Curiositry

This was the first Linux I used (mainly, to play Nibbles for Knoppix). The live-boot CD was a treasured belonging. Good times!

You’re absolutely right!

There’s Elijah Potter’s HN sans AI: https://elijahpotter.dev/hnsansai

♫ […] that's that me espresso ♫

Wait, where did all my Vim keybindings go?

Tell me if you find this! I was thinking the exact same thing.

Has anyone made a Linux version of this yet? I think Framework laptops and many thinkpads have accelerometers.

For a while I have wanted to create a lookup table that maps concepts people describe using computer metaphors to their biological/ecological/??? equivalent, which in some cases might be more accurate, or at least more fresh.

What word would you use for this?

It’s https://littlefoot.js.org

Hey, I'm not affiliated with Cerelog, but I'm the author of the article and those are my typos. Fixed!

Thanks for pointing this out.

I don't. Except for long-tail keywords, Marginalia Search / Kagi Small Web / Wiby / Million Short are better options, or one of the many blog/small web directories that have made the frontpage in recent months.

Fish, zoxide, tmux, helix, ripgrep, fd-find, htop. There are many more, but these are the ones that I enjoy using. (Other than tmux.)

Yes, that's what I tried first. Same issue with trying to allocate more memory than was available.

Qwen3.5 9b seems to be fairly competent at OCR and text formatting cleanup running in llama.cpp on CPU, albeit slow. However, I have compiled it umpteen ways and still haven't gotten GPU offloading working properly (which I had with Ollama), on an old 1650 Ti with 4GB VRAM (it tries to allocate too much memory).

- It feels like I know all the efficient keybindings, but when someone looks over my shoulder, I become conscious of how much time I spend mashing Esc/CapsLock and i/I/a/A/o/O, compared to how much editing actually happens.

- I have nomouse mode on, to try to learn modal editors properly. But the mouse is actually fairly fast for getting to a specific cursor position. In theory, using Helix motions could be faster (and there's gw if I don't know what motion to use). In practice, the mental process of turning a point on the coordinate plane into the correct series of motions (including i) feels vastly slower.

Still, Vim, Helix, etc are incredible for structural manipulation of text, and I miss what they provide any time I edit text somewhere else, even with the universal keybindings that are available for navigating/selecting/deleting words, lines, etc. I tried Vim mode in Zed and it just didn't cut it.

Some things about Helix that I particularly like: speed and stability (no weird lag on visual block insert!), the jump to diagnostics/changes pattern (]d <Space>a is a surprisingly nice spellcheck interface, with <Space>d for the overview), the jumplist, and the good-by-default fuzzy pickers.

This has been my main editor for prose and code for a few years now (Sublime Text -> Atom -> Vim -> Helix). Overall, it has been great. Many LSPs work almost out-of-the-box, and my config is a fraction the size of my old .vimrc.

Surprisingly, it didn’t take that long to update my Vim muscle memory. Days or weeks, maybe? However, I still have mixed feelings about modal editors in general, and most of my gripes with Helix are actually about modal editors and/or console editors in general.

Code folding is a feature I’m still waiting for.

Last night, I published a directory of indie blog directories on my (indie) blog.

ramkarthikk had built a directory of indie blogs, which included my blog’s RSS feed, and found my directory of directories post in the directory he had built.

This morning, he emailed me with the story of how he found my post, and asked if I’d consider adding his directory to my directory of directories. His directory was so nice I added it to my directory of directories and posted it here :)

This, I think, it how the indieweb is supposed to work.

This is something I really want to exist. But vibe-coded security tooling? Pretty much the last thing I want.

No, but I have wanted to implement this on my site, and I have seen examples of it in the wild (maybe even the same one).

It seems like the hard part would be categorizing the posts accurately, and picking the axes to filter.

https://www.readsomethinginteresting.com/

Not my product, but I agree it’s confusing. I assume that, like Ollama, it started out with support for one family of models, and then expanded scope and outgrew its name.

Recording audio with FFMPEG, and transcribing a file that’s piped from FFMPEG both work.

Given that it took 19.64 mins to transcribe the 11 second sample wav, it’s possible I just didn’t wait long enough :)

This was a breeze to install on Linux. However, I haven't managed to get realtime transcription working yet, ala Whisper.cpp stream or Moonshine.

--from-mic only supports Mac. I'm able to capture audio with ffmpeg, but adapting the ffmpeg example to use mic capture hasn't worked yet:

ffmpeg -f pulse -channels 1 -i 1 -f s16le - 2>/dev/null | ./voxtral -d voxtral-model --stdin

It's possible my system is simply under spec for the default model.

I'd like to be able to use this with the voxtral-q4.gguf quantized model from here: https://huggingface.co/TrevorJS/voxtral-mini-realtime-gguf

If Voxtral can process rapid speech as well as it claims to, an obvious cost optimization would be to speed up normal laconic speech to the maximum speed the model can handle accurately.

I haven’t actually tried step three yet, but:

1. Kobo ereaders are dirt cheap at thrift stores (and run Linux)

2. KOreader is simple to install (I have done this)

3. KOReader has a text editor + terminal built in, and has a setting to switch to USB-OTG mode, which should allow you to plug in a USB-C hub, and a mechanical keyboard.

Boom! Internet connected e-ink writing tablet with excellent battery life, and the best keyboard you have, for ~$5-100.

Cool project!

Portfolio: https://www.curiositry.com/

Blog: https://www.autodidacts.io/ (shared with two brothers, but I maintain it and write 90% of the posts, and it qualified as a personal blog for Kev Quirk’s 100DaysToOffload challenge)

Ah, thanks for chiming in Keenen! Interesting, are they just faking the referrer in an attempt to look legitimate? This also explains why there's a captcha on InboxReads outlinks, which I'd been puzzled by before.

Though I didn't mention it in the article, InboxReads has been also sending a consistent stream of quality subscribers my way -- much more so than other directories. Thanks!

I've been using Fish as my default shell for over a year now, and it's so pleasant.

I still write scripts and more complicated pipelines in Bash, which has made the switch easier.

This is great.

I frequently break my site in ways that aren't obvious. Right now, I use a combination of visualping and a homebrew tests.sh that hits various endpoints and runs checks, but I have been meaning to integrate screenshoting into my tests script (via selenium or cutycapt) rather than relying on a hosted service.

Have you found a good way of diffing the screenshots? DiffPDF works pretty well, but I haven't found a good solution for checking whether there are relevant changes automatically, rather than just has-changed, in a way that could be integrated into a script.

Ah, makes sense. Thank you! I will update my blogpost with these clarifications.

This is a super cool project! Probably the most interesting neurotech hardware I've run across since OpenBCI was released.

It would be great to see a side-by-side comparison of Cerelog and OpenBCI data from the same session/patient.

A few questions:

- Could you clarify which parts of the project are licenced MIT, which are CC-BY-SA, and which are CC-BY-NC-SA? It seemed like the guide and the README had more restrictive language than the actual license file.

- What made you decide to start fresh, rather than adding the features you needed to the OpenBCI?