Mauve 👁💜 @mauve@mastodon.mauve.moe

@Lottie I am not blind but the inaccessible PDFs issue is such a pain for me. I use a screen reader for most books and being into stem means lots of equations that are un-pronounceable and figures that are only vaguely alluded to but not labelled. Don't get me started on tables that only make sense visually and lead to my reader listing random numbers for multiple minutes 💀

Low key wish I was a morning person.

Rewatching Spaceballs. Honestly I wish more movies were this goofy

Private heathcare should not be commercially viable. Until the poorest are covered the rich should be made to suffer or else nothing will change.

@dansup A similar model is etke.cc and their hosting. You can give them ssh access and they can set up their docker automations for running services. I fins it quite convenient.

@fleeky yeah sadly fine tuning needs way more resources than I have available. :( Maybe if I get an egpu?

@fleeky I only use locally runnable models so I have no clue how well it compares to the big corpo ones. 🤷 Sadly the code Ibwas using this on is a proprietary codebase so I can't share it. I ended uo deleting and rewriting myself like three times. It was some pretty basic cli arg parsing code tho

@cwebber@octodon.social Interesting. My main frustratiom is thatthe code it write just isn't anywhere near the quality I need and it generates stuff that barely makes sense. I've mostly been exploring this as an alternative to speech to text coding and for doing some quick formatting but so far it's just sucked. Even very junior devs need less hand holding and reviews

Even though I still don't have screen sharing or a KDE dark theme working, I at least have an Emoji picker via wofi-emoji on my niri setup. 💜

https://github.com/Zeioth/wofi-emoji

https://github.com/YaLTeR/niri/

@hermeticvm@convo.casa Cool to hear ty for sharing. Is your code published somewhere I may read?

Today I was trying to get it to generate some cli commands in Node.js (this is with phi3+continue via ollama) but t has trouble adhering to the specific syntax and structure I wanted it to use.

I found a big issue for me has been hallucinated packages or APIs within packages. Maybe doing multi step stuff where inject context from api docs could help there.

@hermeticvm@convo.casa Yeah I'm using continue.dev with ollama and phi3

Previously I used some 7B models that were just too slow.

The starcoder model has been giving okay suggestions once in a while but is super slow still. Deepseek seems to be too dumb to do anything useful.

@cwebber@octodon.social Is this from personal experience and finding out or second hand from what others have said?

Are folks using #ai copilots for generating code successfully? Mine gets confused on syntax a bunch but it's a local model which is general rather than code specific. Like I wonder if I am prompting it poorly or something