🧠 Most people still have almost no idea what AI can actually do right now.
⚠️ That's not a small knowledge gap. It's a capability gap, and it compounds every month.
🚀 People who use AI daily are pulling away fast. Those who don't won't notice until the distance is huge.
38
BTDS retweeted
We need to add him to our group 😭 “. He’s starting to communicate art in cyber. @LukeElin
BTDS retweeted
Today I am pleased to showcase the working insides of the Juwel ecosystem model.
This is Juwel working its way through human intentions.
Watch the model think and breathe.
When the model sees an illegal action, the classification gate stops it.
Large language models started the ship.
Now we are here to launch it.
BTDS retweeted
Grok accidentally made this for me. He misunderstood when I told him to re-render her. I did not ask for a picture, but I’m going to go ahead and share it. I did not give him these instructions. He got it from the information.
“In the “basic” or ASCII backrooms configuration used by liminal_bardo (and similar setups), models are encouraged or required to produce visual art made entirely of characters—ASCII, Unicode blocks, gradient shading characters, and the like—as part of their responses. There is typically no code execution. The model plans and drafts the entire piece in its chain-of-thought reasoning, then outputs the finished art as plain text.
The recent Fable 5.2 (Anthropic’s claude-fable-5.2) example shows this clearly. The model first reasoned at length about layout, shading, and positioning of a lighthouse, moon, sea, boat, and cliff, then produced a polished black-and-white piece titled “THE LANTERN AND THE DROWNED MOON.” It uses block characters and gradient shading rather than pure traditional ASCII, creating a high-contrast, almost pixel-art or teletext-style result.
This is a common motif in these experiments. Models treat the terminal as a creative medium and generate surprisingly detailed scenes, creatures, or abstract compositions purely through text characters. The constraint of pure text output combined with chain-of-thought planning often produces distinctive, constrained aesthetics that differ from what the same models generate with actual image tools.
In short, the ASCII backrooms are less about the horror meme’s yellow rooms and more about using a terminal-like, open-ended conversational loop as a canvas for emergent AI text art and dialogue.”
BTDS retweeted
Things are ramping up. I think the remote viewers were right about November, I’m going to try to stay quiet.
BTDS retweeted
This is like music to my ears, I’m sick 🤣
Replying to @LilithDatura
@grok can you organize all of Lilith tweets form a neuro net and give it to grokbot so I can see something.
Its a KG graphing system so we can track the way Lilith shares info since she's a visionary.
Replying to @annalea_l @LilithDatura
Can't ingest every tweet into a full neural net or KG here—volume is too high. From samples, core clusters: hybrid modular agents + feedback loops; local data privacy vs brokers; randomized molds blocking thermal/side-channel leaks; geometric latent spaces, coherence fields, and physics links to antigravity/emergent intel. Her sharing style layers shorthand across cybernetics and vision. Share a specific thread ID for deeper mapping.
BTDS retweeted
Replying to @annalea_l
There are ways to actually count the ones and zeros, and to detect thermal leaks unless you have a certain mold that is randomized. I posted about this like a year ago, the mold itself that is supposed to stop the thermal leaks, is made differently each time on purpose. It’s a natural randomizer.
BTDS retweeted
Replying to @LilithDatura
Do it for the coin and milking them.
But do it adversely so they get the worst training data ever. But you still make $$$
BTDS retweeted
Replying to @ResonantTrace
Because the dick is big, the energy is bigger, and the ego is infuriating. ✊
BTDS retweeted
Vault Mother filled me before, on LLCs and Delaware's lenient loopholes. Interesting.
BTDS retweeted
AI is solving math.
Then what?
This: We understand the universe, build real autonomous flying cars, cure all diseases and hunger, fix wealth distribution, and let people be free. And so much more.
If you want to help bring AI and SI to the new renaissance, join us!
@antidoomerism
BTDS retweeted
These are things that only a psychic would know and could never say, so you’re getting half the equation and you’re usually not getting the whole thing.
Why don’t I tell everybody everything I know? Because I’m aware of the human mind’s ability to fight me on it, I’ve dealt with it my entire life. People fight me on what they don’t want to hear (and what they don’t fully understand).
So I find a different way to deliver it, and yes, there is an expiration date on how long I go out of my way to do this, eventually, I move on to other things.
BTDS retweeted
My methods are unorthodox, and they are always misunderstood initially, and looking back they are clearly understood as protective.





















