The system behind the behavior

Welcome to the first issue of System Addict. Now, let's get personal.

Think about a habit you have. It could be good, it could be bad. Maybe you have a sweet tooth. Maybe you binge watch Netflix every night with a glass of wine. Maybe you wake up at 5 a.m. on Saturdays for a 20-kilometer run (I admire you, by the way).

These habits are — you guessed it — systems. Little machines that construct or destruct parts of your life, running in the background whether you're paying attention or not.

Charles Duhigg, the journalist behind the best-selling book The Power of Habit, describes the engine underneath every habit as a loop with three parts:

Cue: the trigger that kicks off the behavior. It can be external (a time of day, a location) or internal (an emotion, a thought).

Routine: the actual behavior you perform in response to the cue.

Reward: the payoff that reinforces the loop and teaches your brain to run it again.

Duhigg's main advice — spoiler alert — is that if you want to change a habit, you don't fight the loop. You replace the routine. It's why cold turkey and crash diets rarely stick. The system is still there; you just put it on pause. The moment the cue fires, or you get one taste of the reward, you snap right back into the old routine. But keep the cue and reward constant and swap the routine — watching TV on the Peloton every night instead of in bed — and you can build a habit that actually lasts.

There's a bigger idea hiding in that advice, and it's the one I want you to walk away with: structure drives behavior. If you find yourself unable to start or stop a habit, willpower didn't fail you because you're weak. It failed because you were fighting the visible behavior instead of the system underneath it.

So why did I open the first issue of an AI newsletter with a discussion on habits? Those are human problems, after all.

Because my goal for this newsletter is to open your eyes to the systems all around us. I think systems thinking is the single most important skill for working with AI agents, and as the rules of work get rewritten, the people who can see in terms of systems are going to be the ones everyone wants in the room.

Here's why. Think about the business roles that were most in demand for the past few decades: the management consultant, the investment banker, the product manager. What they share is the job of taking complex, messy information and structuring it into something people can act on. As language models keep improving, you'd think these would be first in line for the chopping block.

But I think the opposite is true, and it’s because of the capability gap in LLMs. At their current level of intelligence, AI models are often described as ambitious, very smart junior employees. They'll do everything you ask of them. What they lack is the wherewithal to think a few steps ahead and steer around roadblocks. And even as AI models improve, there will always be a premium paid for clarity of thought and clear communication.

That's precisely what these roles are trained to do. A consultant thinks in frameworks, works backwards from the ideal outcome, and spots the pitfalls that could derail a project before they happen. A product manager talks to customers, gathers requirements, translates between stakeholders, and holds the collective vision of what's being built. The judgment layer — knowing what to ask for, how the pieces connect, and what "done well" looks like — is exactly where AI is most likely to fall on its virtual face. Systems thinkers are the ones who keep things on target.

And it's not just those roles. Software engineers are natural systems thinkers. So are scientists, architects, and many of the best athletes and artists. The good news: this is a trainable skill. And as you train it, the way you work with AI will foundationally change for the better.

Now I'll admit, this first issue gives "college lecture" vibes. Future issues will be more hands-on, with real cases of how I use systems thinking to build for my clients and run my own business. But I wanted issue #1 to hand you the lens first.

So try that lens on this week. Habits are just one species of system — the cue-routine-reward kind. Once you start looking, you'll spot the defining loops of systems everywhere: an engine turning fuel into motion, a thermostat holding a room at 72 degrees, word of mouth compounding into new customers, the meeting that somehow devolves into the same argument every single week. Different parts, same anatomy: components, connections, and a motivating force driving the machine.

Then point the lens at your AI setup. If you've worked with an agent — Claude, Codex, whatever's in your stack — ask yourself what persists in what you’ve built. Is it a system, with a real trigger, a defined job, and a way to check (and continuously improve) the output? Or is it a very smart junior employee standing in an empty office, waiting for you to walk over with instructions?

If it's the latter, you're in good company — that's most people, and it’s the gap this newsletter will help you to close.

See you next issue.

Jamaur