 the month after they let me out. Memento Mori - "Remember you will die" was a principle I already claimed to live by. It turns out there is a difference between believing something and having it demonstrated.
So I came home and rebuilt how I work instead of resuming it.
I now spend almost all day, every day, working with AI. My output is higher than it ever was in my all-nighter years, and I finish inside working hours. I did not know I could do that. I'm not sure I believed it was allowed.
This is not just my experience. Research from the Federal Reserve Bank of St. Louis found workers using generative AI report saving around two hours a week on average, and one in five report saving four or more. Those hours don't appear from nowhere. They're the nights.
Proof, if you want it: symbiometry.com. While a new team prepares to build our new site properly and in its entirety, I built the current one myself in the interim. AI made it faster and easier to get it to a standard I'm actually happy with. And for the first time in my life, I haven't asked a single person for their opinion on it, ok, maybe one or two. For a lifelong approval-seeker, that's character development.
"Which brings me to my first ask since coming back: if you can spare a click and a minute, have a look, and tell me what you honestly think. Comments open. I can take it. I've survived worse this year."
"And, if AI still makes you scared or nervous, that instinct is reasonable. Fear of new infrastructure is as old as infrastructure. But this is 1910, and the motorcar is not going back in the box. AI is not a tool you might pick up one day. It is infrastructure, like electricity and WiFi: invisible when it works and unarguable once it does."
One honest caveat, though, because this is not a productivity post.
AI can give you your hours back. It cannot tell you what they're for. That part is still yours. And mine. I'm still learning not to hand the recovered hours straight back to the part of me that needs to prove something.
"So, a question, from someone who had to be put flat on his back to ask it:"
Who taught you that the hours were the proof?
I'm writing more about all of this, the proving, the masking, and the stories we work ourselves ill inside, in a book called Reason Wins. But that's for another day. Today I'm just glad to be standing.
Michael "Milo" Lockett
Co-Founder and CTO of Symbiometry, fractional CTO and technical adviser. I write from practical experience across systems, event technology, interfaces, automation and complex delivery.
Based in Windermere, England. Working in event technology since 2013.
Related thoughts
Continue the thread
The number that sounds right
A tool that's right most of the time teaches you to stop checking.
Why AI Assistants Need Better Interfaces
An assistant becomes useful when people can see its context, state, assumptions, actions, and the point at which human approval takes over.
Automation Is Not About Removing People
Good automation removes repeated friction while making judgement, ownership, and the route to intervention clearer.
If this sounds like a problem inside your organisation, we can make it concrete.
Tell me what is happening now and what the organisation needs to be able to do next.