grigio

31 ms·

121.
by grigio 7mo ago
EU destroyed all its industries (Olivetti, Ericsson, Nokia,..) and automotive isn't an exception
122.
Show HN: OpenCode Benchmark Dashboard(github.com)
1 points by grigio 7mo ago|0 comments
123.
OpenCode Benchmark Dashboard – compare different LLM providers / quants / models(grigio.org)
2 points by grigio 7mo ago|0 comments
124.
by grigio 7mo ago
if it doesn't run GNOME Mobile or KDE it isn't an alternative
125.
Accuracy vs. Speed in Local LLMs: Finding Your Sweet Spot(grigio.org)
1 points by grigio 7mo ago|0 comments
126.
by grigio 8mo ago
safety means it works for your gov but not for normies or enemy govs.. so it's impossible
127.
by grigio 8mo ago
Not your infra not your AI
128.
by grigio 8mo ago
yeah i still can't believe many people bought a mac mini just for the claw hype
129.
by grigio 8mo ago
is this a new series?
130.
by grigio 8mo ago
I just follow BlackRock if they stopped to fund green tech, it means the global warming is fixed /s
131.
by grigio 8mo ago
Because the States and the "Green" companies earn more money reselling the entire product again instead just the parts
132.
by grigio 8mo ago
+1 for voxtype with Whisper-base model it is quite fast an accurate
133.
by grigio 8mo ago
In theory I agree between OS/APPS separation but in practice i like the fact that i can use Debian+docker as server and Debian+wayland as desktop Non integrated apps can already be distribuited as .bin, debs, Flatpak or appimage, it se
134.
AMD OpenSIL and Open-Source Firmware Efforts for Confidential Compute(phoronix.com)
2 points by grigio 8mo ago|0 comments
135.
Ask HN: Do you use LLM memory features?
2 points by grigio 8mo ago|4 comments
136.
by grigio 8mo ago
used chromebooks like c720p, but usually older laptops do not have excellent battery life
137.
by grigio 8mo ago
nice project, it seems the only non-broken websites are Github and nono.sh
138.
by grigio 8mo ago
i like the idea is possible to run in a docker container?
139.
by grigio 8mo ago
it is interesting because of Memory, Cron and Telegram integrations, but most of the magic is done by the LLM model and Skills
140.
by grigio 8mo ago
No, Copilot pushed with force in Windows 11 is causing it
141.
by grigio 8mo ago
https://github.com/openai/whisper
142.
by grigio 9mo ago
2025: EU bans Huawei because is China tech 2026: EU bans BigTechs because is USA tech 2027: EU blackout
143.
Qwen3-TTS Works Also on CPU. No Nvidia GPU Required(youtube.com)
1 points by grigio 9mo ago|0 comments
144.
GLM4.7-Flash vs. other Local LLM. Tested also with OpenCode [video](youtube.com)
2 points by grigio 9mo ago|0 comments
145.
GLM4.7-Flash the new Local LLM king at 30B A3B and OpenCode?(grigio.org)
1 points by grigio 9mo ago|0 comments
146.
by grigio 9mo ago
Because yours Windows PC isn't yours
147.
by grigio 9mo ago
Exactly, also the fact that this container is build from a flake.nix instead of a Dockerfile it's quite unusual.
148.
by grigio 9mo ago
It seems cool, will it work in headless mode without X11/Wayland/.. ?
149.
by grigio 9mo ago
Can I also see my impact while I'm gaming and paying with credit card? /s
150.
by grigio 9mo ago
it seems overkill, I made a similar sandbox for opencode where the AI agent can do everything, also install software and it is done as normal user, only a directory is shared with the host if you want https://github.com/grig

More ›