Monthly Archives: November 2025

“AI Wrote That!”

“Some run from brakes of ice, and answer none; And some condemned for a fault alone.” [Measure for Measure, Act 2, Scene 1]

If only we could all write like Shakespeare. It’s sonorous, timeless, replete with metaphor and meaning. Now we have AI slop swilling around the internet. Finding something written by a human, something genuine, something worth reading, is like trying to pick out the orations of Cicero in a sports crowd as they roar for a touchdown. If you let yourself, you could drown in that cacophony of information.

The appearance of generative AI and its effectively infinite capacity to…well…generate has meant that you, poor reader, are now faced with the literary equivalent of a Deli full of lunchmeat—homogenised words with colouring and preservatives.

We need better ways of writing. We need to return to the old ways—a kind of writing where the artist, steeped in craft, can mold and form a narrative or argument and render it in a single draft. I am thinking here, at best, of a cuneiform tablet. But I would settle for ink, quill and velum. That is the true measure of the art.

We blame AI, but things really started to go wrong in the late 19th Century. The combination of wood-pulp technology and the Fourdrinier machine made paper cheap and available. And as paper became more affordable, thinking got lazier. Loose, ill-considered mutterings and on-the-fly musings could now be committed to paper and reworked through multiple drafts. There was no allegiance to de novo refined precision.

László Bíró, inventor of the ballpoint pen, and Marcel Bich, mass producer of the same, need to shoulder some of the blame (1933-1956). Even with the ready availability of paper, the blotches and smudgings of the maladroit kept many wannabe writers out of the market. Some thought they had good ideas, but manual dexterity was a solid benchmark for well-constructed prose.

The manual typewriter became a ubiquitous domestic item in the 1960s. Liquid paper had already been invented, which meant we could all become monkeys at the keyboard, randomly pecking in the hope of producing Shakespeare. These were followed in rapid succession by the electric typewriter and the electric typewriter with correction tape.

Between 1978 and 1983, authorship was no longer bound to paper. WordStar, WordPerfect, and Microsoft Word, running on personal computers, freed the illiterate to create everything from a letter to Grandma or a eulogy to a first novel. Effort and thought were gone. “Writing” was a mindless process of rinse-and-repeat. Spellcheck, grammar check, word suggestions, thesaurus (for the truly illiterate—or as I like to call them, the analphabetic) and “suggestions”.

And here we are—2025. Editors are inundated with crap because everyone is now a writer.

Claude, write me a bawdy Limerick proving the infinity of primes.

A strumpet proved primes never cease
By shagging each one for a piece,
She’d finish the set,
Find one larger yet,
Her clients increased without peace.

OK! It’s not Shakespeare. But it is a curiosity—a two-minute amusement. It’s also worth thinking about how that limerick comes to exist. The GenAIs are not monkeys at a typewriter. They are constrained. They respond to prompts. The outputs are not random. You might get lucky and one of the generative AI engines immediately produces a limerick worth two minutes of your life. The chances are, however, you will get dross, or it will be a proof, but it won’t be bawdy, or it will be bawdy, but it won’t be a proof. You will need to go back and forth with the AI, refining, editing, and selecting. It was your idea—a bawdy proof. You refined and selected. For a five-line limerick, it might not take much time and effort, but it does require some—and that process is creative.

When photography first appeared on the scene in the second half of the 19th Century, it was seen as the end of painting, because all painting was an attempt to reproduce reality perfectly (Not!). And all photography was the perfect reproduction of reality (also Not!). Photography is now accepted as an art form, although not always. The technology, however, is mechanical, and…. Where is the art?

I heard a story told of the renowned art photographer, Robert Maplethorpe. A woman commissioned him to take her photograph. He took dozens and dozens of photos on the day. When the woman returned some weeks later to receive her portrait, she was not entirely happy with it and asked Maplethorpe if she could see the other photos taken on the day. He refused. The other photographs are not “Maplethorpes”, he explained.

The production of the art might rely on a mechanical device—but the composition, the lighting, the post-production, and most importantly, the aesthetic choice is entirely in the hands of the artist. Maplethorpe might have been able to render a portrait in a fraction of the time it would take to paint the same picture—that is a matter of medium, however, not artistic merit.

If Shakespeare be the measure of literary art, then, Houston, we have a problem. Who in 2025 knows what that line from Measure for Measure means: “Some run from brakes of ice, and answer none; And some condemned for a fault alone”?

The Bard himself is unintelligible to the reader—and he is rarely, if ever, translated into modern English. The translation is an affront to the author as artist, which is ironic because Shakespeare almost certainly would have embraced the idea.

If he were translated, we might get any of the three following forms. There is the poetic and adherent: “Some hide in icy coverts, shun the call; and some are judged for but a single fall”. There is the plainer meaning: “The guilty hide and prosper; the unlucky answer once and fall”. And there is the prosaic: “Some people evade justice entirely by hiding and refusing to answer charges, while others are condemned for committing just one offence”.

The problem with AI is not that working closely with it cannot produce things of merit and worth: curated, thoughtful, and illuminating—things artistic and authored. The problem is the volume. We are looking for grains of black sand on a shore of white sand.

To judge “AI Wrote That!” as a dismissive and condemnatory act is as useful as looking at a Maplethorpe and declaring, “That’s a Photograph!”

ps: AI did not write this, except where it did.

The End of the Green-Growth Illusion

COP30 will produce nothing realistic to address a development, climate Ponzi scheme forced on the weak by the strong.

This chart reveals some stark realities about the link between development and carbon. It plots excess CO₂ consumption against GDP per capita, measuring each nation against a fair per-capita carbon allocation. Consumption-based accounting attributes the carbon from manufacturing to the country that buys the goods, not where they’re produced—revealing the true carbon costs of wealthy nations’ consumption patterns. The “fair” CO₂ allocation is based on a distribution of the global carbon sink—the biosphere’s capacity to absorb CO₂ annually. It has a finite budget, and the red lines show what a fair per-capita share of that budget would be for each country.

Every wealthy country is in the upper right quadrant. They are well above the red-line, consuming well beyond their climate share. The relationship is stark: a consequence of the current development strategies rely on generating CO₂. The highest GDP per capita achieved by any country without exceeding the fair CO₂ allocation is $16,000 by Costa Rica.

China, for its modest GDP per capita of $19,000 sits above the 75th percentile. That is, it is an inefficient producer of wealth for every tonne of carbon burnt—and this is after the carbon from exports is removed from the books. India under-consumes relative to its wealth as do Brazil and Colombia.

Norway achieves high GDP per capita whilst remaining near zero and Qatar achieves an enormous GDP per capita while sitting on the 25th percentile. These apparent climate successes hide structural problems: as energy exporters, they offload carbon costs to importing nations through consumption-based accounting. This isn’t a weakness of consumption-based accounting but it shows the fragility of the carbon exporting economies in the face of real commitments to reduce carbon emissions.

City-states like Singapore, Hong Kong, and Malta sit well above the 75th percentiles relative to their GDPs per capita, their high consumption reflecting imported goods produced elsewhere.

The difficulty deepens when we examine the geopolitics of the energy transition. We have a global development strategy that is almost entirely tied to carbon. Wealthy nations continue to burn carbon whilst promising to “decouple” growth from emissions. The promise is that green energy will save us, without us ever having to compromise on our expectations of wealth. The levels of solar-panel and wind-turbine manufacturing required to reduce emissions (not just growth in emissions) are staggering. Every one is looking to China—a single point of failure and an aggressive geopolitical force that (like all powers) will exploit it’s advantage—without credible alternatives. The world’s decarbonisation strategy implicitly assumes that China will continue to expand manufacturing of clean-tech materials at stable prices, and it will do this without entangling its dominance in geopolitical leverage. This is implausible. It is also historically unprecedented to have the world’s future energy system hinge on one state’s industrial capacity and beneficence. That single point of failure will endanger the entire project.

Addressing climate change requires “giving up”. It requires wealthy countries give up on their expectations of continued privilege and comfort at the expense of poorer nations. Poorer nations need to give up their expectations of ever achieving current wealthy nations’ standards of living. Finally, powerful countries need to give up on using climate as a geopolitical pawn in the search for national, strategic advantage.

My argument is, empirically, in the right direction and politically unpalatable. No political system—authoritarian, democratic, or otherwise—has shown willingness to articulate a program of “giving up”. As a consequence, we are sold a vision of techno-optimism that preserves the narrative of growth without sacrifice.

As soon as the bubble of that narrative if burst and becomes a more realistic narrative of growth versus sacrifice, it becomes abundantly clear that the powerful intend to sacrifice the weak.

The logic of the present system leaves no room for a fair transition. The global economy is built on the extraction of ecological capacity from the many to sustain the consumption of the few. The COP process, for all its technical ambition, is designed to protect that hierarchy, not dismantle it. It offers the appearance of collective action while ensuring that no high-income state is required to reduce consumption to anything approaching its fair share, and no low-income state is allowed to question the model it is expected to follow.

Once the comforting fiction of green, painless growth gives way, the reality becomes clear. The world is not preparing for shared sacrifice but for selective survival. The strong will preserve their position by externalising the costs of climate stabilisation onto the weak, whether through border adjustments, constrained development pathways, or the quiet abandonment of nations deemed geopolitically expendable. Under current structures, the future is not collective sacrifice but a brutal form of triage in which the most vulnerable are forced into destructive competition, fighting for space in a system designed to abandon them.

The data for the graph were from the Our World in Data (OWD) CO2 dataset. The graph and the analytic approach is not OWD’s nor endorsed by them. The arguments and conclusion are mine.