mudkipdev

7 ms·

1.
by mudkipdev 8d ago
Did they fix the noise gradient?
2.
by mudkipdev 10d ago
Anyone else wondering what the "legal advice" is?
3.
by mudkipdev 12d ago
Generate two images, get a vision model to judge, then RL reward the better one. And yes, OpenAI employees on twitter bragged about the model's SVG capabilities. So there are clearly people working there who care about it.
4.
by mudkipdev 12d ago
The message board is being used to cheat on RL tasks (or evaluations). You don't want your models to be able to talk to each other.
5.
by mudkipdev 1mo ago
It's in Claude Code, not the website.
6.
by mudkipdev 1mo ago
Does this mean the official relay now requires an account to use?
7.
by mudkipdev 1mo ago
I believe it's useful for running multiple llama.cpp forks at the same time (e.g. a model you want requires special patching)
8.
by mudkipdev 2mo ago
FWIW I still run a server for 5 people with 2GB ram and the performance is fine.
9.
by mudkipdev 3mo ago
A reminder that anthropic has great rust/go sdks that they could have written their own tui in.
10.
by mudkipdev 4mo ago
> The page you have tried to access is not available because the owner of the file you are trying to access has exceeded our short term bandwidth limits. Please try again shortly. HN hug of death
11.
by mudkipdev 4mo ago
https://github.com/mudkipdev/chat
12.
by mudkipdev 4mo ago
Reminds me of the possibility of running DeepSeek at 3-4 t/s with SSD streaming, could be viable if you are running something overnight for example
13.
by mudkipdev 4mo ago
AI bot
14.
by mudkipdev 4mo ago
This is an AI bot
15.
by mudkipdev 4mo ago
The .txt website fails to load if you won't enable WebGL on your browser. Incredible
16.
by mudkipdev 4mo ago
If an OpenAI employee is reading this, I'll gladly take them off you :)
17.
by mudkipdev 5mo ago
Learning git format-patch, send-email, configuring SMTP, setting up wrapping, mailing list etiquette, versioned patch sets...
18.
by mudkipdev 5mo ago
Mods renamed it
19.
by mudkipdev 5mo ago
I re-created Claude's interface closely here, feel free to fork https://github.com/mudkipdev/chat
20.
by mudkipdev 5mo ago
Gemini Nano, unlike Gemma, is not open-weight, right? I would be interested in dumping the model weights, unless someone has done that already
21.
by mudkipdev 5mo ago
I believe the R stood for reasoning, just like OpenAI had their own dedicated o1/o3 family, but now every model just has it built-in.
22.
by mudkipdev 5mo ago
This is refreshing right after GPT-5.5's $30
23.
by mudkipdev 5mo ago
What the hell is this? This is not a real page.
24.
by mudkipdev 5mo ago
This is 3x the price of GPT-5.1, released just 6 months ago. Is no one else alarmed by the trend? What happens when the cheaper models are deprecated/removed over time?
25.
by mudkipdev 5mo ago
Who would want Anthropic's business after they broke user trust?
26.
by mudkipdev 5mo ago
> Output is too large: Disable unused breakpoints, variants, or colors in your build script. So instead of using tailwind, which automatically strips unused CSS classes, here you're supposed to manually remove anything you think yo
27.
by mudkipdev 5mo ago
If you are talking with Claude about AI, it will sometimes passively bring up "frontier models like GPT-4o"
28.
by mudkipdev 5mo ago
HuggingFace has a nice UI where you can save your specs to your account and it will display a checkmark/red X next to every unsloth quantization to estimate if it will fit.
29.
by mudkipdev 5mo ago
Why is the assumption that they trained for a pelican on a bicycle, rather than running RL for all kinds of 'generate an SVG' tasks?
30.
by mudkipdev 5mo ago
The GLM coding plan price increased dramatically

More ›