andhuman

5 ms·

1.
by andhuman 5d ago
This benchmark expects models to not only do the happy path, but also edge cases, even though the ticket they give to the model doesn’t specify if. This to align more to real world tickets. Usually with benchmarks it’s the other way around:
2.
by andhuman 22d ago
Also, handing over the work to AI robs you of learning and understanding. Let’s say you write a ticket for a feature and that description is good enough that an AI agent can implement it from start to finish. The AI will then discover thing
3.
by andhuman 1mo ago
I don’t see why they just don’t allow a smaller model to answer the question while letting the bigger one vet it. The vetting can be asynchronous and can be delivered after a few seconds (if it’s an easy query). If it’s a hard query, the UI
4.
by andhuman 2mo ago
What surprised me is that caffeine gets converted to other compounds that _also_ binds to the same receptors thus causing you to feel less tired. And that compound has its own half life. Now I’m just guessing, but it could be that one is mo
5.
by andhuman 2mo ago
Why conscious thought as described in the article was thought to be signs of a delirious person was because that’s where you could find it in real life, is my guess. If someone blabbing about without a listener, they’re insane. So to descri
6.
by andhuman 3mo ago
Finally an alternative to the big dogs that a company can use. People have been asking for a way to run the Chinese models from a trusted provider. Here GitHub delivered! The performance, if we trust the benchmarks, put it at Sonnet 4.6. Le
7.
by andhuman 5mo ago
The Vibe CLI is really bad on Windows, sure they don’t officially support it, so can’t blame them, but a FYI for anyone wanting to try it. It can’t get find and replace right.
8.
by andhuman 5mo ago
I have good experience with Mistral Vibe.
9.
by andhuman 5mo ago
I just listened to the Matrix OST and that one really gets me into a coding mood!
10.
by andhuman 6mo ago
Oh man, this was the release I wanted to link, as it has a new feature (tiled follow link) that I actually started using right away. A new browser feature I find useful didn’t happen that often for me, so I got excited.
11.
Vivaldi 7.8 Is Here(vivaldi.com)
3 points by andhuman 6mo ago|2 comments
12.
by andhuman 6mo ago
Be kind to others, but start with being kind to yourself.
13.
by andhuman 6mo ago
Isn’t llama.cpp’s implementation of Qwen 3.5 better, or am I misinformed?
14.
Qwen 3.5 small models out(huggingface.co)
18 points by andhuman 7mo ago|3 comments
15.
by andhuman 7mo ago
What helped me get a better understanding of oauth and oidc, was this YouTube video [0]. [0]: https://youtu.be/996OiexHze0
16.
by andhuman 7mo ago
Same setup I’m using! Parakeet and pocket turbo. It’s feels good enough for daily usage.
17.
by andhuman 7mo ago
Have you tired using the run action instead to clean the data? Win+r
18.
by andhuman 7mo ago
I have this nagging feeling I’m more and more skimming text, not just what the LLMs output, but all type of texts. I’m afraid people will get too lazy to read, when the LLM is almost always right. Maybe it’s a silly thought. I hope!
19.
by andhuman 8mo ago
One of the more annoying software that does this is the copilot Office 365 on the web. Every time (!) I open it, it shows a popup on how to add files to the context. That itself would be annoying, but it also steals focus! So you would be t
20.
by andhuman 8mo ago
I built this recently. I used nvidia parakeet as STT, open wake word as the wake word detection, mistral ministral 14b as LLM and pocket tts for tts. Fits snugly in my 16 gb VRAM. Pocket is small and fast and has good enough voice cloning.
21.
by andhuman 8mo ago
It’s uncanny good. I prefer it to pocket, but then again pocket is much smaller and for realtime streaming.
22.
by andhuman 8mo ago
Gave it four of my vibe questions around general knowledge and it didn’t do great. Maybe expected with a model as small as this one. Once support in llama.cpp is out I will take it for a spin.
23.
by andhuman 8mo ago
I’ve tried the voice clinking and it works great. I added a 9s clip and it captured the speaker pretty well. But don’t do the fake mistake I did and use a hf token that doesn’t have access to read from repos! The error message said that I h
24.
by andhuman 8mo ago
I’ve recently bought the LG with 4th generation OLED, and for me that works for long coding sessions (I use it for work). They shifted or did something with the pixel arrangement for this generation just for text legibility.
25.
by andhuman 9mo ago
Interesting experiment. I would hazard to guess that Google is on top when it comes to these sorts of things (spatial ability), then OpenAI and last Anthropic. I would like to see the same experiment using Google’s Live view or whatever it’
26.
by andhuman 9mo ago
The new LG panels are bright enough. I think they’re called 4th generation WOLED.
27.
by andhuman 10mo ago
This is big. The first really big open weights model that understands images.
28.
by andhuman 11mo ago
“ Starting at approximately 16:00 UTC, we began experiencing DNS issues resulting in availability degradation of some services. Customers may experience issues accessing the Azure Portal. We have taken action that is expected to address the
29.
by andhuman 11mo ago
I bet it’s DNS.
30.
by andhuman 11mo ago
Yup! I follow these and I also take walks, those can be helpful to relax and let the mind rest for a bit.

More ›