Mauve 👁💜 @mauve@mastodon.mauve.moe
- Pronouns
- they/them/it
- mauve+fedi@mauve.moe
- Matrix
- @mauve:mauve.moe
- Github/Gitlab/Discord
- @RangerMauve
Mod
Occult cyberpunk. Yap with me about decentralized systems, wearable computing, and biohacking.
Joined Apr 2022
@lvk Mostly for processing streams of data. e.g. I'm using them for a mic input -> ASR -> STT -> LLM loop and it was way easier than the original callback based flow.
Async Generators are by far my favourite #JavaScript feature.
slop dev
Local model was acting more clueless than usual and after stepping away and having a snack I found out it was due to a bug in my harness that kept clearing it's context.
Remember kids: The fastest code is the code you never write. #ProTip
Mauve 👁💜
boosted
This is the first genuinely operational littleFedi instance.
It has been online since 1st July, is single-user (just me), and runs on a Raspberry Pi Zero W powered by NetBSD, directly on its SD card.
It consumes just under 1W.
Its database is this size:
-rw------- 1 little wheel 109289472 Aug 26 15:14 littlefedi.db
That is, just a little over 100 MB. Yes, MB.
Its average CPU load is extremely low.
It is perfectly usable both from the web interface and from Mastodon API apps.
I have 179 followers and 169 followings, and it doesn't bat an eye.
I promise myself to use it more, and I will.
You don't need Big Tech to communicate with others.
You don't need an expensive data center to exist online.
Because we are people, and the bits are just extensions of our voice.
EDIT: littleFedi allows to transform posts into blog posts, having its own SSG (Static Site Generator). The one generated by this post is reachable by clicking here.
Mauve 👁💜
boosted
"Large/long-scale industry studies [of LLM use for programming] show a clear trend – output is up (more code, more commits, bigger diffs), but outcomes don’t reflect that trend. If anything, the average team is taking longer to ship worse software. If ever we needed proof that software development isn’t a production process."
https://codemanship.wordpress.com/2026/08/12/ai-software-development-what-does-the-data-say/
@aeva it'll be an amazing day when linux folks can recreate the utility of Procreate on an iPad Pro 😭💜
Mauve 👁💜
boosted
@argv_minus_one @murdoc @renardboy @KatyElphinstone my "party trick " back in the early 90's and while a sysop of a home BBS system was I could whistle the handshake of a 9600 band modem XD
They'd phone up, I'd answer get their handshake, I'd respond then they'd start transmission and I'd put the phone down XD
Kokofreakbean really lives up to their name. Really nothing like em out there.
Finding signal in a bunch of slop genned noise is a pain.
Remember kids: the best way to lie is to tell the truth!
@futureisfoss the meta right now is buying those mac minis with the "unified memory" so you can use the RAM with their GPU acceleration. Something like a 64 gig box should do the trick. I ended up getting a 128 GB DGX spark alternative cause I didn't want to worry about ram for a few years. 😝
@futureisfoss Hmm tbh it depends on what you're trying to do. I'd say for multi round coding taskd with longer contexts anything less than 27B won't cut it. And even then Qwen 3.6+ seems to be one of the few that actually cut it. For basic stuff like zero shot bash scripts or code snippets 8b models can work well enough, even the 4b gemma models can handle zero shot tasks if you keep their context small. Use vllm or llama.cpp instead of ollama for running inference.
Youtube search is such garbage lately. It's hard to actually get the videos for my search terms. And then harder to find something that's low effort and full of empty words. 🥲 Gotta publish my own blog post about it once I figure it out.
Snake food
My ball python missed its initial strike on its food and now it's too nervous to acrually eat it 😭 Sensitive baby.
@symbiogenic These aren't *amazing* but they're real and get the job done. i sleep with them on the daily. https://www.aliexpress.com/item/1005004258147797.html The mic is so so but not the worst.
Pissed off at how many supposed "bone conduction" headsets are actually "air conduction". One I considered had the usual bone conduction design in the marketing but very obviously had shitty speaker grates in the customer review photos.
- Pronouns
- they/them/it
- mauve+fedi@mauve.moe
- Matrix
- @mauve:mauve.moe
- Github/Gitlab/Discord
- @RangerMauve
Mod
Occult cyberpunk. Yap with me about decentralized systems, wearable computing, and biohacking.
Joined Apr 2022
