
Sign up to save your podcasts
Or


Levent Alpöge of Anthropic announced on Friday, November 20 that he and Claude had found a proof of the Riemann Hypothesis, the most famous open problem in mathematics, and released it as roughly 17 million lines of machine-checkable Lean code instead of a paper. By Saturday the Lean and Mathlib communities had certified both the statement and the proof as correct, while the Clay Institute, Kevin Buzzard and Alpöge himself said human understanding of the argument has not caught up. An all-human rival draft posted from Stanford hours later is the one reviewers are finding problems in.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
At a White House news conference on Friday, November 20, President Trump said U.S. companies will be barred from building data centers in other countries or face 'a tremendous tax' if they do, a policy he said he had not yet told his own officials about and for which no legal mechanism was given. He named a project in Finland (the Washington Times ties it to Google's roughly $15 billion Finnish investment) and urged American towns not to block data centers but to 'make them pay you', citing $50,000 for Louisiana teachers. Those bonuses are real, but pay records obtained by The Center Square show they came from a 1% parish sales tax that spiked during construction of Meta's Hyperion campus, not from Meta. The day before, the Senate passed a bipartisan bill requiring data centers to pay for their power and grid upgrades.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
Between Nov. 12 and Nov. 17 Microsoft 365, the FAA's air traffic network, JPMorgan Chase, Apple services and the NSA's classified internal network all failed. On Nov. 18 Fortune reported that Sen. Mark Warner's office linked the NSA and banking outages to botched patches for vulnerabilities that AI models are now finding faster than organizations can safely fix them; the FBI and the Director of National Intelligence told a House hearing the same day there was no malicious activity. JPMorgan and Microsoft do not tie their own outages to AI-found flaws, and Downdetector's parent says the cluster looks like coincidence.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
On Wednesday, November 18, Anthropic priced its IPO at $83 a share, above its $70 to $78 range, and raised the size to 980 million shares, about $81 billion at a $1.5 trillion valuation, with trading to start Thursday on Nasdaq as ANT. CNBC's sources put demand near $500 billion, a fifth of the shares are reserved for small investors, and one of them told CNBC he emptied his Roth IRA because of the company's 'safety first' emphasis. The same morning the Center for AI Safety and about ninety signatories, from Steven Pinker to the founder of the e/acc movement, published a statement saying Anthropic 'uses safety branding to court investors' and 'is not a safe AI company.'
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
On Tuesday, November 17, Nvidia reported its first quarter since Jensen Huang resigned: revenue of $115.8 billion, up 103%, with about $130 billion forecast for the current quarter. Hours earlier Huang said publicly he would be 'honored' to return as CEO; board chair Harvey Jones opened the earnings call by saying the board 'does not intend to reinstate' him. Interim CEO Colette Kress, the finance chief the Wall Street Journal says was already skeptical of Huang's financing deals, told analysts Nvidia will be 'disciplined and selective' and is 'rigorously evaluating future investments, commitments, and guarantees', after disclosing $118 billion of guarantees and $390 billion of purchase commitments.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
Fed Chairman Kevin Warsh opened a Fed conference in San Francisco on Monday, November 16, by saying staff are gaming out three AI scenarios: a boom shared with workers, a boom in which automation pushes unemployment up anyway, and a bust in which AI disappoints and stocks fall. From the stage he asked whether AI could do to services what the China shock did to goods, cheaper prices at the cost of jobs, while his committee weighs a December rate hike because AI spending is heating up the economy. The Fed's one lever, the price of borrowing, does not obviously reach a worker displaced in a growing economy.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
On Saturday, November 14, SpaceX's launch tower caught a Starship upper stage returning from orbit for the first time, about 22 hours and 14 orbits after liftoff; the booster was lost the day before, so the hoped-for double catch failed. Full reuse is the precondition for the orbital data centers Elon Musk now says are SpaceX's future: he posted 'Data centers in space looking pretty appealing right now' the day Texas elected a governor who promised a data center moratorium. But the same week SpaceX announced more than $74 billion of data centers on the ground, and the Texas Tribune reported that its $24 billion Starbase campus sits on land Brownsville de-annexed at the company's request, outside the 90-day data center moratorium Brownsville adopted on November 4, and inside Starbase, where two of the three commissioners who will vote on its tax break December 9 are SpaceX executives.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
Fortune reported on Friday, November 13 that OpenAI has dissolved its stand-alone robotics team and folded its people into the main research group under Mark Chen and the Applied organization under Greg Brockman; OpenAI confirmed it and says robotics remains core. The stated reason is that general-purpose frontier models such as GPT-6 Astra can now be applied directly to robots: in an Nvidia study an Astra-powered agent scored 74.8% on a simulated manipulation benchmark against 51.0% for the best specialized robot models. Anthropic made a similar claim for its own new model four days earlier, which matters for every worker who was told that hands-on work is the part AI cannot reach.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
On Thursday, November 12, the Irreplaceable Education Fund, Public Citizen, Heat Initiative and the Alliance for Secure AI sued Anthropic, Google, Meta, OpenAI and SpaceX's AI arm in San Francisco Superior Court, asking for an immediate injunction barring them from developing or deploying any AI system more capable than their best one today, and from using AI to autonomously run AI research. The 63-page complaint brings a single claim under California's Unfair Competition Law and builds its case from the companies' own statements, including Anthropic's amended IPO filing two days earlier. No injunction is in force, the companies have not responded, and a founder of one plaintiff told Al Jazeera a real freeze would need government inspections, verification and penalties.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai
Nvidia announced on Thursday November 12 that Jensen Huang has resigned as president, CEO and board member, effective immediately, after 33 years; a regulatory filing says the board accepted the resignation Wednesday, the day the New York Times reported a second relationship, with an Nvidia vice president who reported to him. CFO Colette Kress becomes interim CEO while keeping the CFO job, director Harvey Jones becomes the board's first-ever chair, and a search will consider internal and external candidates. Nvidia stock fell 14% at the open and closed down 4.3% while AMD and Broadcom rose nearly 6%, leaving the S&P 500 down only 0.4%; CNBC reports Huang personally signed off on the financing deals that underwrite much of the AI buildout, and Nvidia reports earnings Tuesday.
Continue the story on unscarcity.ai:
Full episode notes on unscarcity.ai