AV

[AV @TheAvtoritet]

Welcome to my library.

Feel free to browse around.

You can find all my threads in this thread, or via this link: https://t.co/B6sC2a34Fw

[AV @TheAvtoritet]

@njmarko @SpaceXAI Interesting, thanks. I'm testing supergrok now to see if it can be a full send

[AV @TheAvtoritet]

@njmarko @SpaceXAI How much usage do you have on it? Do you have any subscriptions running to compare? Currently on claude x20 plan and gpt x5 plan, but usage went down so bad lately I am considering grok now

[AV @TheAvtoritet]

@bob_hw_store Do you think it’s more useful at faster speeds,lower intelligence and context window/kv cashe, or push to 1M context window with YaRN?

For grunt work that is, given that it can rely on frontier for strategy/important decisions

[AV @TheAvtoritet]

@bob_hw_store I have found a plug willing to ship them to me for $1.2k a pop, was wondering at what price point they lose appeal

been sparring with sol a bit on how qwen would fit in my hermes, probably will pair it with frontier models to do fast grunt work and learn from them🤔

[AV @TheAvtoritet]

@ZubyMusic Happy birthday Zuby!

[AV @TheAvtoritet]

@SpaceXAI @grok can grok subscription be used by Hermes agent like openai's sub, or does it need to run on API like claude

[AV @TheAvtoritet]

@Mho_23 EDITOR

[Miko @Mho_23]

i just built a 24/7 AI video editor inside Claude Code

this was supposed to stay internal but f*ck it i'm leaking the whole thing..

you give it a reference video and your product info. it comes back with a finished AI UGC ad ready to post.

here's how it works:

the Higgsfield video analyzer MCP breaks down any winning ad you feed it. it pulls the structure, pacing, style, and timing from the reference so you're not guessing what made it work.

then a custom Claude skill takes that breakdown and generates all the video clips using Gemini Omni. which then it renders each scene, stitches them together, and adds captions all in one go.

you're not jumping between tabs or scrubbing timelines or waiting on editors anymore. the whole thing runs inside Claude and outputs a finished 9:16 ad

this is faster than any editor could ever do it and costs almost nothing to run..

i'm sharing the full video breakdown, which includes:

> the Higgsfield MCP setup for video analysis
> the custom Claude skill that handles generation and stitching
> the exact workflow for going from reference to finished ad
> how to batch this out at scale

RT + reply "EDITOR" and i'll send it over (must follow so i can dm)

[AV @TheAvtoritet]

@Mho_23 UGC

[AV @TheAvtoritet]

@hthieblot Fallout 2, osrs, Fiesta Online

[AV @TheAvtoritet]

@XdelaC Never seen this myself, but does not surprise me at all 🤣

[AV @TheAvtoritet]

@limitlesstack Appreciated 🤝

[AV @TheAvtoritet]

@BigTimeBlazo @thespearing @AlexFinn I follow, sure, but there will be cases that you can’t be trained on and experience/intuition in the digital realm is the difference between failure and safety

[AV @TheAvtoritet]

@thespearing @AlexFinn Getting openclaw in blue collar hands is like giving your toddler access to your firearms room with open ammo boxes. Sooner or later he’s shooting someone or something.

Doing local properly is like getting into an altercation you’ve done 1000 drills on

https://t.co/aw5szlgB4L

[Nathan Spearing @thespearing]

Programers to blue collar business owners running @openclaw after overcharging for apps they were building with AI.

[AV @TheAvtoritet]

@sudoingX @AlexFinn k2.5 was released with native INT4 quantization, it was literally trained to run at INT4, not compressed after. Moonshot's own benchmarks are all run at INT4. There's no 'full precision' version that performs better.

Oh & you need 4 H200 GPUs to run k2.5 which is over 100k

[AV @TheAvtoritet]

@EXM7777 Copy my openclaw trading bots onto 1 & sell the 9 others for extra trading balance

[AV @TheAvtoritet]

@pipelineabuser Lmaooo

[AV @TheAvtoritet]

@hugomercierooo Twin