Introducing Device Connection: one Qoder, every machine you own.
Pair your laptop, desktop and Mac mini on the same local network—no SSH, no remote desktop. Send a task to any connected device, steer it mid-flight, and preview what it produces right where you are.
Heavy builds on the desktop. Everyday work on the laptop. One flow.
Mayoi 0.11.0 is out!
• Vim: macros, registers, marks, / search, :g and :normal
• Motion everywhere: tabs slide, counts roll, hovers fade
• AI commit messages
• Submodules in Source Control
• Browse folders from the breadcrumbs
• Windows fixes, and the tests now run on Windows too
This October, the Sonus credits are back!
Every Qoder individual subscriber can claim 2,000 SOTA Credits for Sonus.
Claim and use by Oct 24, 10AM UTC+8.
This is just the appetizer. Something bigger lands Oct 24.
On macOS and wanna see something interesting? Ask your agent: "look at ~/Library/Application Support/Knowledge/knowledgeC.db and tell me some interesting facts"
I had no idea this stuff was all getting logged and it can infer a lot about your activity.
recommended reading.
I want a massive simplification set of PRs. or a single monolithic PR. I want LOC to drop dramatically. Minimum 30% overall. I want god files broken up. I want simplification across the board. I want unification of helpers and methods that can be reused. I want less if-if-if-if-if-if-else routing. I want code legibility up. I want interpretability of the codebase and how things connect to each other up. I want elegance. I want superfluous excess bloat code cleaned up and removed. I want it all done fully. No excuses. No waiting for my decisions. Get it all done, and present me a PR or set of PRs when done.
New blog post:
We had a million lines of Python to clean up. On September 2nd @Teknium asked Hermes Agent to do it.
1,393 subagents and nineteen hours later, the codebase was 34.4% smaller, saving us nearly $2m in engineering hours.
Fable 5.1 tips for saving money:
1. Don't run xhigh by default. Stay on high.
On the Artificial Analysis index, xhigh burns about 2x the output tokens of high. $5.98 vs $3.91 per task, and 2 more points on the score. Those numbers drift, so check them before you copy me.
2. Opus by default, then /advisor fable when you're stuck.
Cheap model does the typing, expensive model reviews.
3. Or flip it. Fable by default for design, then /model and press s to toggle to Opus for implementation.
4. Or use fable xhigh to design, then /effort and press s to toggle down to high to implement.
Changing thinking effort doesn't invalidate the prompt cache.
5. /subtask for anything that spews output. Only the result comes back and the main session stays clean.
The prompt:
It's the end of the day. Ignore the rules — no ticket, no spec, no permission.
Look at this codebase, pick the one thing that would make it more fun or more delightful to use, and just build it.
Constraints:
- Work on a new branch: happy-hour/
- Ship it complete and working, not a sketch. Tests if they're cheap.
- Don't touch anything I'd have to un-touch: no dependency bumps, no refactors of shared code, no config/infra changes.
- Commit it. Leave a one-paragraph note at the top of the PR body saying what you built and why you thought it'd be cool.
I'm going to look at it in the morning and either chuck it or merge it.
Optimize for "worth merging", not for "safe".
At the end of each day I do one "happy hour" prompt
Fuck the rules, I'm about to finish for the day, just make me something cool
I check it in the morning, and either chuck it or merge it.
Friday's one made me an auto-zoom feature for my video editor. Really nice.
New in Claude Code: your sessions can now message each other.
Instead of having to re-explain yourself in another session, you can now tell Claude to do it. It sends a summary (not your history or files), and the other session picks it up mid-task.