“Intelligence is 100 percent human. AI is a tool created by humans to distill, digest, and distribute intelligence.” — Keith Teare
The working class died this week — at least in Palo Alto. Delivering the eulogy in our regular That Was The Week tech summary is my co-host Keith Teare. “Humans create intelligence,” (whatever that means) the Silicon Valley-based entrepreneur tells us. And so, in our AI age of supposedly abundant intelligence, he pronounces, human knowledge “should not be trapped inside experts, institutions, or companies.”
Check your pockets, everyone. Silicon Valley has another freebie for you. With AI, the entrepreneur promises, intelligence is democratized. Everybody gets it. We will all have the intelligence of a Nobel laureate at our fingertips. Even Keith.
And so he attacks Daron Acemoglu, the Nobel Prize-winning MIT economist who has called for a “pro-worker AI.” But, for Keith — a council-estate kid from Yorkshire whose lifetime ambition was to evacuate the working class — this is “complete bullshit.” Acemoglu’s ideas, he says, are an example of the “fetishization of workers” when, in fact, we should be celebrating the end of the “working class.”
What Acemoglu is calling for in his pro-worker AI manifesto is more government planning for today’s transition to the AI epoch. But Keith disagrees. So I asked him three times what government should do while AI kills the working (and middle) class. “Allow it to happen,” he finally answers. “A good upheaval.”
Good? The former “worker” will lack jobs, wages, healthcare, housing. Even food in an America now eliminating food stamps. No matter. Let them eat intelligence.
Five Takeaways
• Humans Create Intelligence. Keith’s editorial thesis distinguishes individual intelligence — where experts live, and always will — from the collective sum of everything all humans, living and dead, have ever contributed. That collective stock was once locked in encyclopedias, libraries, and universities; for the first time, AI can aggregate, distill, digest, and distribute it, at a price falling toward everyone. Knowledge, he writes, “should not be trapped inside experts, institutions, or companies” — but note the fine print: experts don’t disappear in this democratization. If anything, they get elevated: the expert reading an AI’s output about viruses understands it very differently than the rest of us.
• The End of the Age of Heroes? Noah Smith’s much-shared essay argues that AI ends the era of the mathematical hero — and that’s fine, since most people (truck drivers, financial advisers, executive assistants) never got to be heroes anyway. Keith’s rebuttal turns on his central distinction: AI and intelligence are not the same word. There is no evidence, he argues, that AI creates new knowledge — it understands and distributes the existing stock. Innovation still takes individuals, and those individuals now start from a far higher floor, leveled up to everything already known. Heroes don’t go away; they multiply. In the world of AI, he suspects, every single teacher becomes one.
• “What Even Is Pro-Worker AI?” The week’s main event: Daron Acemoglu — via Yascha Mounk’s Persuasion interview and an Atlantic essay, with What Happened to Liberal Democracy out next week — wants AI agencies, grant programs, and public competitions to build “pro-worker AI.” Keith’s verdict: “complete bullshit.” The middle-class “fetishization of workers” is paternalistic; the wage is a temporary power relationship between employer and employee; and the end of the working class is precisely the progressive outcome — says the council-estate kid from Yorkshire whose aspiration was not to be working class. Pressed three times on what government should do amid the upheaval, Keith finally answered: “Allow it to happen… a good upheaval.” Though swap workers for people, he conceded, and he’d almost entirely agree — every teacher a hero, even in East Palo Alto.
• Bandwagons and Silences. Regular people are being arrested protesting data centers; Erin Brockovich — a Keen On guest some years back — is assembling class actions; Ezra Klein has begun folding anti-big-tech language into abundance. A politician-led bandwagon, Keith argues, regressive but keyed to genuine local concerns. The stranger fact is the silence on the other side: neither Altman nor Amodei nor Demis Hassabis is making the public case that AI benefits everybody — astonishing, Keith says, and the vacuum Acemoglu is trying to fill. Hassabis himself stepped aside at Google this week — a scientist returning to science as Sergey Brin becomes AI czar — while the Nobel-winning AlphaFold team has been quietly broken up. “Something strange is going on there.”
• A Drama in a Teacup. Is the AI economy real? Ed Zitron’s stat — 70 percent of Amazon, Microsoft, and Google’s AI revenue comes from OpenAI and Anthropic — is two-thirds right, says Keith, and no problem at all: beneath the concentration, the money comes from some two billion distributed users paying real subscriptions, and the revenues are sustainable. The Aschenbrenner postscript, via Porter Stansberry’s post of the week: he bet the chip layer (Samsung, SK Hynix) when the value sat a layer up, got the timing wrong more than the thesis, sold to Citadel at a discount — and kept his Anthropic shares, remaining a multi-billionaire. As for the coming reality check: Anthropic and OpenAI will IPO only when public capital beats private, and SpaceX’s wobble from $135 to $108 — through a 20 percent lockup release — counts as no catastrophe. Public markets, Keith reminds us, don’t determine the success of the underlying business.
About the Co-Host
Keith Teare is the publisher of That Was The Week, the essential weekly tech newsletter, and founder and CEO of SignalRank Corporation. A serial entrepreneur — co-founder of, among others, EasyNet and RealNames — he was present at the creation of the UK internet and has spent four decades at the intersection of technology, capital, and ideas. He joins Keen On America every Sunday to make sense of the week in tech. His AI-assisted book in progress is titled Who Owns Intelligence.
References:
• That Was The Week — Keith’s newsletter, including this week’s editorial, “Humans Create Intelligence.”
• Noah Smith — “The End of the Age of Heroes,” on what happens to human ambition when the machines do the math.
• Daron Acemoglu — the Yascha Mounk interview at Persuasion, the Atlantic essay on pro-worker AI, and What Happened to Liberal Democracy, out next week.
• The Financial Times — “Google’s AI shakeup boosts Brin as DeepMind’s Hassabis steps aside.”
• Ed Zitron — on the 70 percent of hyperscaler AI revenue that flows from OpenAI and Anthropic.
• Porter Stansberry — post of the week, on Leopold Aschenbrenner’s losses, Citadel’s discount, and the drama in a ...
