iiinova — where your digital self keeps evolving

Ben Johnson @ersatzben.com · 30/09/2026

London AI × Science Hackathon 👾 Science Works × Inherent Labs • Oct 16-18 Lots of scientific software is outdated, hard to use, and isn't interoperable – let alone making use of (or ready to interface with) AI. Or it's expensive. Join us to build better for the community

102

Ben Johnson @ersatzben.com · 20/07/2026

Sunday: an LLM disproves an 87-year-old mathematics conjecture, seemingly on the prompter's whim Monday: the UK govt dept responsible for thinking strategically about AI + research in the round is disbanded, seemingly over a petty civil service turf war. Good one. 🇬🇧🦶🔫

001

Ben Johnson @ersatzben.com · 23/06/2026

M’cr would almost certainly argue that they could do more with more powers (e.g. over skills, energy and rail) and more money - possibly a big increase in devolved R&D spending. There could be opportunities here…

100

Ben Johnson @ersatzben.com · 23/06/2026

“Manchesterism” is going national. What might a Burnham premiership mean for science? Devolved R&D? There are opportunities & risks. We’ve been thinking about this at @science.works – here’s a taster of our thinking… 🧵

111

Ben Johnson @ersatzben.com · 31/07/2025

Our analysis shows why public R&D is so vital. Knowledge spillovers mean private firms systematically underinvest in R&D. The evidence shows social returns from private R&D can be more than 4x higher than private returns, creating a clear rationale for public investment.

220

Ben Johnson @ersatzben.com · 31/07/2025

What’s inside? 🛠️ ➡️ In-depth, evidence-based profiles of 9 major R&D policy tools. ➡️ A comprehensive framework for evaluating returns on R&D investment. ➡️ An interactive calculator to generate benefit-cost ratios (BCRs) for R&D spending.

110

Ben Johnson @ersatzben.com · 31/07/2025

The challenge: Driving innovation is crucial for the UK’s growth and security, but policymakers face a complex choice from a wide array of policy tools. Our toolkit is designed to help navigate these critical decisions.

110

Ben Johnson @ersatzben.com · 21/07/2025

The narrow path ahead requires political will, new leadership, and sustained focus. But the prize is huge: a reformed ATI delivering decisive national advantage through sovereign AI capabilities. Turing cracked Enigma for Britain and her allies. His Institute should do the same for our time.

110

Ben Johnson @ersatzben.com · 21/07/2025

Success isn't guaranteed. Organisational reform is hard. We propose a 36-month implementation plan with clear phases and KPIs. Critically, EPSRC must use its funding leverage to enforce change, or risk policy failure.

110

Ben Johnson @ersatzben.com · 21/07/2025

Why now? AI is reshaping warfare globally. Ukrainian forces use AI for artillery. US DARPA trials AI-piloted F-16s. China tests swarming autonomous drones. The UK needs sovereign AI capabilities - not dependence on overseas commercial labs.

110

Ben Johnson @ersatzben.com · 21/07/2025

The SoS letter provides a clear mandate: transform ATI into a vital D&NS asset. This returns the Institute to Alan Turing's legacy. His work at Bletchley Park shortened WWII by 2-3 years. The Institute bearing his name should match that ambition and sense of national purpose.

110

Ben Johnson @ersatzben.com · 21/07/2025

The standout exception? Its defence & national security work - which has been supporting the UK defence and intelligence communities with behind-the-scenes breakthroughs. Ministers now want D&NS to be the main focus.

110

Ben Johnson @ersatzben.com · 21/07/2025

Since 2015, the Turing has received £200m+ in govt funding but struggled with mission drift, thinly spread research, and governance misaligned with national priorities. Research has visibly not kept pace with the AI frontier.

110

Ben Johnson @ersatzben.com · 21/07/2025

🚨 NEW REPORT 🚨 The Alan Turing Institute has visibly struggled, and is now at a critical juncture. The Secretary of State has intervened. What happens next? A new report for @britishprogress.bsky.social by me and @jujulemons.bsky.social sets out a plan: 🧵

178

Ben Johnson @ersatzben.com · 28/04/2025

In the essay, I argue that the AI microsite: - updates the grey literature tradition for a web-native world - uses minimalist design to preload seriousness and inevitability - blurs fact, forecast, and fiction into persuasive campaigns - subtly engineers memory, not just belief

102

Ben Johnson @ersatzben.com · 28/04/2025

We’re so used to ambient doubt now – about models, outputs, and expertise – that the simple act of building a clean, unbroken microsite reads as a gesture of good faith. “Look: no ads, no tricks, just this argument.”

101

Ben Johnson @ersatzben.com · 28/04/2025

But AI microsites are about more than explaining ideas. They exude trustworthiness at a time when trust itself is precarious.

101

Ben Johnson @ersatzben.com · 28/04/2025

The microsite format gives writers a powerful toolkit: - Singular visual focus - Total control over pacing - Authority through design - A subtle feeling of inevitability In short: it performs seriousness.

101

Ben Johnson @ersatzben.com · 28/04/2025

What is it about the microsite that makes it so suitable for AI discourse in 2025? Let's dig into it 🧵

123

Ben Johnson @ersatzben.com · 16/04/2025

The opportunity is real. So is the risk. Get UKDI right, and we don’t just defend better – we grow better. Miss this chance, and we stay stuck – while others charge ahead. Let’s choose speed, sovereignty, and strategic clarity.

130

Ben Johnson @ersatzben.com · 16/04/2025

What’s needed: • A concentration of truly exceptional people – paid properly • Radical autonomy: freedom from civil service pay bands, procurement rules, and hiring constraints • Political cover: a serious Chair reporting directly to PM + Defence Secretary

130

Ben Johnson @ersatzben.com · 16/04/2025

UKDI should: • Identify classified capability gaps • Translate them into solvable tech problems • Fund & scale UK solutions fast • Kill failing ideas ruthlessly • Drive adoption via direct pull from military users

120

Ben Johnson @ersatzben.com · 16/04/2025

UKDI must avoid the trap of reforming everything and achieving nothing. It should instead be designed to bypass system failure. If UKDI gets a few key things right, it can show the rest of the system what’s possible.

130

Ben Johnson @ersatzben.com · 16/04/2025

But to ensure the UK catches up, UKDI must not be just another R&D funder. It must be a mission-led, high-agency engine, with expert Technical Directors empowered to run bold, fast, outcome-driven programmes. Think DARPA meets Vaccines Taskforce – with full political support.

120

Ben Johnson @ersatzben.com · 16/04/2025

That’s a problem in peacetime. But in the coming years it could prove catastrophic. Geopolitical risk is rising. Autonomy & AI are reshaping conflict. If we can't develop & deploy at speed, we’ll fall behind. The Government's announcement of UKDI is therefore a welcome move.

120

Ben Johnson @ersatzben.com · 16/04/2025

The UK has world-class research & growing defence investment – but promising technologies stall before they get to the front line. Why? Institutional inertia, slow procurement, and a lack of follow-through on promising techs.

120

Ben Johnson @ersatzben.com · 16/04/2025

🚨 NEW PAPER KLAXON 🚨 @britishprogress.bsky.social's first research report is out! It is all about the Government's announcement of a new defence innovation unit – UK Defence Innovation. Here’s how to build this unit for maximum impact on security and prosperity: 🧵

194

Ben Johnson @ersatzben.com · 31/12/2024

You know annual recap emails have gone too far when you get one from your fucking toothbrush company. Guys, it’s teeth. Dial it back.

061

Ben Johnson @ersatzben.com · 09/12/2024

This one paragraph from this story made me chuckle… Travel agent’s chatty AI call-handlers take off www.thetimes.com/article/b56c...

000