Valentin

Dev Web et prof d’info, créateur de framework et de librairies en tous genres. Keyboard lover ⌨️🥰

[Valentin @c4software]

Petite expérimentation de l'après-midi, intégration d'un système de « réécriture / résumer » via LLM dans mes supports de cours.

Évidemment, l'intégration est « On Device » pour utiliser #llama 3.2 3B.

L'utilisation de #WebLLM est vraiment très simple !

https://t.co/nqd7Yfq6Ap

[Valentin @c4software]

@pjournel L'idée c'est plutôt un nœud mobile. Sans forcément de gps ni d'écran.

[Valentin @c4software]

@vaab @FrameworkPuter I have the same issue, but I have a Samsung SSD 990 Pro (only when resuming from hibernate)

[Valentin @c4software]

@chokotrix @outsource_ It's seems we have gain with mlx, I will try with my M3 pro (same bandwidth as the Strix) to see.

[Valentin @c4software]

@chokotrix @outsource_ Interesting, I've never read this, but it seems logical, you're right.

[Valentin @c4software]

@FrameworkPuter Hey, I'm owning a Framework 13 (HX 370). When the upgrade kit will be available, do I need to upgrade the screen too? I'm only interested in the new bottom chassis and keyboard.

[Valentin @c4software]

@steeve Clairement mérité ! Je vous suis depuis le début : qualité constante, sujets bien traités et toujours très agréables à écouter. Bravo !

[Valentin @c4software]

@outsource_ I think in my setup it's not working (or do nothing)

[Valentin @c4software]

@outsource_ Interesting, in my setup, I have the speed of the 27b (~12t/s Strix Halo 128gb).

The draft model seems loaded (visible in the logs)

[Valentin @c4software]

@pierrelezan @LottoLabs Intéressant.

[John O'Reilly @joreilly]

New version of llmfit (https://t.co/iWbiqETlgZ) today that supports Qwen3.6 models

[Valentin @c4software]

@pierrelezan @LottoLabs Tu l'installe toi-même ?

[Valentin @c4software]

@FrameworkPuter Hi,

Is it possible to have only the bottom kit ?

The idea is to reusing the screen of my current 13 370 and transfer my current mobo to the now bottom case kit + keyboard.

[Unsloth AI @UnslothAI]

Qwen3.6-27B can now run locally! 💜

Run on 18GB RAM via Unsloth Dynamic GGUFs.

Qwen3.6-27B surpasses Qwen3.5-397B-A17B on all major coding benchmarks.

GGUFs: https://t.co/ykKgwh2zI9
Guide: https://t.co/ITLNq20WJp

[VideoLAN @videolan]

VLC is onboard the @NASAArtemis mission!

[Chaofan Shou @Fried_rice]

Claude code source code has been leaked via a map file in their npm registry!

Code: https://t.co/jBiMoOzt8G

[Daniel Hnyk @hnykda]

LiteLLM HAS BEEN COMPROMISED, DO NOT UPDATE. We just discovered that LiteLLM pypi release 1.82.8. It has been compromised, it contains litellm_init.pth with base64 encoded instructions to send all the credentials it can find to remote server + self-replicate. link below

[Marques Brownlee @MKBHD]

From the Macbook Neo video

[Valentin @c4software]

@XKamidh Moi, c'est pas mal je trouve.

[Valentin @c4software]

@ssamat Nice feature, working great here, but it would be great to have an option like in Dex to make the phone act as a touchpad.