Jonathan Stokes

Health inequalities researcher @UofGlasgow. Systems science, economic determinants, place-based policy. Leading OPTIMA project. Views my own

[Jonathan Stokes @DukeStokes]

Some communities have everything they need on their doorstep. Others have almost nothing.

We built a dashboard to make that visible, linking service provision and health needs at neighbourhood level across Great Britain.

Explore your neighbourhood 👇
🔗 https://t.co/9wGOdq1RjW

[Jonathan Stokes @DukeStokes]

@PreventionPodc @BrineHealth @JimBethell Prevention keeps being framed as a priority, but budgets and accountability still reward downstream activity. I’d be keen to hear what would make upstream, place-based prevention genuinely investable over 5–10 year+ horizons.

[Jonathan Stokes @DukeStokes]

The more I work on 20-minute neighbourhoods, the less I believe proximity is the point.

A GP, park or shop nearby only matters if people can actually use it, afford it, and feel it's for them. Equity is in the mechanism, not the map. What do you reckon?

[Jonathan Stokes @DukeStokes]

@DHSCgovuk "20,000 fewer A&E visits" is the easy sell. The harder question: do single patient records help the people with the messiest, most fragmented care, or just (maybe) smooth the journey for those who could already navigate it?

[Jonathan Stokes @DukeStokes]

The new @Spotify agent integration is awesome! I've wired mine into a daily paper-briefing: agent picks a paper from my Notion queue, drafts a personalised researcher-grade script, runs it through text-to-speech, and uploads to a personal podcast I listen to on the nursery run 🎶

[Jonathan Stokes @DukeStokes]

Everyone's reading today's local election map as party drama. Fine. But councils are where buses, housing, air quality, leisure and public health budgets actually live. If prevention is local, why do we still talk about it like Westminster theatre or NHS responsibility?

[Jonathan Stokes @DukeStokes]

We keep asking the NHS to fix what housing, transport, wages and planning keep breaking. Then we call prevention difficult. It isn't difficult; it's inconvenient because the levers sit outside the health system.

[Jonathan Stokes @DukeStokes]

“Ministers should force food firms to make their products healthier, introduce minimum unit pricing of alcohol in England, as Scotland has done, and tackle drug-related harm, Dixon urged.” - seems strange to use Scotland as exemplar here. Gap widening https://t.co/GTIkWGjbFc

[Jonathan Stokes @DukeStokes]

@openclaw (running GPT-5.5) still recommends Opus as orchestrator rather than switching to GPT as the go to…

[Dan ⚡️ @d4m1n]

GPT 5.5 benchmarks

BUT w/o hiding the ones where they lose to Opus 4.7 lol
also check vs Mythos

marketing bastards got me again

[Jonathan Stokes @DukeStokes]

@JujuliaGrace This is the right question to be asking. We've been building a dashboard that links service provision data with health needs at neighbourhood level. You can see where the gaps are. Lots of services affecting health outcomes are also privately run: https://t.co/BPnBXYzRiV

[Jonathan Stokes @DukeStokes]

@DHSCgovuk @UEL_News Really encouraging to see this model take shape. We've been working on something complementary: a dashboard that maps service provision against health needs at neighbourhood level.

Might be useful for identifying where these hubs are needed most: https://t.co/BPnBXYzRiV

[Jonathan Stokes @DukeStokes]

A lot of policy "evaluation" is, unfortunately, just politely documenting damage after the big decisions are already locked in. If public health only turns up once implementation starts, it’s not upstream, it’s clean-up. Too harsh?

[Jonathan Stokes @DukeStokes]

If your prevention strategy can be delivered entirely by the NHS, it probably isn’t prevention. It’s earlier treatment with better branding. Prevention lives in things like income, housing, transport and work. What am I missing?

[Jonathan Stokes @DukeStokes]

@petergyang Agree. You can get it performing a bit better with some back and forth, tell it old workspace was designed for Claude and ask it what would optimise for GPT. Seems to need to write itself a lot of automated scripts.

Still not nearly as good, I also switched back!

[Garry Tan @garrytan]

Using OpenClaw is basically is like driving your own Ferrari (that you have to be a mechanic for yourself) and it's broken down all the time, but gives you the time of your life

vs driving a reliable Honda (Hermes Agent)
vs riding the bus (Claude / ChatGPT)

[Jonathan Stokes @DukeStokes]

6) The flexible model selection and cron job automation is where it really shines. Things running in the background 24/7 that still no other tool can quite do right now. Outstanding when used alongside @AnthropicAI Claude Code, a tool @causalinf and others have been highlighting

[Jonathan Stokes @DukeStokes]

Anyone else using @OpenClaw as an academic/researcher? Inspired by @AlexFinn I bought a Mac Mini and set up Atlas: my always-on AI personal research assistant. It's saving me hours of admin every week. A brief thread on what it actually does in case useful to others too 🧵

[Jonathan Stokes @DukeStokes]

5) The boring admin stuff too: help with formatting of papers for different journals, pulling citation data, compiling papers for evidence summaries. The kind of work that used to eat entire afternoons