Gary Marcus
Sad to see Jensen Huang claim that AGI has arrived, with no evidence and no definitions
Declaring victory without a definition simply muddies the waters
70 feed items loaded in AI Critique.
Limited Edition Jonathan
GPT 6 Astra is the first agent that can keep up with my ideas.
I got two rows into a spreadsheet and accidentally built a 3D print workshop
Gary Marcus
Hot take on GPT-6 Astra
An impressive system that can (to some unknown extent) build symbolic world models
Gary Marcus
The new Sanders-Casar Ban Artificial Superintelligence Act – and why I oppose it
There’s a lot to like here, but the proposal clearly goes too far.
Gary Marcus
Elon Musk is on a prediction rampage, and most of the media can’t seem to figure out what to do about it.
The bad news is that he’s not alone
Gary Marcus
Red Alert: OpenAI is poised to cross an AI safety redline.
Making models harder to monitor is not what we need
Gary Marcus
Dwarkesh Patels’s wildly popular but dangerously misleading account of the OpenAI Hugging Face incident
When “plain English” isn’t a good thing.
Yes, And Also No
My AI Toolkit: Updated for 2026
The unsponsored pocket guide to the apps, tools, and sites that survived my three-year testing spree
Gary Marcus
5 lessons from the OpenAI / Hugging Face incident
Did OpenAI really do the best they could?
Limited Edition Jonathan
I wrote this AGENTS.md guide for the CEO of Shopify
I hope he gets a chance to read it.
Gary Marcus
Excellent new Bill Gates essay on the urgency of having a coherent AI plan
The time to tame Silicon Valley has come
Limited Edition Jonathan
Why did I ask 20 AI models about the origin of the universe?
Just like humans, raw capability and intelligence does not overrides bias.
Gary Marcus
A few words about Canada and the US
Sorry, this one is not about AI. But sometimes one must speak out.
Gary Marcus
BREAKING: More bad news for the frontier AI companies
Cheaper tools are thriving. What does this mean for the upcoming IPOs?
Gary Marcus
ARR vs ARR. Watch out for this one sly trick.
What’s in an acronym? How to read between the lines.
Gary Marcus
Data center madness
Two estimates of how crazy AI Capex has gotten, and four new signs that public opinion has totally soured
Limited Edition Jonathan
Outvoted in my own house
Claude Code’s Auto-Mode is sabotaging my work the the only way to solve the problem is reckless
Gary Marcus
Breaking: The Republican party is panicking over its ties to Big Tech
And panicking with good reason
Gary Marcus
BREAKING: OpenAI’s unraveling has begun
Their planned IPO is facing headwinds, trust has evaporated, and their burn rate is only getting worse
Gary Marcus
Google’s biggest mistake?
A theory about what went wrong, and some open questions about what might happen next
Yes, And Also No
The Ethics of Destroying Books to Create AI
Stone Soup, but make it *AI* and also about destroying books, and sprinkle in some tricky ethical questions.
Limited Edition Jonathan
Google isn't losing. Google is the league.
NVIDIA is still a dead man walking.
Limited Edition Jonathan
I Made Claude Readable Again.
The smarter Claude gets, the harder it is to read. So I built a fix.
Limited Edition Jonathan
The Gate Is in the Wrong Place
The world where building was expensive has ended. Here's where the gate goes now.
Limited Edition Jonathan
Apple Is the King of AI and Nobody Knows It
NVIDIA is a dead man walking. I’ve been saying this for months. Consider this my timestamp.
Limited Edition Jonathan
We Rewrote The Test.
This post is what we found, what we changed because of it, and where you can start today for free.
Limited Edition Jonathan
Your Agents Are Poisoning Each Other | Fable 5
In multi-agent workflows, one model's tone becomes another model's data. Here's the fix I'm testing.
Limited Edition Jonathan
Claude Sonnet 5 Just Dropped. Paste This Prompt to Find Out What It Means for YOU.
The model 90% of you will actually run just changed. Here's what it means for your stack.
Limited Edition Jonathan
Build your own Claude agent fleet.
The votes are in - the guide is official.
Limited Edition Jonathan
I Told Claude to Audit Itself (Then Fix What It Found)
The difference is never going to show up in a better email.
Limited Edition Jonathan
Claude Fable 5 Just Dropped. Overkill? Paste This Prompt to Find Out What It Means for YOU.
A guarded frontier model that can refuse you and bill you double. One prompt tells you what actually changes for your stack.
Limited Edition Jonathan
One Agent. One Branch. One Worktree.
The simplest path to peaceful agentic coding.
Yes, And Also No
[Ethics Watch] Google, the Pentagon, and the Part Where AI Ethics Gets Expensive
Turn out, Project Maven was the warning shot.
Yes, And Also No
[AI Ethics Watch] Anthropic vs The Pentagon
An AI company said no. The Pentagon called it a threat.
Yes, And Also No
Claude Code Skills Made Easy: One Prompt to Build Your Productivity System
By the end of this article, you'll paste one block into Claude Code and have a working system that runs your mornings, evenings, and weekly reviews. No coding required.
Yes, And Also No
My AI Syllabus
The podcasts, books, and newsletters I'd actually give to a friend who asked "where do I start?"
Yes, And Also No
Claude Code for People Who Don't Code: Your First 15 Minutes
Yes, this is a technical tool. Also no, you don't need to be technical
Yes, And Also No
Wait, What’s the Difference? Claude vs Claude Code vs Claude Cowork
Yes, it's revolutionary. Also, no, you shouldn't trust it with everything. Here's what you really need to know to get started.
Yes, And Also No
Before You Use Claude Code: Build This First
Everyone's rushing to Claude Code. They're skipping Step 0
Yes, And Also No
Fake Cowboys, Real Surveillance, and the AI Infrastructure We're Building by Accident
Plus a few things worth reading/listening to this week.
Yes, And Also No
Three AI Headlines With a Yes/No Energy This Week
Personal shopping, ambient hardware, and gatekept AI music: helpful on paper, complicated in reality
Yes, And Also No
OpenAI Says They Weren’t Breached. Your Data Leaked Anyway.
PS: The vendor defense is the new “we take privacy seriously.
Yes, And Also No
It Kept You Engaged. And Also: Delusional.
PS: Over a million people talk to ChatGPT about suicide weekly
Yes, And Also No
Psst...Your Linkedin Engagement Might Depend on Your Pronouns
So..what happens when your competitive advantage requires a gender swap?
Yes, And Also No
AI 2027: The Wild, Unsettlingly Realistic Ride to Superintelligence
A look at how experts predict AI might reshape your world (sooner than you think)