There’s a concept in computer science called explore vs. exploit. Exploitation means using what you know to get reliable returns; exploration means trying new things at the cost of those returns. Most algorithms skew too hard toward exploit. Humans have also been known to do this – including me. The known path is comfortable.
My last job, I was deep in exploit mode. I had playbooks. I refined them. I ran them. I wrote them up into a book. But often I worried I wasn’t growing. Part of why I left – and chose the particular shape of what came next – was to return to a place of genuine learning. I had enough concreteness that I could exploit, but also space to explore.
So what does that actually look like? A few things I’ve been in the middle of this season:
Formal learning. I’m currently enrolled in LSE’s MBA Essentials Course. I’ve long felt under-equipped on business fundamentals. Every time I was asked for an opinion on an acquisition, I would panic that I didn’t know how to reason about it. Structured training has been helpful for me to prioritize time for learning – I did the Co-Active coaching training a few years ago, but that was immersive; a few days I could slot in when it made sense. MBA essentials is 8-10 hours a week for 10 weeks and that felt like a different ask entirely – not one I felt up for adding onto a more than full time job without organizational support. Prioritizing it for myself, finally filling in a gap I’ve known about for years, feels good. Some combination of learning and building confidence – I’m not sure how much of which yet – but so far, it’s interesting.
Getting back to the work. As you get up the org chart the job stops being the work and starts being the work that makes the work happen. It’s necessary and done right it’s impactful and I do enjoy it, but I missed the feeling of actually building. For the past few months I’ve been back to the code again, building tools, co-building a course platform – I pushed a fix to production this morning! In a time of huge change, being closer to the work feels important. I want to understand this shift deeply because it will be affecting our whole industry for some time to come.
Deep collaboration. The best periods of my career have always featured a great collaboration. But the higher up you go, the lonelier it gets. Coming back to a real, balanced, give and take collaboration feels like a glass of cold water on a hot day. I love it. We negotiate to our strengths, balance getting stuff done with having fun. I also learn a surprising amount just from being close to another person’s workflow.
Fractional CTO work. This is a different kind of leadership than I’ve done before. In some ways it’s like running a large org – you have to be deliberate about your time because it’s limited. The strategic work (my favourite) is more visible when every decision counts and resources are limited. In other ways it’s very different – more hands-on with the actual infrastructure decisions, closer to an IT function than I’ve ever been (not a strength for me, I’m working on it).
Social media. For me, Twitter was everything in one place – community, self-promotion, ambient industry conversation – not that I had to be so clear about that. It worked so well for me and I missed it for a long time, without doing anything about figuring out how to fill the holes it left. I miss the community most, and I’m slowly trying to find it again. Given everything I’m doing, I need to figure out self promotion that doesn’t give me the ick.
AI workflows. Having space to actually play with this – without the constraints and overhead of an organization – has been revelatory. What works, what doesn’t, where it helps and where it’s just noise. The context matters enormously for perceived impact: if you use AI to do a mid version of someone else’s job inside an org, it’s insulting and will cost you in credibility. If you use it to do a mid version of something you otherwise couldn’t do at all, it gets something done and helps you learn. Jean and I are building a course platform, running cohorts, developing new material… with plenty of other things going on for both of us. We’ve used AI to create efficiency and leverage, to build something we couldn’t otherwise have built, while retaining the time and attention to make sure the pieces that should be human are.
I do really love running bigger teams. But you end up constrained – by org structure, by politics, by the cost of starting anything new, the difficulty of changing direction when you’re already moving, the morale hit of killing something even when it’s the right call. In a time of this much change, the number of people becomes its own constraint; change management could easily be the whole job right now, and not an enjoyable one.
It’s good to be free for a while. A hill-climbing algorithm will always find a peak – but it might not be the highest one. Sometimes you have to go down before you can go higher. I don’t think careers have to go up and to the right, and this season is making that clearer every day.
If everything is changing, you may as well choose the change you want.
