All posts

Book notes - Death to the Org Chart

Luke Girgis makes the case for rethinking every role from the ground up and hiring only for what can't be automated. It's written for the C-suite, but I took a lot from it, and it's a 90 minute listen.

In one line: rethink individual roles from the ground up, and hire for what can't be automated, so you stay lean and move fast.

Why I picked it up

I've known Luke since we were teenagers, back when he was a performing hip hop artist. Life took us in different directions. I moved to the US, and Luke traded performing for the corporate world. We've stayed connected on LinkedIn and other platforms, though, and I've followed his career from a distance. I'm continually impressed by his success and the way he leads.

So when I saw he'd published a book, I picked it up partly out of loyalty. I wasn't expecting it to land as squarely as it did. It turned out to be one of the most relevant things I've read for the work I'm doing right now.

The big idea

Most companies grow by adding boxes to the org chart. A new problem shows up, so you hire for it, and over time the structure becomes the thing you manage instead of the work.

The book flips that around. Rather than asking "who do we need to hire?", it asks you to start from the work itself. What does this role actually do day to day? Which parts of it could be automated today? What's left once you take those away? The answer to that last question is what you hire for: the judgment, relationships and creativity that tools can't yet replace.

Done well, the result is a smaller team that moves faster, carries less overhead and spends its time on the work that matters most.

What stuck with me

Push AI tooling to its limits

It's easy to treat AI as a side experiment: a few people trying a chatbot, a pilot that never quite becomes how the team works. The message I took away is that this undersells it badly. The real gains come from pushing the tools as far as they'll go and finding out where the limits actually are, rather than where we assume they are.

For an engineering team, that's a useful challenge. It isn't about using AI because it's fashionable. It's about being honest about which parts of our work still need a person, and which parts we keep doing by hand out of habit.

Optimize what you can now

The second idea is about speed and ownership. It's tempting to wait for the right vendor, the right platform or the next big project before improving how a team works. The book makes a strong case for acting on what's within reach today, and for questioning how much we rely on traditional services and external dependencies.

Every dependency is something you don't control and something that can slow you down. Bringing more of that capability in-house, now that the tooling makes it practical, is a big part of how a small team keeps pace with much larger ones.

Who should read it

The book is aimed squarely at C-suite executives, and founders will get the most out of it. But I took a lot from it as an engineering leader, and I think anyone who shapes how a team is structured, or who decides what to build versus buy, will find something useful here.

It's also short. The audiobook runs about 90 minutes, which makes it an easy listen on a commute or a long walk, and one I expect to come back to.