One of the things I was happiest about at a recent internal hackathon was something I didn't do. I didn't put my screen up on the monitor.
That probably needs some explaining.
My natural instinct in these situations is to get involved. If there is a problem to solve and I think I can help, I want to get into it. I like making things. I like understanding how they work. And if a group of people are gathered around trying to make something happen, it would be quite normal for me to end up with my laptop connected to the big screen, driving.
This time I made a conscious effort not to. Other people took the lead. They worked through things, got stuck, tried things, talked to each other and gradually turned an idea into the foundations of something useful.
I was involved. I wasn't absent or pretending not to know anything. But I wasn't at the front of the room controlling the work. And it felt good.
I'm not going to turn declining to use an HDMI cable into a leadership framework. What interested me was noticing how deliberate the decision had been. Because doing it myself would have been easy. Probably easier.
That is something I've been thinking about quite a lot recently when people talk about bringing people along. There is a tension in leadership between getting something done and allowing enough room for somebody else to become capable of doing it. The frustrating thing is that both can look like good leadership.
Sometimes people genuinely need help. Sometimes the quickest way through a problem is to use the experience in the room. Sometimes standing back while everybody struggles would be pointless theatre.
But there are other times when I'll just do it is really about me. I know how to do this. I can make this go faster. I enjoy this bit. Give me the keyboard.
AI has made that instinct much more interesting. I've found that knowing a little about a lot has become increasingly useful. I can move into areas where I understand enough of the problem to ask reasonable questions, and AI can help me get much further than I could have done alone.
That's enormously powerful. It also means I can interfere in far more things.
For someone like me, who already wants to jump in, coding agents, research tools and AI make that easier. A gap in my knowledge that might once have forced me onto Experts Exchange can sometimes be crossed in an evening.
There is a lot to like about that. I am building things now that I simply would not have attempted a few years ago.
But personal capability and team capability are not the same thing. That was what I noticed at the hackathon.
If I'd taken over, perhaps we'd have got somewhere slightly faster. Perhaps not. What I'm fairly certain of is that it would have changed what everybody else was doing.
Instead, people had to work things out. There was collaboration because collaboration was actually required. People learned because there were things they genuinely needed to learn. And the emerging product belonged to the people building it rather than simply being something demonstrated to them.
I have heard slow down to speed up enough times for it to risk becoming management wallpaper, but this felt like a practical version of it. The slower bit is tolerating the fact that somebody else will not necessarily approach the problem as you would.
They may take a route you wouldn't take. They may need longer. They may ask questions whose answers seem obvious to you. And, inconveniently, they may also do something better. That's easy to forget when experience gives you a strong view of what should happen next.
AI adds another complication because it makes the boundaries around expertise much less obvious. I don't think deep expertise suddenly stops mattering. If anything, using AI has made me more conscious of the value of knowing enough to recognise when something is wrong, incomplete or missing the point.
But knowing a little about a lot suddenly feels like a real strength when it's augmented with AI. Being able to connect technology, organisations, people, data, process and whatever else is involved in the problem can be incredibly useful when a tool can help with some of the depth on demand. That feels like a strength. It might also make restraint more important.
If I can participate credibly in more domains, there are more opportunities for me to become the person everything routes through. That's not what I want.
I'm starting to think part of leading people is creating more capability around you rather than continually demonstrating your own.
The hackathon was useful partly because something tangible came out of it. I like that. I am still a builder and I don't think leadership requires becoming somebody who stands at the edge applauding while never touching anything.
I'll put my screen up again. There will be times when that is exactly what is needed.
But I noticed how good it felt not to this time, and how much happened with somebody else at the front. I'm trying to remember that.