mythz

10 ms·

151.
by mythz 1y ago
Was looking forward to this for a long time as I think Linux could do with a clean break from the Legacy built around Gnome/KDE/X11. But it's taking so long to get to release my main concern is now that a small company doesn&
152.
Llms.py – Lightweight Open AI Chat/Image/Audio Client and Server(github.com)
1 points by mythz 1y ago|0 comments
153.
by mythz 1y ago
Team Qwen keeps cooking! qwen2.5VL was already my preferred visual model for querying images, will look at upgrading if they release a smaller model we can run locally.
154.
by mythz 1y ago
zoxide alongside fzf, eza, bat and starship were my killer CLI productivity tools I discovered after ditching Windows for Fedora. I have it aliased to `cd` so I don't really notice when I'm using it until moving to a Terminal that
155.
by mythz 1y ago
Kind of glad that I missed the Cursor hype then rapid slide into enshittification, since I didn't think its goal of replacing an IDE would be a winning strategy and instead went for IDE plugins/extensions AI assistants like Augmen
156.
by mythz 1y ago
> The trade-off is fewer features, but also less surface area and almost zero “lock-in” — you can literally view-source and drop it into a page. You should be able to drop in code dynamically on any existing website with JS Modules, I sh
157.
by mythz 1y ago
Using the full vue.js doesn't have a build step either. I started out developing UIs using petite-vue, unfortunately ended up rewriting it to use full Vue.js after running into too many of its limitations and issues, which given that i
158.
by mythz 1y ago
Also less power efficient, takes up more PCI slots and a lot of software doesn't support GPU clustering. Already have 4x 16GB GPUs which is unable to run large models exceeding 16GB. Currently running them different VMs to be able to m
159.
by mythz 1y ago
Really confused why the Intel and AMD both continue to struggle and yet still refuse to offer what Nvidia wont, i.e. high ram consumer GPUs. I'd much prefer paying 3x cost for 3x VRAM (48GB/$1047), 6x cost for 6x VRAM (96GB/$
160.
by mythz 1y ago
I mostly use neovim for editing files remotely and have come to miss AI Autocomplete but was pleasantly surprised to find my preferred AI IDE Tool also maintains a Neovim plugin [1]. Not as many features as its VS Code and Intellij/Rid
161.
by mythz 1y ago
I've owned several laptops over the years and have come to regret each of my non Apple laptop purchases which never got more than 3-4 years life without some hardware failures, Surface was the worst which died after 10 months of low sp
162.
by mythz 1y ago
What's the issue with nginx-proxy? We've used it for years to handle CI deploying multiple multiple Docker compose Apps to the same server [1] without issue, with a more detailed writeup at [2]. This served us well for many years
163.
by mythz 1y ago
> How is Rider v. VS? Rider is far better than VS for everything apart from Desktop UI Apps and perhaps Blazor WASM hot reloading, which is itself far behind the UX of JS/Vite hot reloading, so I avoid it and just use Blazor static
164.
by mythz 1y ago
ok issue is with ollama as gpt-oss 20B runs much faster on 1x 5070 Ti with llama.cpp and LM Studio: llama-server = ~181 tok/s LM Studio = ~46 tok/s (default) LM Studio Custom = ~158 tok/s (changed to
165.
by mythz 1y ago
Same, was just after a small lightweight solution where I can download, manage and run local models. Really not a fan of boarding the enshittification train ride with them. Always had a bad feeling when they didn't give ggerganov/
166.
by mythz 1y ago
Getting great performance running gpt-oss on 3x A4000's: gpt-oss:20b = ~46 tok/s More than 2x faster than my previous leading OSS models: mistral-small3.2:24b = ~22 tok/s gemma3:27b = ~19.5 tok&#
167.
by mythz 1y ago
gpt-oss:20b is ~14GB on disk [1] so fits nicely within a 16GB VRAM card. [1] https://ollama.com/library/gpt-oss
168.
by mythz 1y ago
Good to see Node is catching up although Bun seems to have more developer effort behind it so I'll typically default to Bun unless I need it to run in an environment where node is better for compatibility.
169.
by mythz 1y ago
why not? const data = await (await fetch(url)).json()
170.
by mythz 1y ago
Yeah can confirm it looks great on my 4K
171.
by mythz 1y ago
I like it, works great in Fedora. Looks like a great addition to the available image editors on Linux. IMO image editing is one of the few areas lagging in Linux. Currently alternating between Pinta, Krita, Gimp and photopea.com. Will switc
172.
by mythz 1y ago
There was a time when it was unbelievably frustrating to navigate the bunch of marketing pages required to find the cost of a newly announced model, now I just look at OpenRouter to find pricing.
173.
by mythz 1y ago
Whilst it's not economically feasible to self-host, using premier OSS models like Kimi K2 / DeepSeek via OpenRouter gets you a great price with a fallback safety net of being able to self-host should the proprietary model Co'
174.
by mythz 1y ago
I'm definitely on the opposite side where despite being in the nascent era of code assisting agents, they've become a productivity multiplier and I shudder thinking about how productive my last 25 years of programming could'v
175.
by mythz 1y ago
Many years ago when I heard Munich City was switching to Linux, I remember thinking it was a reckless decision that was doomed to fail, and turns out it eventually did as they announced in 2020 they were moving back to Windows. Which is in
176.
by mythz 1y ago
I've been turned off with OpenAI and have been actively avoiding using any of their models for a while, luckily this is easy to do given the quality of Sonnet 4 / Gemini Pro 2.5. Although I've always wondered how OpenAI could
177.
by mythz 1y ago
Cool, I've tried a few but settled on using EraX NSFW to do the same.
178.
by mythz 1y ago
Using just an Ollama VL Model (gemma3/mistral-small3.1/qwen2.5vl) or a specific library?
179.
by mythz 1y ago
Good value but a 12GB card isn't going to let you do too much given the low quality of small models. Curious what "home AI" use cases small models are being used for? It would be nice to see a best value home AI setups under
180.
by mythz 1y ago
Good to see with that war chest they'll be competing in this space for some time yet - which we'll need more of to keep the pricing down (aka subsidized). Never tried Cursor since I'm not prepared to leave my IDEs, but still

More ›