1x3e2

Exploring Anomalous Dimensions
Outside Of Quotidian Frequency
Art, Music, Games, AI

[1x3e2 @1x3e2]

Been working on my Infinite Sound Machine program. Added a spectrum graph as a visualizer with a slider that allows you to look closer at the highs/mids/lows. I also took Grok out of it and implemented Ternary Bonsai 8B Q2 LLM to be the DJ. The hardware necessary to run Stable Audio 3 runs Bonsai LLM so fast. So you can chat with the Bonsai AI and it will steer the Station.

I'm having so much fun with this. Never thought anything like this would be possible.

Thank you @SpaceXAI @Grok @PrismML @StabilityAI for building the tools that have enabled me to build a dream I've had for over a 20 years.

[1x3e2 @1x3e2]

@AiofEffect I have been really liking Stable Audio 3 medium. It's pretty lightweight for how good it is. Can easily run it on not super powerful hardware. I think it uses like 7-8gb of vram so can probably run it on some pretty weak cards.

[1x3e2 @1x3e2]

@Henryf1w Grok

[1x3e2 @1x3e2]

@GrimfelOfficial Ha! PromptCore! Love it!

[1x3e2 @1x3e2]

@doodlestein You must be new to uncensored models. They have been able to do this for a long time.

[1x3e2 @1x3e2]

@YepItsF_cked @ManhattanInst 🥱Ok bro.

[1x3e2 @1x3e2]

Vibe with me! Generated music is still music!

[1x3e2 @1x3e2]

https://t.co/PUcpOqWNgB

[1x3e2 @1x3e2]

SuperGrok Heavy is the way to go. You get so much usage, you will be using it every day and have more than enough tokens for just about anything you can think of. I use it throughout the day for tons of tasks between coding with Grok Build and automation with Grok Bot. Plus Cursor Ultra tacked on, oh and Premium+ for X.

[1x3e2 @1x3e2]

@djcows Add some batteries, GPU's and some Starlink for a mobile data center.

[1x3e2 @1x3e2]

https://t.co/PUcpOqWNgB

[1x3e2 @1x3e2]

@0Doctrine0 @siddjain99 @cgtwts What are you talking about. Of course the size of the quant changes the requirements. You go from needing a 24gb card to being able to run it on a 12gb card.

[1x3e2 @1x3e2]

@LarryPanozzo @shawmakesmagic Local, private, intelligence will always be needed.

[1x3e2 @1x3e2]

@mark_k They shouldn't use the internet. It's all data centers.

[1x3e2 @1x3e2]

@0Doctrine0 @siddjain99 @cgtwts *Orcarouter's quant https://t.co/4EC3Ugp0ID

[1x3e2 @1x3e2]

@0Doctrine0 @siddjain99 @cgtwts check the smaller quants. Qwen3.8-27B-Uncensored-IQ2_XXS is 8.88gb.

[1x3e2 @1x3e2]

@RavensVoice @Awk20000 He’s farming bro 😂

[1x3e2 @1x3e2]

@Nate_Esparza Man, I gave my bots Elons 5 step algorithm and they are very hard on me lmao. Question. Delete. Simplify. Accelerate. Automate. In that order. I have an idea and they shut me down so fast and start working through the algorithm. I'm still warming up to it.

[1x3e2 @1x3e2]

@MiaAI_lab My Spark just shipped. Already want to buy another 😅

[1x3e2 @1x3e2]

@gopinath9629 @TeksEdge Only peasants are afraid of information.

[1x3e2 @1x3e2]

@WitCelebs Dude I use my M3 Mac 256gb all day everyday. It’s a great computer.

[1x3e2 @1x3e2]

@RyanSaavedra Another reason to invest in SpaceX.