Yes. They record almost every talk. You’ll find the relive (unedited recording of the live stream) on https://media.ccc.de and some time later the edited recording also on media.ccc.de and also on YouTube.
There also also live translations to English (and some other languages) for talks in german.
It ends at 1am their time. The conference itself never stops there are people there around the clock. (I wish I could go!! I went to 28c3 in Berlin and 29c3 in Hamburg they were amazing)
Probably. In previous years there were streamdumps available immediately somewhere (though figuring out where took me usually a day or so) and a re-live version on media.ccc.de a bit later. (Usually hours, but from time to time a day or so.)
See also https://github.com/jj-vcs/jj/wiki/JJ-Con-2025 for a list of all the talks, their slides, and some additional notes. I think the slides are also linked from youtube, but the website definitely has them.
> Find the (minimal) AI prompt that lead to the generation of this article: [...]
The minimal AI prompt that likely led to the generation of this article could be:
"Write a technical blog post comparing JPEG XL and AVIF image formats, focusing on how JPEG XL's bit-depth-agnostic float-based encoding is superior to AVIF's integer-based approach, with emphasis on perceptual quality, HDR handling, and workflow simplicity."
This prompt captures the core elements:
- Technical comparison between two image formats
- Focus on JPEG XL's unique "ignoring bit depth" design
- Emphasis on perceptual vs. numerical quality
- Discussion of HDR and workflow benefits
- Tone and structure matching the published article
The prompt would have guided the AI to produce content that:
1. Explains the technical difference in encoding approaches
2. Demonstrates why JPEG XL's method is better
3. Provides real-world implications for users
4. Maintains the author's voice and technical depth
5. Follows the article's structure and emphasis on "perceptual intent" over bit precision
This is so meta, we're using AI to generate feedback loops between a prompt, the AI generated content, using AI to recreate the prompt used to generate the content, etc. Spiraling to unreadable slop - unreadable to real humans anyway.
Soon enough the AI will invent a format for communicating with instances of itself or other AIs so that they can convey information that a client AI can translate back to the user's personal consumption preferences. Who needs compression or image optimization when you can reduce a website to a few kB of prompts which an AI engine can take to generate the full content, images, videos, etc?
I find JJ to make things much easier for me than Git:
- No staging needed, edit your commit directly in your working directory
- Rebasing is automatic, I regularly edit historic commits and the messages
- jj absorb will automatically split and merge hunks into related ancestry commits
- Defer conflict resolution; many conflicts are resolved later by splitting/squashing/moving commits/hunks further (possible only for first-class conflicts)
- No interactive rebase needed, since jj commands automatically rebase while editing historic commits, it feels like constantly living within an interactive rebase
- Easily refer to the same revision (change) across rebase commands by using a jj change_id, without scribbling down Git commit hashes that change across a rebase
- Easily undo (+redo) your last commands through the jj operation log without having to backup the entire git repo at each step
And jj-fzf (https://github.com/tim-janik/jj-fzf/) gives me every useful jj command with a single hotkey (e.g. Ctrl-P for push, Ctrl-N for a new commit with 1+ parents, Alt-B to edit bookmarks/tags, Alt-R to rebase, Alt-Z for undo, …) while browsing the log with commit+diff preview.
It is not too surprising that many HTML or JS based watermarking schemes are very easy to remove, some even via a script. For watermarking schemes to be effective, the watermark payload needs to be embedded in the content. Here's one approach that works for images:
https://github.com/tim-janik/imagewmark/
I hacked up a slim alternative localpilot.js layer that uses llama-server instead of the copilot API, so copilot.el can be used with local LLMs, but I find the copilot.el overlays kinda buggy...
It'd probably be better to instead write a llamapilot.el for local LLMs from scratch for emacs.
For a quick impression, the jj-fzf README has a number of screencasts that showcase tasks easily accomplished with jj (using fzf to preview the jj log and key bindings to invoke jj commands):
We are experiencing interruptions in multiple public GitHub services. We suspect the impact is due to a database infrastructure related change that we are working on rolling back.
