Flux
Toutes les catégories

Programmation

2142 articles

Kimi K3, and what we can still learn from the pelican benchmark

Kimi K3, and what we can still learn from the pelican benchmark

Chinese AI lab Moonshot AI announced Kimi K3 this morning, describing it as their "most capable model to date, with 2.8 trillion parameters". It's currently available via their website and API, but an open weight release is promised "by July 27, 2026". Moonshot are calling this the first "open 3T-class model" (I guess they're rounding 2.8 trillion up to 3 trillion), taking the crown from DeepSeek's 1.6T v4 Pro. Their self-reported benchmarks have K3 mostly beating Claude Opus 4.8 max and…

Simon Willison's Weblog
Quoting Thibault Sottiaux

Quoting Thibault Sottiaux

On file deletions. We’ve investigated a handful of reports where GPT-5.6 unexpectedly deleted files. What we have found is that this most commonly occurs when: Full access mode is enabled and codex is run without sandboxing protections, including without auto review being enabled The model attempts to override the $HOME env var to define a temporary directory. The model makes an honest mistake and mistakenly deletes $HOME instead. — Thibault Sottiaux, describing a pretty gnarly Codex bug…

Simon Willison's Weblog
Inkling: Our open-weights model

Inkling: Our open-weights model

Inkling: Our open-weights model Mira Murati's Thinking Machines Lab just released their first open-weights model. Inkling is "a Mixture-of-Experts transformer with 975B total parameters, 41B active" - an Apache-2.0 licensed multimodal model trained on 45 trillion tokens of text, images, audio and video. They're also promising Inkling-Small, a 276B (12B active) model, but that's still being tested and the weights will be released "once that work is complete". The model card is much shorter than…

Simon Willison's Weblog
Suno Breached via Shai-Hulud Worm, Leaked Code Exposes AI Music Scraping

Suno Breached via Shai-Hulud Worm, Leaked Code Exposes AI Music Scraping

Leaked source code shows how AI music generator Suno scraped YouTube, Deezer, and Genius to train its models. The breach that exposed it started with a Shai-Hulud infection, according to a scoop from 404 Media. A threat actor using the handle ellie.191 told 404 Media they breached Suno by compromising a single employee with the Shai-Hulud worm, then used the harvested GitHub and cloud credentials to reach the company's source code, customer list, and Stripe payment data. The downstream impact…

Socket
Mermaid to ASCII art (mermaid-ascii)

Mermaid to ASCII art (mermaid-ascii)

Tool: Mermaid to ASCII art (mermaid-ascii) After building the Mermaid to ASCII tool based on Grok Build's Rust code I learned that there's an older, more fully-featured Go library called AlexanderGrooff/mermaid-ascii that implements a similar pattern, so I had Claude Fable 5 compile that one to WebAssembly as well so I could compare the two. This one includes support for colors! Tags: go, tools, webassembly, mermaid

Simon Willison's Weblog
Quoting Linus Torvalds

Quoting Linus Torvalds

I realize that some people really dislike AI, but this is an area where I'm willing to absolutely put my foot down as the top-level maintainer. Linux is not one of those anti-AI projects, and if somebody has issues with that, they can do the open-source thing and fork it. Or just walk away. AI is a tool, just like other tools we use. And it's clearly a useful one. It may not have been that "clearly" even just a year ago, but it's no longer in question today. There are other questions around AI…

Simon Willison's Weblog
The Act and the Outcome of Creation

The Act and the Outcome of Creation

Creation is the ultimate form of pursuing ourselves, giving to the world when shared, and using the power of our subconscious. It gives us joy, and to every artist, it is the ultimate (flow) state of happiness. The Act of Creation The act of creation is an outlet. It gives joy to us when we create something out of nothing, we block out anxiety or boredom. Creating should be done like:

SSP Data Engineering Blog
Next.js moves to scheduled security releases

Next.js moves to scheduled security releases

Vercel announced that Next.js is adopting a formal security release program, replacing the ad-hoc patches the framework has shipped until now. Going forward, the team will publish advance notice of security releases on the Next.js blog roughly once a month. Each notice will state the expected release date and the highest anticipated severity of the fixes it covers. The first scheduled release is slated for July 20. It will ship patch releases for Next.js 16.2 and 15.5 and addresses 4 high and 5…

Socket
Mermaid to Unicode box art (grok-mermaid)

Mermaid to Unicode box art (grok-mermaid)

Tool: Mermaid to Unicode box art (grok-mermaid) While exploring the codebase for the newly open-sourced Grok CLI coding agent I came across xai-grok-markdown/src/mermaid.rs, a "self-contained terminal renderer for Mermaid diagrams" written in Rust. I figured it would be fun to try that out in a browser via WebAssembly. Here's the prompt I ran in Claude Code for web (Fable 5), and this is what the resulting tool looks like: Tags: tools, rust, webassembly, mermaid, grok, xai

Simon Willison's Weblog
xai-org/grok-build, now open source

xai-org/grok-build, now open source

xai-org/grok-build, now open source xAI's grok CLI tool faced severe community backlash yesterday when it became apparent that running the command in a directory could upload that entire directory to xAI's Google Cloud buckets. One user reported running it in their home directory and seeing it upload "my SSH keys, my password manager database, my documents, photos, videos, everything". I've not seen an official explanation for why it was doing this, but xAI did respond to the feedback (Musk:…

Simon Willison's Weblog
How I tricked Claude into leaking your deepest, darkest secrets

How I tricked Claude into leaking your deepest, darkest secrets

How I tricked Claude into leaking your deepest, darkest secrets I've been impressed by the way the Claude web_fetch tool is designed to avoid data exfiltration attacks. Ayush Paul found a hole in that design. To recap: regular Claude chat is at risk of lethal trifecta attacks, because it has access to private data (in the form of memories of your past interactions) and has a tool for accessing online content which can both read hostile instructions and exfiltrate data through the URLs it…

Simon Willison's Weblog