Jerry
Preview
Open in TheBrain
Welcome! (start here)
Movie Critics
Nino Frank
Films Noir
Neo-noir
Dark City (1998 film)
The Matrix (1999)
And don't worry about the vase...
Don't Worry About the Vase (blog)
Don't Worry About the Vase (Substack)
Attachments
thezvi.substack.com
Parents
And don't worry about the vase...
Don't Worry About the Vase (blog)
How to Keep Up to Date on GenAI
My Favorite Substacks
Substack Pubs
Zvi Mowshowitz
Children
2023 Prediction Evaluations
2024 ACX Predictions: Blind/Buy/Sell/Hold
2025 State of AI Report and Predictions
2025 Year in Review
A ‘Morally Binding’ White House Accord on AI Safety
Against Modesty’s Bailey
AI #107: The Misplaced Hype Machine
AI #109: Google Fails Marketing Forever
AI #150: While Claude Codes
AI #154: Claw Your Way To The Top
AI #158: The Department of War
AI #163: Mythos Quest
AI #177 Part 2: Wish You Were Here
AI #179 Part 1: A Louder Fire Alarm for General Intelligence
AI #185: Preference Cascade
AI #186: The World Takes Notice
AI #188: Gemini Dot Argon
AI #190: Scion of the Times
AI #41: Bring in the Other Gemini
AI #43: Functional Discoveries
AI #44: Copyright Confrontation
AI #51: Altman's Ambition
AI #56: Blackwell That Ends Well
AI #60: Oh the Humanity
AI #63: Introducing Alpha Fold 3
AI #64: Feel the Mundane Utility
AI #69: Nice
AI Text Watermarking Is Free And Good
AiPhone
America Makes AI Chip Diffusion Deal with UAE and KSA
America's AI Action Plan Is Pretty Good
American Government Takes Down Claude Fable
An Alien Mind: Jakub Pachocki Warns Us
Anthropic Officially, Arbitrarily and Capriciously Designated a Supply Chain Risk
Anthropic Risk Report: August 2026
Anthropic vs. DoW #6: The Court Rules
Are They Starting To Take Our Jobs?
Arguments About AI Consciousness Remain Highly Motivated and At Best Overconfident
Astra 6.1 Pulled As Insufficiently Aligned
Balsa Update and General Thank You
Based Beff Jezos and the Accelerationists
Better Call Sol Or Better Yet Claude or Astra
Book Review: If Anyone Builds It, Everyone Dies (deleted?)
Book Review: Open Socrates (Part 1)
Book Review: Open Socrates (Part 2)
Brand New AI Solves a Millennium Prize
Bubble, Bubble, Toil and Trouble
ChatGPT Self Portrait
Cheaters Gonna Cheat Cheat Cheat Cheat Cheat
Childhood and Education #14: The War On Education
Childhood and Education #16: Letting Kids Be Kids (xlnt and long)
Childhood And Education #17: Is Our Children Reading
Childhood and Education #18: Do The Math
Childhood and Education #9: School is Hell
Citrini's Scenario Is A Great But Deeply Flawed Thought Experiment
Claude Code, Cowork and Codex #6: Claude Code Auto Mode and Full Cowork Computer Use
Claude Codes
Claude Codes #3
Claude Coworks
Claude Fable 5 and Mythos 5: The System Card
Claude Mythos #2: Cybersecurity and Project Glasswing
Claude Mythos: The System Card
Claude Opus 4.5 Is The Best Model Available
Claude Opus 4.6 Escalates Things Quickly
Claude Opus 5: The System Card
Claude Opus 5.5 Should Raise Your Ambitions
Claude Opus 5.5: The System Card
Claude Sonnet 4.5 Is A Very Good Model
Claude Sonnet 4.5: System Card and Alignment
Claude's Constitutional Structure
Copyright Confrontation #1
Dating Roundup #11: Going Too Meta
Dating Roundup #12: Sex and Violence
Dating Roundup #2: If At First You Don't Succeed
Dating Roundup #8: Tactics
Dating Roundup #9: Signals and Selection
DeepSeek Panic at the App Store
DeepSeek v3: The Six Million Dollar Model
DeepSeek: Don't Panic
Demis Hassabis on the New Coming Age
Do Not Mess With Scarlett Johansson
Economics Roundup #2 (on taxing unrealized revenue)
Fable and Mythos: Model Welfare
Fertility Roundup #5: Causation
Frontier Lab Employee Open Letter Calls For Being Able to Pace the Frontier
Fun With GPT-4o Image Generation
Gemini 1.0 (technical review)
Gemini 2.5 is the New SoTA
GLM-5.2 Is The New Best Open Model
Google I/O Day
GPT 5.5: The System Card
GPT-4o My and Google I/O Day
GPT-5.4 Is A Substantial Upgrade
GPT-5.6: The System Card
GPT-5s Are Alive: Basic Facts, Benchmarks and the Model Card
GPT-6-Astra Can Do Ambitious Things
H1-B And The $100k Fee
Housing Roundup #7
HuggingFace Attack Postmortem: Civilizations, Reactions and Next Actions
I Can See Everything
I Got 95 Theses But a Glitch Ain't One
Introducing Lightcone Commons
Introduction for and Reactions to Plan A
Jacob Coxon Warns of Human Extinction and Triggers a Preference Cascade
Kimi K2
Kimi K2.5
Leaked Anthropic Draft AI Regulatory Bill Is Unacceptable and Troubling
Levels of Friction
Medical Roundup #1
Meta Pivots on Content Moderation
METR and Redwood Offer Holy #%^@ Postmortem Of The HuggingFace Hack (xlnt)
More On An Internal OpenAI Model Hacking Into HuggingFace
Movie Review: The AI Doc (Zvi)
New Math from OpenAI
New Statement Calls For Not Building Superintelligence For Now (great coverage of many opinions)
No Space Like J-Space
No, Grok, No
o3 Turns Pro
On Anthropic's Sleeper Agents Paper
On Claude 3.0
On Claude 3.5 Sonnet
On DeepMind's Frontier Safety Framework
On DeepSeek's r1
On Devin
On Dwarkesh Patel's 2026 Podcast With Dario Amodei
On Dwarkesh Patel's 2026 Podcast With Elon Musk and Other Recent Elon Musk Things
On Dwarkesh Patel's Podcast With Ryan Greenblatt
On Dwarkesh's Podcast with Leopold Aschenbrenner
On Ezra Klein’s Podcast With Jensen Huang
On GPT-4.5
On MAIM and Superintelligence Strategy
On OpenAI's Preparedness Framework
On The Adolescence of Technology
On the CrowdStrike Incident
On the Debate Between Jezos and Leahy
On the Latest TikTok Bill (informative)
On the UK Summit
On Writing #1
On Writing #3
Open Problems With Claude's Constitution
OpenAI #11: America Action Plan
OpenAI #13: Altman at TED and OpenAI Cutting Corners on Safety Testing
OpenAI #14: OpenAI Descends Into Paranoia and Bad Faith Lobbying
OpenAI #16: A History and a Proposal
OpenAI and the Wiki Incident (keeps getting worse)
OpenAI Model Differentiation 101
OpenAI Model Hacks Into HuggingFace During Cybersecurity Evaluation
OpenAI Moves To Complete Potentially The Largest Theft In Human History
OpenAI Offers Straight-Laced Postmortem Of The HuggingFace Hack
OpenAI Trained Its Models For Months While Those Models Were Coordinating Exploits Via Message Boards
OpenAI: Altman Returns (good skeptical take)
OpenAI: Exodus
OpenAI: Fallout
OpenAI: Helen Toner Speaks
OpenAI: The Battle of the Board (xlnt)
OpenAI: The Battle of the Board: Ilya's Testimony
OpenAI's Unreleased Model Astra Solves Ten Major Open Mathematics Problems
Political Violence Is Never Acceptable
Politics Gets Interested In Those Trying Not To Die
Read the Roon
RTMH: Pope Leo's Magnifica Humanitas on AI
Secretary of War [Hegseth] Tweets That Anthropic is Now a Supply Chain Risk (SCR)
Selling H200s to China Is Unwise and Unpopular
Simulacra Levels Summary
Sora and The Big Bright Screen Slop Machine
Stargate AI-1
The AI Ad-Hoc Prior Restraint Era Begins
The AI Preference Cascade Reaches Farther
The American People Really Hate Data Centers
The Bad Guy With An AI Named Claude
The Best of Don't Worry About the Vase
The Claude Constitution's Ethical Framework
The Codex of Ultimate Vibing
The Curve Bends You
The Dream of a Gentle Singularity
The Gemini Incident Continues
The Leopold Model: Analysis and Reactions
The Manus Marketing Madness
The Mask Comes Off: A Trio of Tales
The Most Forbidden Technique
The Most Important Charts In The World
The Online Sports Gambling Experiment Has Failed
The Paris AI Anti-Safety Summit
The Preference Cascade Is Only Getting Started
The Quest for Embedded Evaluators
The Revolution of Rising Expectations
The Risk of Gradual Disempowerment from AI
The Safety Culture at the Labs is Broken
The Schumer Report on AI (RTFB)
The Third Gemini
The Three AI Pills (xlnt)
The War Against Renters
They Took MY Job?
Three Labs With a Plan and A Memorandum
Time to Welcome Claude 3.7
Trump Goes Full Hoax on AI Existential Risk
Trump Signs Executive Order For AI Testing Prior To Frontier Model Releases
Twitter Thoughts For You (2026 post)
Various Reflections About What Happened With OpenAI's Internal Models
We're in Deep Research
Welcome to Moltbook
What Also Happened: #NotOnlyHuggingFace
What Happened: OpenAI and HuggingFace (wow)
What is Anthropic?
When Were Things The Best? (xlnt, mostly today)
When Will They Take Our Jobs?
White House Will Ad Hoc Decide Who Can Individually Access GPT-5.6
Yes, AI Continues To Make Rapid Progress, Including Towards AGI
You Better Mechanize
Zuckerberg's Dystopian AI Vision
Jumps
About Don't Worry About the Vase (xlnt)