What Is the AI Singularity? The Glitch Pitch’s Core Claim

Strip the robot footage and the January 27th countdown out of the Infinite Money Glitch presentation and one intellectual idea remains underneath: the technological singularity. James Altucher’s model Press pitch is a stock story built on top of a concept from 1960s futurism, and the concept is worth understanding on its own terms, because the pitch’s biggest numbers are all extrapolations from it. Our full teardown covers the presentation; this piece explains the idea it is built on. We walk through the specifics in our Technological Singularity: From Good to the Glitch piece.

The Idea, Properly Credited

The singularity is the scenario in which AI improves AI, and the improvement loop runs faster than human oversight. The originator the presentation cites is I.J. Good, the British cryptographer and statistician who worked beside Alan Turing at Bletchley Park. In 1965, Good described what he called the “intelligence explosion”: once a machine becomes smart enough to design better machines, the design loop compounds beyond human prediction.

Altucher recounts it accurately in the pitch: “So Good called this the Intelligence Explosion. And today scientists call it the Technological Singularity.” The naming lineage is real and the concept has been standard vocabulary in AI research for six decades. What is new is not the idea; it is the evidence that the loop has started turning.

The Evidence the Pitch Uses

Three exhibits carry the argument, and we checked all of them.

AI now writes AI. The presentation’s statistics: “At Google, its own AI now writes 75% of the company’s code. At OpenAI, the company behind ChatGPT, AI writes 80% of its code. And at Anthropic, more than 90% of the company’s code is now written by its own AI, called Claude.” These figures come from the companies’ own leadership statements through 2025; the exact percentages move quarter to quarter, but the direction, AI systems doing the work of building AI systems, is documented.

The task horizon is compressing. The METR research institute measures how long a task horizon an AI model can complete reliably. Altucher’s summary: “In 2019, that number was doubling about every seven months. By 2026, the doubling time had shrunk to just four months.” That matches METR’s published trend line. A compounding loop that doubles its reach twice a year is the mathematical shape of Good’s explosion, arriving not as a single event but as a slope.

A central bank is now modeling it. The most striking exhibit: the Dallas Fed, in a June 2025 research piece on AI and productivity, explicitly analyzes the singularity scenario, defining it as one where “AI eventually surpasses human intelligence, leading to rapid and unpredictable changes to the economy and society.” A regional Federal Reserve bank formally modeling the intelligence explosion is new, and the presentation is right to treat it as a marker.

Where the Pitch Stretches

Here is the honest boundary. The Fed paper models multiple scenarios, and its singularity case is the aggressive one. The Fed’s own benign scenario, in the paper’s words, produces “a difference in GDP per capita in 2050 of only a few thousand dollars, which is not trivial but not earth shattering either.”

Altucher quotes the Fed’s definition accurately, then extends the Fed’s most aggressive scenario line into his own chart: “By 2035, it climbs to around $390 trillion, almost four times the size of the entire world economy right now. And by 2040, US output alone could approach $2 quadrillion in a single year.” He flags his method as following “the Fed’s own data set, adjusted into today’s dollars,” which is more disclosure than most presenters offer. But the quadrillion figures are his extension of the extreme case, not the Fed’s forecast, and the presentation’s framing sometimes invites you to forget which is which. A careful viewer should not.

Why the Robot Is the Delivery Mechanism

The pitch’s final move connects the singularity to Optimus: intelligence that stays in a data center improves software; intelligence with hands improves labor, the largest slice of the economy. Musk’s claimed version, quoted in the pitch: the glitch “could grow the global economy by a factor of ten or maybe 100.” The robot builds robots, every robot teaches every other, and the loop crosses from bits to atoms.

That is why the investable surface of this idea is the physical AI supply chain, which our physical AI explainer covers, and why the pitch’s reports center on suppliers rather than frontier labs. For the presenter’s own track record with big calls, see our James Altucher profile.

The singularity is a real research concept with real evidence accumulating under it. The quadrillion extrapolations are one optimist’s math on top of a central bank’s aggressive scenario. Both things are true, and knowing exactly where the first ends and the second begins is the difference between reading a pitch and being read by one.

NewsletterVetter is an independent publication. We receive compensation from some of the services we review through affiliate links. Nothing on this site is investment advice. Always do your own research.