Milton Keynes marathon 2025
I was supposed to run the Brighton marathon last month, but bug between a charity’s CRM and the marathon meant I didn’t actually have a place. Lots of people very kindly sponsored me, so to make sure I delivered my side of the deal I ran Milton Keynes.
I hadn’t done a very convincing job of training for Brighton because of other commitments. I thought a month more might mean I could make up for that, but that was wishful thinking. The result was I didn’t really have a plan or a target time, so I just ran what felt Ok and didn’t push it too much.
That was liberating and had a nice run as a result. I’m also limping considerably less than after the London marathon last year. I still did 3:43, which is pretty convincing and only 10 minutes slower than London.
I got a Bolt to the start because the busses didn’t seem to be running. Bolt seems to have basically the same coverage as Uber now, so I’m choosing Europe.
Milton Keynes, as viewed from the satnav, doesn’t seem to exist. It’s liminal connective tissue, like corridors in a hotel laid, over the countryside. It does connect some very beautiful places though. I really liked running through all the under passes too, you could almost imagine the roads and the cars weren’t there.
Administrative Fairness Lab Conference 2025
I was in York for the Administrative Fairness Lab’s conference. Somehow I’ve never visited before.
The photo above is from York Minster’s stone mason yard, cathedral’s being things that are never done.
One of the sessions at the conference, by Maria Lee and Sam Guy was about due process in the planning system. One of the key ideas I took away was that removing things like consultations doesn’t remove the contested viewpoints they exist to air. Those still exist and some process are there as a safety value.
There was also lots of talk, as there is everywhere, about AI. But this was good talk, rather than the same old same old.
Jed Meers presented findings from prototyping chat bots and automated transcription that included more or less human intervention and how that might impact people’s sense of procedural fairness. Tom Tyler referenced similar tradeoffs and blended human-automated systems.
My dream interdisciplinary mashup at the moment is public sector designers, rule of law people researching administrative fairness, and psychologists researching mentalization/epistemic trust. Somewhere in there lies some of the answers on what decent public AI looks like.
Standing inside the machine
Wooden cogs in a windmill spinning fast with a staircase in the background
I was in the Netherlands to deliver a keynote at the User Needs First conference. While I was there I visited the Zaanse Schans open air museum. Windmills and other industrial buildings were transported and reconstructed there in the 60’s and 70’s. There’s also a museum of industry, mostly chocolate and biscuits.
Compared to most preserved windmills in the UK, many of the windmills at Zaanse Schans actually work and it was a very windy day. I’ve been in power stations and factories, but I’m not sure I’ve ever had the feeling of being inside a machine. Big, but you can still walk about it at human scale.
It’s just jogged a memory of The Mouse Mill episode of Bagpuss where the mice pretend to make a biscuit factory. The reveal being that there is only one biscuit and it just goes round and around. For some reason that really stuck with me as a kid.
I’ve been going all in on Obsidian the past few months. Normally open-source software that is endlessly configureable is a total mess, but with obsidian it feels less like configuring, more like being inside the machine.
Public services should work much harder for the public. However, the new UK government isn’t going to meet its aspirations for digital and data unless it resets the public sector’s approach to design.
The approach that grew from GOV.UK and the Government Digital Service was, if not flawed, at least incomplete at source. It ended up prioritising utilitarian simplicity, at the cost of designing government out of the way, and shunned technology as inconsequential to the design of services. The public have been engaged as consumers, not as part of a democratic society, in a way that fundamentally misunderstands the nature of what makes public services public.
Minimalism can’t scale to the types of services that do much more for the public, it can’t meet the demands of accountability. Trust in a public institution is a function of their ability to deliver the public good they are tasked with, and public services only get better if there is a public understanding of them. But designing government out of the way degrades the public image of government. Shunning technology means conversations about better use of data, remain stuck in the rut of ‘data sharing’ and critical infrastructure is too easily dismissed as solutionism.
A focus on utility leads us down a route to ‘personalised’ and ‘simplified’ services which, while OK for the minimal interactions that some may with the state, are useless at explaining the messy and demanding.
User needs, as a guiding principle was, it turned out, poorly suited to situations where the needs of government and the public were at odds with each government, or where people wanted more from government than a simple outcome.
Much of this was already obvious in 2013/14. Designing the digital account for the Universal Credit digital account, it was abundantly clear that the approach to design that worked for GOV.UK and was spreading across government was fundamentally unsuited to services that used automation, intentionally placed burdens on the public through policy choice, and used data from across government. As was the need for greater transparency and accountability. But as design practice spread across government, the focus on simplicity took on a life of its own, developing into what, at times, felt like a tyranny of design, where anything that distracted from the proximate user need was impossible to justify. The idea that digital public services needed to be more than transactional was lost.
Given there’s a UK election coming up, there’s going to talk about how to get more tech talent into government. Most is probably going to be encouraging ‘tours of duty’ from the tech industry. That short-term engagements are the way to get people with digital skills in to government. I thoughts I’d share my experience from 2011 on the precariousness that creates.
I originally joined the civil service on a 2 year fixed term as a civil servant. That was extended once (after a stressful period of uncertainty). Then when it should have rolled over into a full-term contract, I was told it wouldn’t. Just a letter in the post. No one anywhere in the civil service/gds could or would help. Nor could I transfer, despite trying. I was trying to move house at the time, in the middle of what turned out to be a series of miscarriages, and undergoing mental health treatment that was the result of the GDS transformation projects. At what felt like the last minute I was told it could maybe rollover if I did a sort of internal contracting role, but there was nothing concrete and the uncertainty of that had already taken its toll, so I quit.
I never wanted to be a contractor, I’d spent years before working on government things outside of government and played a not insignificant hand in several successes (GOV.UK, Universal Credit, Service Standard, GaaP, etc). I built built everything around the assumption of remaining a product manager in the civil service. But contracting was where I found myself having to absorb with covid, unpaid parental leave, several episodes of cancer in the family, deaths and a everything that life throws at you. Treating digital skills an exception, or something to be passed onto people already in management positions in a short bursts forces this kind of precariousness and costs on to people. If you want digital people in government, give them stable jobs and responsibilities.
I’m running the London Marathon in April. I’m doing it for Barnado’s who helped our family through the adoption process and beyond to give something back.
You can sponsor me here and I’ll match and sponsorship up to the target.
The aim is to get as close to 3 hours 30 minutes as I can.