Welcome to May 21, 2026

The Singularity has just cut out the middleman in Erdős’s quip that mathematicians are “machines for turning coffee into theorems.” An internal OpenAI model has disproved Erdős’s longstanding planar unit distance conjecture in discrete geometry, contradicting decades of belief that square grids were optimal. Notably, this came from a general-purpose model, not a math-specialized one. Mathematician Arul Shankar marveled that the model’s chain-of-thought tried “a vast array of ideas from a wide range of mathematics” before honing in methodically with what he called “original ingenious ideas.” Sam Altman called it “a kinda big milestone” evoking “complicated feelings,” while Noam Brown noted that less than a year ago frontier models were merely at IMO gold level. Epoch AI’s Yafah Edelman has pulled her median for solving most Millennium Prize Problems forward to 2032. The frontier is moving so fast that the White House is quietly briefing labs on an imminent executive order pushing 90-day pre-release notifications for frontier models, treating new releases like FDA submissions.

While math cooks, the application layer is fanning out into both the mundane and the absurd. Google is rolling out Gemini-powered conversational ads inside AI Mode and Search, generating tailored creative for queries about, say, making your home smell like a spa. On the unauthorized end of the spectrum, hobbyists are using Seedance 2.0 to “fix” the Harry Potter cinematic universe by violently dispatching the unpopular characters.

All of this runs on silicon that cannot be built fast enough. Seagate’s CEO concedes new factories would simply “take too long” relative to AI demand. Nvidia just posted a record $81.6 billion Q1, up 85% year-over-year, even as Jensen Huang acknowledges Nvidia has “largely conceded” the China AI chip market to Huawei. The real bottleneck has migrated from logic to power and concrete. SpaceX’s newly acquired xAI division is buying another $2.8 billion of turbines and Anthropic is now paying SpaceX $15 billion per year for compute, with Chief Compute Officer Tom Brown confirming Anthropic is scaling onto GB200 capacity in Colossus 2 through June, placing Anthropic in the surreal position of bankrolling its rival’s landlord. Not everyone is on board, however. St. Charles City, Missouri, just voted to effectively ban large-scale data centers, a reminder that the Singularity still has to clear local zoning meetings.

The natural escape hatch is straight up. SpaceX is preparing for the twelfth flight test of Starship as soon as today, while filing an IPO prospectus claiming a $28.5 trillion total addressable market, roughly the entire US GDP, spanning Starlink broadband and mobile, X advertising, AI infrastructure, and a Tesla-collaborated AI agent platform named Macrohard meant to emulate an entire AI-run software company. Orbital computing rival Jeff Bezos agrees data centers in space are “very realistic,” but called Musk’s 2-3 year timeline “a little ambitious,” a sign that the orbital-compute debate has quietly collapsed from physics to scheduling. Either way, compute itself is preparing to leave the planet.

Meanwhile, the wetware is being upgraded too. Startup Bexorg is now restoring some functions to intact brains from deceased donors, hoping to build a better drug development testbed for neurodegenerative diseases, and quietly redrawing the line between mortuary and laboratory bench. If compute is heading to orbit, cognition is heading back from the grave.

Capital is reorganizing around all of this at fantastic speed. A new wave of $37-100 billion in philanthropic funding is about to become liquid as the OpenAI Foundation’s 26% stake and Anthropic founders’ 80% giving pledges mature, a 6-17% boost to annual US philanthropy. Sam Altman is offering every YC founder $2 million in OpenAI tokens instead of cash via SAFE, betting on what he calls “tokenmaxxing startups.” With federal AI legislation foundering, OpenAI’s top lobbyist is pursuing a backup “reverse federalism” strategy, shaping state laws the industry can live with. Anthropic, for its part, expects 130% revenue growth to $10.9 billion this quarter and its first operating profit, defying every skeptic of the AI boom. Authorship itself has been demoted from fact to forensic question. The Commonwealth short story prize winner “The Serpent in the Grove” was immediately accused of being AI-generated upon publication, an ongoing referendum on whether human-only literature is even verifiable. Intuit is cutting 17% of its workforce, roughly 3,000 employees, to sharpen its AI focus. Meanwhile, OpenAI is preparing an imminent filing for an IPO, possibly within days.

Cogito, ergo IPO, the Singularity’s last theorem.