By Tom Watson in weeknote —

Weeknote 30

Meeting people off the internet, kicking off a new project, doing some building, thinking out loud on the North East and Public Interest AI, running through bogs

Weeknote 30
High force absolutely booming! A photo from my run this week. It were boggy and lovely.
Do you ever doubt yourself or you know feel like you don't know what you are doing?

I was asked this this week and my simple response is...

All. The. Fucking. Time

Doubt is natural, healthy even (when it doesn't become crippling). Doubt helps us to question. Doubt is part of learning. It's finding the tools and methods to make use of that doubt, embracing uncertainty, and find your way through. It's not always easy. Find your methods, your people, your rebellious streak, your running through bogs, whatever it is that works for you.

Also, no, I don't always remember my own advice, but I try.

What I did

  • Got to a draft of a data and AI strategy for an organisation I’ve been working with. Always nice, after lots of discovery and sensing to get to something real. Something to point at and say, is it this? If it’s not what is it? A bit of scenario testing to go along with the draft…does it help with x decision? If y happens how does this help, how would we respond, adapt? When the board says we should 'do AI' here's the strategy that says how and why we probably shouldn't in this case. I thought a lot about Open Strategy as we got to this stage. There is a lot of talk around building skills within AI at the moment, but what about building strategic capabilities? I think Open Strategy and some of the things I'm working on with Collective Resilience could help here. 
  • Kicked off a new project with Sheffield Futures as they embark on a three year data transformation project. We’ll properly start in January with the project but it was important to speak to a member of staff who is leaving at the end of the year. 
  • Met with three people who I feel I sort of know from social media or other networks. Been trying to do this more recently.
  • Firstly met with Iona. Heard about her journey to arrive at The Decelerator. I’m a big fan of the approach and I’m hoping to bring some of the thinking, insight and practical approaches from The Decelerator team into the Organisational Resilience programme next year. 
  • Met with Nicola as we talked about evaluation and social value. I’ll admit to sometimes being frustrated with the evaluation world, too strict when we need flexibility, too delayed when we need the ability to adapt, too focused on proof and metrics. Don’t get me wrong, evaluation is useful, but I think we need different approaches for different things. So I really appreciate Nicola’s view on evaluation as an evaluator, her focus on social justice especially. One of the reasons I wanted to talk to Nicola was to get some views as I focus on developing the Signals app. Her insight was really helpful, and I think now having had a few chats around this I’m in a good place for some development and then initial launch early next year.
  • Also met Jamie, in actual real life, as we went for coffee. Jamie does some excellent writing and thinking about the North East tech sector (you should read his stuff) and now was an especially interesting time as he has just started a secondment to the North East Combined Authority as a Technology and AI Advisor. I’ll be interested to see where he gets to with it, but more importantly, what he learns from it.
  • I spent a big chunk of the week making progress with the Collective Resilience guide, and the People pillar workbook for the Organisational Resilience programme. This involved really fleshing out my ‘5 modes’ framing (which I possibly bored Jamie with!). I intended to get the people pillar finished this week, but I’m not quite there. Next week. 
  • I found a couple of free evenings to do some building. I began building a research api end-point for The List which will allow research and trends analysis. I also prototyped a dashboard using this end point. There are now close to 500 changes on the list, so we are starting to really be able to do some useful analysis. 

I also spent some time on Open Recommendations

  • I added some new models to try for both chat and analysis and now allow users to select which one they want to use. There are now around 12 different models to choose from, some large, some smaller, some open, some closed, some greener. While I’d purposely built in flexibility to model usage from the start, I wanted to extend this flexibility. I want to see which models people use, what the performance is like, and also begin to understand the token usage/value created, which will, I hope, allow me to add in some environmental indicators for the site.
  • I also added both a bulk upload feature allowing users to upload up to 10 sources at a time. This meant adding a new end point and polling.
  • and I added a ‘private’ section allowing users to upload sources to a private area, and give them the choice to search and chat with either private sources, public sources, or a mix. 
  • Still on my todo list, is bringing in streaming and chained responses and really improving the whole interaction with sources.

What I thought about

Following the chat with Jamie I sat in a coffee shop (fueled by the most incredible hazelnut croissant - pink lane bakery obviously) and wondered what I'd like to see in the North East when it came to tech and AI. Now, of course, I recognise have a slightly different stance on this than probably most people, but still I allowed myself to imagine the North East as a global leader in Public Interest AI & Tech. There had been a couple of pieces I read this week about regional ecosystems thinking too small. And so I wondered - What if the North East was on par with Barcelona, Amsterdam, Tallinn - not just as a tech hub, but as the place where democracies come to build the AI that serves their citizens? What if we strategically focused on Public Interest Technology and AI as our distinctive proposition?

The ingredients are already here. The region has just secured a £30 billion AI Growth Zone. There are companies here that are already leaders in infrastructure technology, like Sage. There is Atom Bank, reimagining digital banking. We have world-class data facilities at National Innovation Centre for Data. Then there are emerging companies like Literal Labs who are helping to build and train the worlds most efficient AI models and the UK's largest data centre infrastructure emerging in Blyth and Cobalt Park. We have lower costs than London, collaborative culture, and maybe, genuine hunger to prove ourselves.

But here's what perhaps we haven't fully grasped yet: infrastructure, companies, and skills aren't enough. You need a coordinated strategy about what you're building toward.

Barcelona didn't just become a tech hub. It became the technological sovereignty capital, the place that showed cities how to take control of their digital infrastructure, how to require open source, how to put citizens before corporations. Amsterdam became the digital rights leader, the place that proved you could have smart cities without surveillance capitalism. 

Each had a clear answer to: "What are we the world's best at?"

So what could ours be?

What if the North East could be where the UK and the world builds Public Interest AI?

Not AI for advertising optimisation. Not AI for surveillance. Not AI for extracting value from users' data. But AI for healthcare that serves patients. AI for education that reduces inequality. AI for government services that work for citizens. AI for democratic participation. AI that embeds equity, transparency, and human rights from the ground up.

But there’s no money in that is there?

I mean if you think that you clearly haven’t been paying attention to the land grabs by the larger AI companies. This is perhaps a £170+ billion market opportunity. UK public services spend £26 billion annually on technology, with 70-85% wasted on legacy systems. The EU has allocated €61 billion specifically for e-government transformation. Both desperately need alternatives to extractive commercial tech. This isn't just about doing good for the sake of it, it's the biggest under-served market in technology - if you want high growth companies maybe this is where we focus. And this isn't to say that this is all we focus on at the exclusion of anything else. But maybe having a defining goal would move us away from hoping for a unicorn to drag us along?

And this approach might actual realise what the UK government calls "digital sovereignty" - building UK-owned, UK-controlled technology for critical national infrastructure.

But as Jamie wrote in this week's tech digest: "The North East is at a moment where AI doesn't just present a technology shift for us, I see this as a whole systems shift for us. I'm talking skills, culture, processes, incentives, infrastructure, and behaviours, all evolving (at pace!) together."

He's right. And more than that it requires a different way of thinking about the problem itself.

If we're not careful we'll just throw some solutions at a wall, accelerator here, funding pot there, skills program over there. But we haven't done the hard work of "what is the real problem?" Why are we building an AI ecosystem? What purpose does it serve? Who does it serve? What makes it resilient when markets shift and technologies change?

Right now, we're trying to ACT without having properly done SENSE, IMAGINE nor DESIGN. We're building infrastructure without clarity about what it's infrastructure for nor how we understand if/when we are doing the right things.

This is where Jamie's point about coordination becomes critical: "I firmly believe, as AI accelerates, the region doesn't just need 'more' activity. It needs coordination, connection and clarity. For a long time I've tried to ensure the North East's tech ecosystem is not just active, but aligned, accessible, and genuinely greater than the sum of its parts."

The question isn't whether we can build AI infrastructure. We're already doing that. The question is: will we build just another tech cluster competing on cost and hoping for anchor companies? Or could we build something genuinely differentiated - the place where technology serves democracy, citizens, and the common good?

And hey look, maybe this is too ambitious, maybe we can't compete with that other region which does this...which is where exactly?

Interesting things