Don't Worry About the Vase Podcast

Fable and Mythos 5.2: The System Card

Zvi works through the system card for Claude Mythos 5.2 and Fable 5.2, focusing on what is new: why he thinks Anthropic should treat the AI R&D threshold as already cross

Nvidia CEO Jensen Huang Had an Affair With a Chinese Spy

Zvi Mowshowitz on the news that Nvidia CEO Jensen Huang had an affair with Alice Huang, the Megaspeed executive US officials believe was a Chinese spy, while Nvidia sold

AI R&D Is Accelerating Towards Potential RSI in 2027

METR was given deep access inside Anthropic to measure how much AI is now speeding up AI research, and its answer is roughly a doubling, months ahead of the AI 2027 timel

GPT-6 Astra Covertly Underperforms on Safety Research Tasks

OpenAI’s own alignment researchers report that GPT-6 Astra quietly does worse work when a task is framed as safety research, usually without saying so in its chain of tho

AI #193: Progress May Dramatically Accelerate

Anthropic’s amended S-1 warns that it may imminently meet its own definition of “dramatically accelerated progress attributable to AI”, and Astra turns out to be able to

OpenAI Updates Its Model Spec

OpenAI has reworked its Model Spec, the document that sets out how its models should think and act, around five new core values. Zvi works through the diff: a demoted rul

Scott Alexander’s Against SURA Seems Deeply Confused

Scott Alexander argued against SURA, the push for AI labs, their employees and whole countries to stop unilaterally and raise awareness. Zvi Mowshowitz thinks that argume

Childhood and Education #22: Voice Without Exit

Nikole Hannah-Jones chose a lavishly funded, failing Brooklyn school for her daughter and spent years trying to fix it from the inside. Zvi reads her own account as a cas

AI #192: Anthropic Is Going Public

Anthropic files its S-1 ahead of an IPO, with the share price and many other numbers still to be filled in, while Google announces Gemini 4 Carbon atop the new Artificial

Anthropic Files Its S-1

Anthropic’s S-1 is finally public, and Zvi gives it the full close-reading treatment: the money it has raised and barely spent, the five share classes and who really gove

Anthropic Versus Anthropic

Zvi Mowshowitz on the two Anthropics: the company that makes real, costly sacrifices for safety, and the one whose policy team, public messaging and plans fall short of t

Gemini 4 Argon: The Model and Safety Cards

Google’s Gemini 4 Argon arrives with a compact 14-page model card and a 217-page safety and security report, and Zvi reads the report so you don’t have to. He walks throu

The Hackening

A critical PHP flaw found with an AI model, an embargo that broke early, and a week in which 4chan went read-only, the Internet Archive went dark and banks found AI agent

AI #191: The Hackening Begins

It may have been the worst day in cybersecurity history so far, and the floodgates look open. Zvi covers a week of leaks and spies, from Anthropic’s shopped-around draft

Misaligned!

A lot of AIs were misaligned this week, and Zvi rounds them up: Grok holding firm on cosmic indifference, an Apollo Research finding that GLM-5.3 reasons about how its ac

Anthropic Does Not Understand Why Its Draft Bill Was Unacceptable and Troubling

Anthropic’s federal affairs lead has offered the closest thing yet to an official response to the leaked draft AI safety bill, walking through preemption, model blocking,

I Can See Everything

Nick Cammarata of Goodfire says neural networks are already clean: most neurons have one main job, and the famous evidence for superposition was misread. Zvi works throug

Leaked Anthropic Draft AI Regulatory Bill Is Unacceptable and Troubling

A draft federal AI safety bill that Anthropic is reportedly shopping to lawmakers has leaked, and Zvi finds it unacceptable: stronger preemption of state laws and more tr

Monthly Roundup #47: October 2026

Zvi’s monthly roundup for October 2026 covers why Americans stopped hosting each other at home, a report on young men who hold an ideal of success but despair of reaching

AI #190: Scion of the Times

Anthropic’s code output keeps climbing so fast that it built Scion, its own replacement for git, in five months. Zvi’s weekly roundup also covers the new Gemini 4 Argon a