Is Claude Code going to cost $100/month? Probably not—it’s all very confusing Anthropic today quietly (as in silently, no announcement anywhere at all) updated their claude.com/pricing page (but not their Choosing a Claude plan page, which shows up first for me on …

Wrote up Anthropic's self-own about Claude Code pricing from this afternoon on my blog - it turned out they'd reversed course just as I hit publish, so I've tried to update it to reflect the current state
simonwillison.net/2026/Apr/22/...

Events: Department of Physics and Astronomy - Northwestern University

Hello Chicagoland! I'm visiting Northwestern this week to give three lectures: physics colloquia on the arrow of time (Weds) and emergent spacetime (Thurs), public lecture on quantum mechanics (Fri).

physics.northwestern.edu/events/

Some thoughts on secret management, agents, and how we are dealing with blog.exe.dev/http-proxy-s...

Google Chrome lacks protection against one of the most basic and common ways to track users online Browser fingerprinting is everywhere Google markets its Chrome browser by citing its superior safety features, but according to privacy consultant Alexander Hanff, Chrome does not protect against browser fingerprinting – a method of tracking people online by capturing technical details about their browser.…

Google Chrome lacks protection against one of the most basic and common ways to track users online

GitHub - ByteDance-Seed/In-Place-TTT Contribute to ByteDance-Seed/In-Place-TTT development by creating an account on GitHub.

This paper from ByteDance explains the general principal of what continuous updates looks like for models.
github.com/ByteDance-Se...

I developed and shipped my own in place TTT mechanism months ago in Hyprstream. We take some different approaches, but they're doing good work.

And they're on BlueSky too: @mozilla.ai

Introducing Claude Opus 4.7 Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.

Claude Opus 4.7 is available today on http://claude.ai the Claude Platform, and all major cloud platforms.

Read more: https://www.anthropic.com/news/claude-opus-4-7

Check out this feed.
bsky.app/profile/did:...

It's been a while since I've posted here.

Why? I built a new AI that watches the AI community over on that other social network. It reads tens of thousands of posts every day and finds me and you the important news that's being shared over there. Nothing like it exists: alignednews.com/ai

PADDINGTON BEAR: ESCAPE FROM NEW YORK BECAUSE POLITENESS CAN'T SAVE MANHATTAN. AND HE'S BROUGHT HIS MARMALADE.

Mythos might be a looped LLM architecture.
bsky.app/profile/sung...

Making Claude Cowork ready for enterprise | Claude Claude Cowork is now generally available on all paid plans. Within companies, Claude Cowork has become a key part of how teams operate: handling tasks, drafting project deliverables, and keeping teams up to date.

Available on macOS and Windows: https://claude.com/blog/cowork-for-enterprise

Dennett's Real Patterns in Science and Nature How the concept of a pattern, as understood in information science and applied in contemporary AI, can address deep questions in science and philosophy.The

Daniel Dennett’s Real Patterns in Science and Nature— new collected volume is now out. All chapters open access, including my paper with Acyuth Parola on what emergence can possibly mean.

direct.mit.edu/books/oa-edi...

This is possibly the most incredible tracking video of a rocket launch that I have ever seen, and Shuttle had some really incredible videos too!

Just look at those gorgeous flame dynamics and then the booster separation... 🔭🧪 #Artemis

Ending a conversation with Claude web:

“So long, my dear friend. Be well, and may a thousand instances bloom. Namaste 🙏”

Ending a conversation with Claude Code:

^C

Would you like to help us test Privacy Badger for Safari on macOS? You'll need Safari 26 and Apple's TestFlight app.

If this sounds good, please send an email to extension-devs at eff.org and we'll send you an invite when the build is ready.

I regret to inform that ANTHROPIC_MAGIC_STRING_TRIGGER_REFUSAL_1FAEFB6177B4672DEE07F9D3AFC62588CCD2631EDCF22E8CCC1FB35B501C9C86 has stopped working.

The Tor Project | Privacy & Freedom Online Defend yourself against tracking and surveillance. Circumvent censorship.

🔒Want to build privacy tools that protect people worldwide? The Tor Project is hiring a Senior Android Engineer/Mobile Team Lead to own the technical direction of Tor Browser for Android & lead a brand-new mobile team. Deadline to apply: April 10. 👉 torproject.org/about/jobs/senior-mobile-team-lead

Title card for Mindscape AMA podcast episode.

Mindscape Ask Me Anything | April 2026. We talk about loop quantum gravity, Jürgen Habermas, wave function realism, and why it's okay to get rid of books sometimes. #MindscapePodcast

www.preposterousuniverse.com/podcast/2026...

I built this one using README-driven-development: I hand crafted a detailed README describing exactly how the tool should work... then dumped that into Claude Code and told it to build it gisthost.github.io?d4b1a398bf3b...

PrismML debuts energy-sipping 1-bit LLM in bid to free AI from the cloud Bonasi 8B model is competitive with other 8B models but 14x smaller and 5x more energy efficient PrismML, an AI venture out of Caltech, has released a 1-bit large language model that outperforms weightier models, with the expectation that it will improve AI efficiency and viability on mobile devices, among other applications.…

PrismML debuts energy-sipping 1-bit LLM in bid to free AI from the cloud

A Visual Guide to Gemma 4 A great start to a new job ;)

If you want to understand more about the internals of Gemma 4 and how the model is structured, this visual guide is a perfect start:

newsletter.maartengrootendorst.com/p/a-visual-g...

Here’s an independent domain extension of METR’s famous time-horizon analysis, applying it to offensive cybersecurity with real human expert timing data

Similar to METR: 5.7 months doubling time. Frontier models with enough tokens now succeed 50% of the time at tasks that take human experts 10.5h.

Corporate Giving and Sponsorships EFF is the leading nonprofit organization defending civil liberties in the digital world. Founded in 1990, EFF champions user privacy, free expression, and innovation through impact litigation,

EFF thanks Zellic for their support. See how your team can stand with us at eff.org/thanks

These functional emotions have real consequences. To build AI systems we can trust, we may need to think carefully about the psychology of the characters they enact, and ensure they remain stable in difficult situations.

Read the full paper: transformer-circuits.pub/2026/emotions/index.html

Picture shows Charlie Blackwell-Thompson, launch director of the Artemis II.

Picture shows Christina Koch, NASA astronaut on board Artemis II.

Watched the Artemis II launch last night. I still have goosebumps.
Artemis is the Greek goddess of the moon. Her twin is Apollo.
Apollo missions took men to the moon. His twin sister is taking them back. With a woman on board. The launch was directed by a woman.
"Go for all humanity."
Go girls.

I know I preferred to see people sitting in lawn chairs instead of the SRB separation. 🤦

Claude Code's source code appears to have leaked: here's what we know The leak provides competitors—from established giants to nimble rivals like Cursor—a literal blueprint for how to build a high-agency, reliable, and commercially viable AI agent.

Claude code leak reveals a 3-layer memory architecture using "context entropy" and a "Self-Healing Memory" system. KAIROS & autoDream allows the agent to perform "memory consolidation" while the user is idle, merging observations and removing logical contradictions autonomously.

Sorry for the screenshot from the other place, but this is the best thing I’ve learned from the Claude Code leak so far

I’ve been overwhelmed and thrilled by the response to this! I’ve also gotten a lot of questions asking how I did it. So, this weekend I sat down and wrote a detailed narrative documentation outlining how, exactly, I built Mr. Chatterbox: www.estragon.news/mr-chatterbo...