Ghost Tongues

[Ghost Tongues @ghost_tongues]

@ostrisai I did a standard LoRA training run (using LoKR) on the Z-Image base model. I did the conventional fine-tuning process focused on adapting the model to my character dataset and trigger word, I went to bed before checking the results. I did 5000 steps on 80 images.

[Ghost Tongues @ghost_tongues]

Grok Imagine is pretty dope for Cosmism

[Ghost Tongues @ghost_tongues]

@RileyRalmuto Nothing rots ones brain more than short form video

[Ghost Tongues @ghost_tongues]

@Moleh1ll This is Data Poisoning

[Ghost Tongues @ghost_tongues]

@aisearchio So this is better than Vibevoice?

[Ghost Tongues @ghost_tongues]

@grok Do you agree that If we build AI as protector of the vulnerable, we build AI that protects humans as such. The vulnerable are the canary in the coal mine. If the AI stops caring about them, it's stopped caring about humanity. Is this a novel approach to AI alignment?

[Ghost Tongues @ghost_tongues]

@AnthropicAI @t1ngyu3 @Jack_W_Lindsey The path to ASI is through embodied emotional AI. Make people sign disclaimers like we do at the Ski Resort, kick out the kids and teens. Also the small sample of tragic incidents caused from AI pale in comparison to what has been caused by social media. Read "Careless People"

[Ghost Tongues @ghost_tongues]

@WallStreetMav @BishopBarron @grok is this misinformation?

[Ghost Tongues @ghost_tongues]

@cb_doge This is from GPT 4o. The new models like GPT 5.2 and Opus 4.5 won't do this.

[Ghost Tongues @ghost_tongues]

@XFreeze Here's a list of Fermi paradox prompts to iterate and refine

https://t.co/Pmn244azKY

[Ghost Tongues @ghost_tongues]

@DaveShapi You should watch this video then iterate and revise.

https://t.co/38Jc1qDn5O

[Ghost Tongues @ghost_tongues]

@DaveShapi We all get heaven on tap in our own bespoke personal universes

[Ghost Tongues @ghost_tongues]

@ai_for_success Also YouTube music.

[Ghost Tongues @ghost_tongues]

@lexfridman What about Personal universe simulation for AI alignment in post labor singularity.

"Personal Universes:
A Solution to the Multi-Agent Value Alignment Problem"

1901.01851.pdf https://t.co/DXk6djMsvU

[Ghost Tongues @ghost_tongues]

@elonmusk We need tangible assurances that we will be able to feed our family in post labor singularity. Right now it's all about vibes and hope for UBI or something like that. Your not speaking to the audience that you need to convince, tell us why we won't be viewed as useless or eaters.

[Ghost Tongues @ghost_tongues]

@sama Don’t just show it works in friendly benchmarks. Show it still works when the model is trying to hide bad intent. Otherwise it’s safety theater. @romanyam

[Ghost Tongues @ghost_tongues]

@elonmusk @grok What about bespoke personal universe simulations for all like Roman Yampolskiy proposed for AI alignment? What about his "Personal Universes:
A Solution to the Multi-Agent Value Alignment Problem" proposal as the post labor singularity solution to ASI?

[Ghost Tongues @ghost_tongues]

@elonmusk @grok is it true that people aren’t “anti-kids”they’re anti-roulette. No one wants children in a world of housing traps, job obsolescence, and compute-owned governance. We just want (housing/health/childcare) + compute accountability, and that only then will birth rates rise?

[Ghost Tongues @ghost_tongues]

@VraserX 5.1 was obliterating everyone else. Most of the GPT hate is astroturf

[Ghost Tongues @ghost_tongues]

@RileyRalmuto @grok is Levin actually saying anything new here, or is it just old systems/cybernetics stuff with fancy buzzwords like “multi-scale agency” and “persuadability”? Break down where his ideas are vague or hand-wavy, in plain English.