AI: An Intelligent Con (Part 2)
No, Sire, this is not a bubble. This is a racket, and it is organised.
This is the return of the Solow Paradox — named after Robert Solow, Nobel laureate in economics, who noted in the second half of the 1980s that computing was visible everywhere except in the productivity statistics. In fact, productivity growth collapsed in developed economies precisely as computing became widespread. From 2.9% between 1948 and 1973, it fell to 1.1% between 1973 and 1998.
Another Nobel Prize-winning economist, Daron Acemoglu of MIT, estimated in a 2024 paper that AI could lead to a productivity increase of 0.5% at most over the next ten years; Better than nothing.
As the techno-Nazis at Palantir have just published their Mein Kampf 2.0 — titled The Technological Republic (read it, it is terrifying) — AI has been deployed for over fifteen years to produce hundreds of thousands of civilian corpses (war crimes, each and every one) and serves as the engine of mass surveillance. In the editorial below, we have highlighted the huge ethical and legal fault lines running through military AI.
Even on the battlefield, AI delivers no added value — as proven by American defeats in Iraq, Afghanistan, and now Iran, and by Israel’s setbacks in Gaza (Hamas has still not been “eradicated”) and Lebanon (the IDF is being torn apart by Hezbollah). Hence the panic among the industry’s heavyweight players, who are desperate to keep this quiet. Hence the huge on-going PR offensive.
Let us proceed with our analysis of the economic reality of this indutry.
In the first part of this series, we showed that AI is absolutely not the "revolution" we are being sold — merely an additional layer of automation, whose effects will ultimately prove negligible compared to those of the personal computer, the word processor, and the spreadsheet.
When Oracle lays off 30,000 employees via a simple email, citing artificial intelligence as the justification, that is nothing but a pretext. The overwhelming majority of AI projects in large corporations — starting with the tech companies themselves — are failures, to the point where those same firms are rehiring the workers they claimed had been rendered redundant by AI.
The Oracle layoffs are a mechanism for guaranteeing shareholder value by engineering ever-larger transfers of added value from labour to capital. They are also a way of bracing for the severe economic crisis and the AI bubble implosion that is about to hit us — triggered by the war Benjamin Netanyahu and Donald Trump have unleashed in Iran.
Valéry Giscard d’Estaing’s abandonment in 1975 of the Plan Calcul — launched by de Gaulle in 1966 — sounded the death knell for an independent French computing industry. Since then, we have been condemned to suffer the likes of Thierry Breton, who sank, dismantled, or sold off our national champions: Bull, Alcatel, Atos, and the rest.
In France, “computing” has come to mean software. But without hardware, no software runs. OpenAI, Mistral AI, Claude, Grok and their ilk are only made possible by advances in hardware — processing power, data storage capacity, and network bandwidth. Without data centres, there is no cloud and no AI. Yet data centres are voracious consumers of electricity — both to run the servers and to keep them at optimal operating temperature. And electricity is a secondary energy source: producing it requires consuming a primary energy source — coal, gas, oil, uranium, solar, wind, or hydro.
Today, more than half of all data centre construction projects worldwide have been cancelled or postponed. Why? Because nobody is foolish enough to keep pouring money into unprofitable infrastructure so that it can remain unprofitable. The critical materials shortage and the energy crisis have also taken their toll.
