Skip to content

Hot I.P.O Summer + What Is A.I. Doing to Math? + HatGPT

Hard Fork

1h 10m
Transcript

Original podcast

Where to listen

Summary

Written by a model from the transcript; not yet reviewed.

Kevin Roose and Casey Newton open with the AI IPO race: SpaceX plans to sell shares at $135 for a $75 billion raise at a $1.75–2 trillion valuation, Anthropic filed a confidential S-1 with ARR jumping from roughly $1 billion in January 2025 to $50 billion, and OpenAI is expected to file soon. They discuss consequences for San Francisco inequality, a coming wave of effective-altruism philanthropy (Anthropic co-founders pledged 80% of their wealth to charity with a 3-to-1 matching program), and whether public-market pressure will push AI labs to cut safety corners.

Guest Kevin Hartnett, author of The Proof in the Code, explains why mathematicians are sounding the alarm. OpenAI's solution to the unit distance conjecture was deemed publishable in the Annals of Math, marking a shift from AI doing high school math to top-tier research. Approximately 800 mathematicians signed the Leiden Declaration calling for disclosure of AI use in proofs. Hartnett describes three camps: Terry Tao sees AI as a jetpack, one mathematician predicts it will end the profession in two years, and another dismisses it as a stupid thing.

HatGPT covers a robot company caught training in a rented Airbnb (the owner is suing for $12,383.50), Trump's executive order giving the government 30 days to review new AI models, George Santos's Kalshi insider-trading investigation, Meta AI's chatbot being used to hijack Instagram accounts, and a United flight forced to turn around because a passenger's Bluetooth speaker was named Bomb.

  • SpaceX's IPO at $135 per share would raise $75 billion and value the company at $1.75–2 trillion, making it the largest IPO of all time. 4:24
  • Anthropic's annualized revenue run rate went from about $1 billion in January 2025 to $50 billion, which the hosts call unprecedented growth in Silicon Valley. 9:06
  • OpenAI's proof of the unit distance conjecture was unanimously considered publishable in the Annals of Math, the top journal in mathematics. 38:19
  • The Math Archive now bans users for a year if they upload PDFs containing unedited AI prompt metadata. 43:47
  • Approximately 800 mathematicians signed the Leiden Declaration, calling for disclosure of AI use in proofs and warning that AI produces plausible but unreliable arguments. 41:55

Highlights

  • 4:24

    SpaceX's Record-Breaking IPO Numbersdata

    SpaceX plans to sell shares at $135 each, raising $75 billion at a valuation between $1.75 and $2 trillion, making it the largest IPO of all time. The company now includes XAI and X alongside rockets and Starlink.

  • 9:06

    Anthropic's Unprecedented Revenue Growthdata

    Anthropic's annualized revenue run rate went from about $1 billion in January 2025 to $50 billion. Kevin recalls visiting the company in 2023 when it was a small group of safety-obsessed people in a walk-up office.

  • 15:47

    Anthropic's Philanthropic Pledgesdata

    All Anthropic co-founders pledged to give at least 80% of their wealth to charity, and the company matches employee charitable pledges three-to-one for some early employees. The hosts estimate this will produce more philanthropic capital than the Gates Foundation annually.

  • 37:48

    AI Solves Top-Tier Math Probleminsight

    OpenAI solved the unit distance conjecture, considered one of the most important Erdős problems. The proof was deemed publishable in the Annals of Math, signaling AI can do absolutely top-tier research rather than just high school math.

  • 40:27

    Three Camps in Math on AIinsight

    At the Institute for Advanced Study, Hartnett encountered two mathematicians with opposite views in one afternoon—one dismissed AI as a stupid thing while another predicted it would put mathematicians out of business in two years. Terry Tao sits in the middle, seeing AI as a jetpack for thought.

  • 41:55

    Mathematicians' Leiden Declaration on AIstrategy

    Approximately 800 mathematicians signed the Leiden Declaration, expressing concern that AI produces plausible but unreliable arguments and calling for new rules about disclosing AI use in mathematical proofs. Terry Tao, who is excited about AI, also signed it.

  • 54:18

    Robot Training in Airbnb Homesstory

    A San Francisco startup called The Bot Company was caught training robots in a rented Airbnb without the owner's knowledge, leaving the house trashed with cables taped to walls. The owner is suing for $12,383.50.

  • 1:02:09

    Bluetooth Speaker Named Bombstory

    A United Airlines flight from Newark to Mallorca had to make an emergency landing after a 16-year-old's Bluetooth speaker, named Bomb, triggered security concerns that the crew could not resolve despite asking all passengers to turn off their devices.

About this episode

“We are on track to see what might be the three biggest I. P. O.s of all time.”

Transcript

0:00
1:10:25

13,047 words · 5 speakers

Topics

  • AI IPOs
  • Artificial Intelligence
  • Mathematics
  • Effective Altruism
  • AI Safety
  • Prediction Markets
  • AI Regulation
  • San Francisco Tech

Page updated · transcript and summary are machine-generated