iiinova — where your digital self keeps evolving

Peter Evans-Greenwood @peter.evans-greenwood.com · 29/09/2026

My latest post, "What Replaces the Proxies." The one where I use a 90-second BGP hijack from 2008 to explain why nobody can settle whether AI is transformative.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 22/09/2026

My latest post, ‘Argue with a Row’. The one where I show the working, not just the answer.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 15/09/2026

My latest post, ‘Which Way it Runs’. The one in which I trace a liability cap back to 1957 and watch a forty-year-old story about fear collapse.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 09/09/2026

My latest post, 'Who Killed the ROI?' The one where I treat 'is AI paying off' as a five-suspect investigation instead of a yes/no question.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 03/09/2026

“Why can’t industry scale AI” and “why doesn’t AI show up in productivity stats” are the same story told twice.

010

Peter Evans-Greenwood @peter.evans-greenwood.com · 03/09/2026

My latest post 'What the Barrel Bought'. The one where I go looking for the institution that's supposed to be holding the cost of the renewables transition's growing pains, and find it's nobody's job right now.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 26/08/2026

My latest post, 'The Report That Got Checked'. The one where I argue the Deloitte scandal was never really about AI.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 19/08/2026

My latest post, "At the Bottom of the Shaft." The one where I take AI through the same test that would've told you a beam engine was forty years from a railway—and get back "maybe, but unlikely."

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 11/08/2026

Interesting: alignment training looks like it's importing human cognitive failure modes along with human-like fluency.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 11/08/2026

My latest post, ‘Three Questions, in Order’, the one where I the one where I explain why electrification took four decades to move a single productivity chart, and what that means for every AI ROI argument happening right now.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 04/08/2026

My latest post, 'The Cost of Knowing'. The one where I name the new failure mode: a critique that sounds rigorous, aimed at the wrong target.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 29/07/2026

My latest post, 'Nobody Was Slow'. The one where I explain why a 20–30% productivity jump can sit inside twenty-three years of nothing.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 26/07/2026

Boards of Canada’s Inferno is their most controlled yet. They seem to be finally masters of their chosen tools and they’re working out what they want to say. The next album could be amazing.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 23/07/2026

My latest post, ‘Nobody Designed It’. The one where I find the closest historical parallel to my driver’s network isn’t a commons—it’s eleventh-century traders in the Mediterranean.

100

Reposted by Peter Evans-Greenwood

Rodney Brooks @rodneyabrooks.bsky.social · 15/07/2026

"AI...making things expensive for the rest of us. Large language models...are so resource-hungry that tech companies...purchasing 70% of the world’s supply of high-end computer memory, causing a shortage." The opposite of what we expect from new technologies. www.theatlantic.com/technology/2...

theatlantic.com

Generative AI Is an Engineering Disaster

A shockingly inefficient trillion-dollar project

43710

Peter Evans-Greenwood @peter.evans-greenwood.com · 15/07/2026

My latest post, ‘Why the Terrain Looks Like This’, the one in which I argue Baldwin’s third unbundling hasn’t stalled — it’s still waiting for a vocabulary that doesn’t exist yet.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 07/07/2026

My latest post, ‘Why Productivity Waves Have a Shape’. The one where I suggest that the housing crisis and the AI wave have the same underlying cause—and it’s not the technologies.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 07/07/2026

The last few years have been a kind of natural experiment. Many people assumed intelligence was a scalar quantity—if you increased the amount of “intelligence” available, economic output & social capability would increase proportionally. That prediction has not held up well.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 30/06/2026

My latest post, 'We Called It Agency', the one where I introduce you to Catherine Helen Spence, who spent fifty years being corrected by the world and won, and ask why we've been calling something else by her name.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 16/06/2026

My latest post, 'What Institutions Actually Do'. The one where I show why Britain's superior electrical science didn't build the better grid. Capability wasn't the question. Liability was.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 09/06/2026

My new post 'The Squirrel and the Feeder'. The one where I explain why the squirrel is smarter than the venture capitalist. (It has no self-concept to protect.)

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 07/06/2026

Inferno feels less like ‘another Boards of Canada album’ and more like Boards of Canada finally asking the questions that were always implicit in their work.

010

Peter Evans-Greenwood @peter.evans-greenwood.com · 26/05/2026

My latest post, 'At the Edge of the Thinkable'. The one where I follow Galileo borrowing a chain to think with, and what his error reveals about how concepts actually form.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 20/05/2026

Is it a year now that the vendors have all been declaring that “the era of AI experimentation is behind us”? It’s become boilerplate on all their announcements.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 19/05/2026

My latest post, 'Two Questions on LLMs'. The one where I name the reason LLM advice keeps contradicting itself, and the two questions that resolve it.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 13/05/2026

History is full of partially assembled futures: - airships, - nuclear power, - expert systems, - Concorde, - virtual reality (first wave), - metaverse infrastructure, - blockchain…

030

Peter Evans-Greenwood @peter.evans-greenwood.com · 12/05/2026

My latest post ‘What It Was Always For’. The one in which I explain why Edison made the current and Ford made the miles, and what that has to do with why Bruce Springsteen’s masters sold for $500 million and a venue in Blackpool charges £10 on the door.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 07/05/2026

LLMs are strange as latent space is higher dimensional, so everywhere is close to everywhere else. Training of all forms is output shaping, and if the shape you create likes goblins, then goblins will appear everywhere.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 05/05/2026

My latest post, 'Showing the Joints', the one where I argue that LLMs aren't primarily writing tools. They're reading tools—and that's the shift nobody is talking about.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 28/04/2026

My latest post, 'Where the Costs Are'. The one in which I argue that AI currently looks a lot more like the nail gun than the dynamo—and explain why that matters for where you look.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 22/04/2026

The nail gun transformed construction productivity. Nobody built a monopoly on nail guns.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 21/04/2026

My latest post: 'When Grammars Can Be Imported'. The one where I argue that Australia's twelve-thousand-mile handicap was the thing that built the capacity. Britain's proximity was the problem.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 07/04/2026

My latest Substack post, ‘The Relationship is the Skill’, the one where I show how the same AI output was a verification crisis for one person and a prospecting tool for another, and neither of them was wrong.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 31/03/2026

My latest Substack post ‘The Empty Chair’. The one where I argue the chair is ready, the preparation is sincere, and the guest won’t be in the room where the invitation was written.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 29/03/2026

The Physical AI playbook is looking a lot like the Metaverse one

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 27/03/2026

The consciences debate continues in NOEMA, with Barton Friedland’s ‘What The AI Consciousness Question Conceals‘ attempting to take down Anil Seth's 'The Mythology of Conscious AI'.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 24/03/2026

My latest Substack post, ‘The Price Was Never the Point’. The one where I show why BMW’s heated seat subscription failed and Texas Instruments’ identical move forty years earlier became a business school classic. Same logic. Opposite outcomes. The difference is structural.

100

Peter Evans-Greenwood @peter.evans-greenwood.com · 23/03/2026

First Words argued the attention economy ended the subsidy, not the reader. Hollywood ran the same playbook—institutional identity replaced earned commons—and got the same result. Hail Mary is the market correcting. open.substack.com/pub/thepuzzl...

open.substack.com

First Words

Venue used to carry the writer. That credit is gone. What remains is the writing — and the moment before the reader decides whether to stay.

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 20/03/2026

Writing is not thinking Writing is the externalisation of thought into symbols and manipulation of those symbols. That's not thinking. It's just one cognitive loop among many. buff.ly/bRz4Bzh

buff.ly

Writing is not thinking – PEG

PEG

000

Peter Evans-Greenwood @peter.evans-greenwood.com · 20/03/2026

It’s a failure of quality. ‘Hachette/Orbit’ was attached to a market signal, not to judgment. The imprint was laundering the Goodreads score into a prestige credential and downstream quality control failed. No wonder the publishers are in trouble. AI is just a convenient villain.

000