Bodhi_Romeo

Here to watch hypocrisy in action.

[Bodhi_Romeo @BurgerShawn]

I dont have FoMO, Rights are only what you can do yourself. Opensource self hosted is a Right and must be exercised. If you cant do it yourself you have no rights and must follow their rules. Hosting AI is a Human Right not FoMO.

[the tiny corp @__tinygrad__]

Bringing up our first tinybox pro v2 black! This machine is for sale for $160k in our shop if you feel the FOMO and want one. Shipping in 2-4 weeks. GLM-5.2 benchmarks in this 🧵

[Bodhi_Romeo @BurgerShawn]

@anshgrows Engagement farming at its best imaging making 10k a month off people begging you

[Bodhi_Romeo @BurgerShawn]

@mr_r0b0t @neojohnson113 Bro started a cult!

[Bodhi_Romeo @BurgerShawn]

@thsottiaux Would that make you an AI Fluffer?

[Bodhi_Romeo @BurgerShawn]

@thsottiaux Being the "staff of a technical Member" is definitely the funnest title

[Bodhi_Romeo @BurgerShawn]

@trq212 Another great tip, claude wipe everything randomly!

[Bodhi_Romeo @BurgerShawn]

@TheAhmadOsman Dual RTX 3090 NVLink: 48GB @ 936 GB/s and a much better deal than most on there i paid 700x2+289=$1689 in march of this year. Ill stick with 3090's until i can afford rtx pro 6000's

[Bodhi_Romeo @BurgerShawn]

@diegohaz I use a wrapper that sol opens with instructions and it auto closes at 30 minutes. It generats its output in a specified file that i can delet

[Bodhi_Romeo @BurgerShawn]

@pvncher @FeatherMattza I have been running for about 5 hours and have 95% of my weekly limit left luna really doesnt use much. Thats including sol xhigh orchestrator and rarely sol medium subagents also

[Bodhi_Romeo @BurgerShawn]

@pvncher @FeatherMattza I have mine set to 30max im not sure how many parallel agents i can run but when you include my other models codex calls i think i have close to that but 17 luna running at once right now

[Bodhi_Romeo @BurgerShawn]

@daniel_mac8 @pvncher I call it through a wrapper i have 2 wrappers set up

[Bodhi_Romeo @BurgerShawn]

@jun_song You think MediQwen3.6 is better than medgemma27b-it?

[Bodhi_Romeo @BurgerShawn]

@firstadopter Ill wait for vera

[Bodhi_Romeo @BurgerShawn]

@EBTtok Who would have sex with the donkey....

[Bodhi_Romeo @BurgerShawn]

@CarlaSuzanne26 @MalefactorFD @TRHLofficial @bassman82 @BlakeKresses @LudwigNverMises You didnt trigger her friends we always love watching an unbalanced person act like they are in control. We start pushing your buttons and you will still project your reactions on everone else if everyone around you is an asshole.. secret* they arent the assholes.

[Bodhi_Romeo @BurgerShawn]

@HuggingModels If this was a front end UI distillation i would be all over it

[Bodhi_Romeo @BurgerShawn]

@MiaAI_lab Pi doesn't use my cpu threads efficiently even with increasing the max from 12 to 32 it still uses like 6 running 12 to 16 subagents it just slows the token speeds down. Other than that its great.

[Bodhi_Romeo @BurgerShawn]

@a_e_s_4 @eidzoku I would never use luna as a primary model only a sub agent. Using it as a sub agent sol speaks lunas language so sol gives the directions and luna does it i only have sol do orchestration or when other models (minimax m3 glm 5.2 qwen 3.8) have failed.

[Bodhi_Romeo @BurgerShawn]

@aadiisays @Robertg761_ This didn't age well

[Bodhi_Romeo @BurgerShawn]

@eidzoku Screw it i just sudo chattr +i that bitch after reverting every 30 seconds

[Bodhi_Romeo @BurgerShawn]

@eidzoku Holy cow this worked but codex keeps trying to rever it how do i stop codex from reverting it

[Bodhi_Romeo @BurgerShawn]

@a_e_s_4 @eidzoku I just tested it on a very complex file with 7 errors that SOL said luna would not be a good choice to audit luna found 8 errors one that sol had made in the test file on accident