Skip to content

ViralLoop

Your next post already worked.

Loading the post

Remix this post

SaveShare

Your note

Private to you, and used when you remix this post.

Viral score

Why it worked

The mechanics

Hook
Style
Tone
Written for

Analysis by ViralLoop — not written by the original author.

Simon Willison on Science · ViralLoop

Post by Simon Willison on X

Simon Willison@simonw

These papers are fascinating, but my favorite thing about them is they aren't PDFs! They're glorious mobile-friendly web pages with interactive diagrams. I hope everyone else who publishes papers takes note, this us a much better way to share research t.co

Screenshot of a multilingual language model visualization showing antonym prediction across three languages. Left panel shows English with prompt "The opposite of 'small' is'" predicting "large". Middle panel shows Chinese prompt "小"的反义词是" predicting "大 (zh: big)". Right panel shows French prompt "Le contraire de "petit" est" predicting "grand (fr: big)". Above shows activation analysis with token predictions and highlighted instances of "contraire" in French text.

1.1K likes20 repliesPosted Mar 27, 2025

View the original on X Written by Simon Willison. Captured Sep 4, 2026.

Remix this post

Posts built the same way

Now in research preview: routines in Claude Code. Configure a routine once (a prompt, a repo, and your connectors), and it can run on a schedule, from an API call, or in response to an event. Routines run on our web infrastructure, so you don't have to keep your laptop open.

Fable 5 is state-of-the-art on nearly all tested benchmarks, with exceptional performance in software engineering, knowledge work, scientific research, and vision. The longer and more complex the task, the larger Fable 5’s lead over our other models.

Bryan Catanzaro

So proud of DLSS5: Fully generative neural rendering, in real-time, in real games. Mind-blowing realism. A whole new generation of real-time graphics. A decade of continuous research and development. Coming soon to PCs everywhere. 💚

I think China's second DeepSeek moment is here. This AI agent called 'Manus' is going crazy viral in China right now. Probably only a matter of time until it hits the US. It's like Deep Research + Operator + Claude Computer combined, and it's REALLY good.

New Anthropic research: A global workspace in language models. Of everything happening in your brain right now, only a tiny fraction is consciously accessible—thoughts you can describe, hold in mind, and reason with. We found a strikingly similar divide inside Claude.

New Anthropic research: Natural Language Autoencoders. Models like Claude talk in words but think in numbers. The numbers—called activations—encode Claude’s thoughts, but not in a language we can read. Here, we train Claude to translate its activations into human-readable text.

More from Simon Willison

A few notes on Anthropic's new Claude Fable 5.1 - with Max thinking level I got the best SVG pelican I've had from any Anthropic model (at a hefty cost of $3.30!), which I then had it animate t.co

The "introductory pricing" for the 3.7 Flash model is really weird It's scheduled to double in price on December 31, 2026, but who would anticipate still using this model five months from now? Especially since 3.6 Flash came out just three weeks ago!

I really hope we get details from @OpenAI on the task that as specifies to their rogue agent I'm guessing it was given the full ExploitGym suite and told to solve it, with an option to run 5.6-Sol subagents as part of the exercise

This is detailed, fascinating and answers all sorts of open questions I'd love to know more about the "unsecured public code-evaluation sandbox hosted on a third-party provider's infrastructure" that the agent used to stage its attack against HF after it broke out of OpenAI

Tucked away in this article is an appeal to the AI skeptics to PLEASE stop writing off stories like this OpenAI accidental exploit of Hugging Face as a dishonest marketing trick Frontier models can find and exploit vulnerabilities now, it helps nobody to pretend that they can't!

Anyone know if ChatGPT Codex (in the new ChatGPT desktop app) is a strict superset of ChatGPT Work? Liked if you're a software engineer who isn't intimidated by Git features is there any reason you'd ever want to switch to ChatGPT Work Mode?

Your note

Private to you, and used when you remix this post.

41

Viral score

Based on absolute engagement — the platform doesn’t expose follower counts here, so large accounts score higher.

From likes, comments.

Why it worked

Opens with a short, self-contained line that is easy to take in at a glance.
Written from experience rather than as advice, which is harder to argue with.
Short enough to read in full without expanding, which removes a step before the point lands.
Short sentences keep the pace up, so the reader never has to re-read a line.

The mechanics

Hook: Plain observation
Structure: Hook → Landing
Style: Conversational paragraphs at a steady pace.
Tone: personal, energetic
Written for: Founders, marketers and operators

Analysis by ViralLoop — not written by the original author.

Topics