A K I O

[A K I O @z3syx]

@Blackfrost_AI @ViC305 ggufs?

[A K I O @z3syx]

@mrexodia does the mcp work with IDA Pro 9.3?

[A K I O @z3syx]

@johnnyrodgersis there's always people who take things so seriously.

[A K I O @z3syx]

@BottleCapAI thanks, we been waiting for this one.

[A K I O @z3syx]

@EdszSeren0 @ModelScope2022 in intelligence maybe no. in speed definitely yes.

[A K I O @z3syx]

Excited for the Qwen4 27B!

[A K I O @z3syx]

@zeldro_ @Badtheorylabs yes

[A K I O @z3syx]

@NVIDIAGeForce I want to run bigger local models in my pc
#RTXPowersPlay

[A K I O @z3syx]

@OrcaRouter add it🙏

[A K I O @z3syx]

@cat3lyst true

[A K I O @z3syx]

@calvinnwq @NousResearch congrats

[A K I O @z3syx]

@PrismML @ollama please add support for PrismML’s Bonsai 2 PQ2_0 / PTQ1_0 GGUF formats.

Right now Ollama fails on "ollama show" with:

unsupported tensor "output.weight" size overflows

and Hermes/tool-calling also gets rejected as “does not support tools.”

I'm using ollama 0.34.2 rn.

[A K I O @z3syx]

@witcheer I like this idea better cause for low end users like me, running on 64k context and the local models eating up 25-35k context right away before starting to write anything.

[A K I O @z3syx]

@oinkspinkpalace @EMostaque I don't think that'd be the case if u run it on low context window

[A K I O @z3syx]

@ivanfioravanti qwen flash next is really good and almost anyone can ran run it if they have enough ram and a fast nvme ssd!

[A K I O @z3syx]

@CaryPalmerr @Teknium @elonmusk my LinkedIn account got hijacked a few days ago and turned into a random woman's profile till i noticed it and recovered it.

[A K I O @z3syx]

@LLMSherpa @anthonyronning thank you!

[Sherpa @LLMSherpa]

@anthonyronning https://t.co/5kNe4ZHaAU

Pirateface good.
Huggingface bad.