David Shapiro’s Substack

Below is a transcript of a conversation I had with the Fourth Founding Project, which is techno-progressive group interested in the future of politics, humanity, technology, and government in the United States. They reached out to me due to my work on Post-Labor Economics and my upcoming work on Structural Realism. The transcript has been cleaned up from our recording. David Shapiro: I would start with the premise, which is that automation has been advancing for decades.

In this article I discuss: The dying hype of AI The problem of “Shadow IT” in enterprises AI burnout and addiction Hilariously bad security practices with agents In my last post, I mentioned that the “show is over” and implied that the Gartner Hype Cycle is still deflating. I think that is true.

Does anyone else get the impression that, outside of a few prestige tasks, Claude just isn’t that bright? Drafting fiction and nonfiction, generally pretty brilliant. ChatGPT is better at nonfiction because it doublechecks everything, whereas eventually Claude just gets lazy and aims for “vibes plus sounds good, plus I can never admit when I’m wrong unless the user really corners me” But even Claude Fable on Max will still utterly destroy some tasks.

In chess, there’s a concept called the early game, the midgame (or middle game), and the endgame. We just crossed the boundary into the middle game with the creation and subsequent restriction of Anthropic’s Mythos/Fable class models, and OpenAI’s GPT 5.6 Sol, Terra, and Luna models. Let’s unpack what this does and does not mean.

I’m an AI maximalist, which in an age where AI’s popularity is dropping off a cliff, makes me a bit of a contrarian. At least in the broader cultural conversation. So here’s what I mean by “AI maximalism” and why I think it’s a reasonable position. Rather than try to give a rigorous pseudo-academic definition, I’ll use analogy. I’m an AI maximalist in the same way I’m an electricity maximalist. It doesn’t mean I want to put AI into every little thing just because.

I’m ready to sign the divorce papers. Labor and capital need an amicable separation agreement. That’s what I’m laying out today. What started as just Post-Labor Economics has evolved into the Labor Zero (L0) movement, a growing coalition of thinkers, builders, and ordinary people who understand that automation represents a categorical threat to our present way of life, as well as an opportunity to abolish the drudgery of labor as we know it.

A pair of articles went viral on Twitter yesterday, and the reason they are resonating is more interesting (to me at least) than the articles themselves. The first article, “The Window Has Closed” by Andrew Curran, says that any nation which missed the Fable/Mythos milestone will inevitably fall further and further behind. Think of it this way—the US has declared export restrictions on that model, just as Leopold Aschenbrenner predicted several years ago.

There’s a popular turn of phrase when criticizing LLMs—“the world is not made of words.” This critique is levied at AI models as an argument that they cannot possibly have “world models.” But first question: what is a “world model”? When you ask most people, what they describe is somewhere between a cognitive architecture and a sensorimotor feedback system.

The problem is not the idea that the public should share in AI’s upside. The problem is trying to seize half of today’s model labs while missing the rest of the value chain. Bernie Sanders has proposed a new American AI Sovereign Wealth Fund, funded by a one-time 50% tax paid in stock by the largest AI companies. Not a 50% tax on profits. Not a windfall tax on revenue. A 50% equity tax.

A member of my Labor/Zero Skool community asked a great question recently, and I wanted to give it a proper answer. The short version is: I think a lot of people are still confused about what “AI fluency” actually means. It is not just using ChatGPT once in a while. It is not just vibe coding. It is not just knowing a few prompts. To me, the real questions are: What is basic AI fluency? What is it not? What practical AI skills should every professional start developing?