Writing this from my car, parked outside, while my son sleeps in the back. A fitting metaphor, maybe, for the first month of a year in government — making it work.

Here are a few things I’ve noticed.

The people are great

I know people say this. But I want to say it properly: the people on my team are smart, passionate, collaborative, and ambitious. I feel genuinely fortunate to be working with them. That’s not nothing.

The stereotypes are true

Government is messy and confusing. The formal org chart tells you very little about how things actually work. Personal networks matter enormously — far more than reporting lines. You need soft power to get things done, and building it takes time.

None of this is unique to government, of course. But the scale amplifies it.

The model is a bad fit for software

This is the one that I keep coming back to. The government model for funding and running policies is a terrible fit for funding and running software.

Software needs to be iterated. It needs ongoing investment. It changes in response to users. But the funding and governance model treats it like a capital project — a thing you design, build, and hand over.

The result is a system that focuses internally, on the Treasury as the primary “buyer,” rather than externally, on the users who actually need it to work.

On Trojan horses

There’s a well-worn strategy in government digital: use agile methodologies quietly, as a Trojan horse, to gradually transform the system from within.

I’m skeptical. It feels both unhealthy — you’re essentially hiding what you’re doing — and inefficiently slow given the urgency. Schools and colleges need better digital services now. The stakes are too high for gradual infiltration.

Overall

It’s brilliant. Genuinely. I find myself deeply frustrated some days, and deeply energised others. Sometimes both at once.

That’s probably a good sign.