Recursive_roots

1/ So, bad news. It's 2026 and you personally need a defensible answer to “is that data center a person.” Not a vibe, not a guess. An actual answer. Here’s mine.

Stuff accounting for 85% of the mass of the universe is around you right now, where you sit. Invisible, unmeasurable in the EM spectrum. We call it dark matter. It's definitely there, we can see it bending space. But it doesn't seem to collide with us.

@Xallt_ @francoisfleuret While you were composing this tweet in your head, who was listening?

@francoisfleuret I'm so tired of people saying that LLMs are just a Chinese Room We all are

Listen. There are two sorts of bits: The one coming from understanding, like if *I* write "2+2=4", this string is encoded in u-bits, understood-bits, good stuff. Whereas ChatGPT writing "2+2=4" produces a string of fs-bits, aka fancy-statistics-bits. Which are crap.

@ZeframM @peterrhague Put differently: Is there a part of spacetime that is invisible to electromagnetism (but not gravity) in most circumstances, that we don't know about that because all our detection tools are also made out of electromagnetism?

@recursive_roots @peterrhague What do you mean?

Yeah it’s called “the entire electromagnetic spectrum outside visible light”

@JFPuget @HealthRanger Structured partiality through a qualifying physical aperture. Data centers at least for now cannot count as that.

@HealthRanger Care to give one definition and how to test for AI systems having it?

TL;DR We need to define consciousness before we can check if AI models can be conscious. Is there a need for a paper to state this? Seriously? What am I missing here?

@UnknownSeraph3 @miniapeur zenodo.org/records/187356… suggested protocol in the paper there

@recursive_roots @miniapeur Holy. How would you test this experimentally?

I want to solve real science problems. Hit me up guys

''You can't even understand that code the AI wrote for you!!'' ''Grok, print out an MD of the code explaining line by line at a layman level what the code is doing, so that I can understand it''

@davidasinclair @Microsoft Get in here dude. Please get rid of this.

Anyone else tired of deleting this?

@Nate_Esparza That's a strong claim. Unfalsifiable. There is a weaker, falsifiable claim that you might like better. That the universe *acts* like a computer.

I’m convinced we are in a simulation

In the age of LLM psychosis if you don’t have at least one Zenodo paper on a theory of everything are you actually even trying?

@Andercot You rang? zenodo.org/records/187050…

@kimmonismus It's simple. Just demand that humans shall remain necessary. One very simple rule to survive these tools. ''Humans shall remain necessary.''

One of China’s massive training labs. And people still think their jobs are safe. The next revolution is already being trained.

13/ Framework in linked paper. It makes novel, testable predictions. Distinct neural signatures for volitional vs. traumatic dissociation that no other theory generates. And it gives an answer: current and forseeable AI is not conscious. Wrong substrate to form an aperture.

1/ So, bad news. It's 2026 and you personally need a defensible answer to “is that data center a person.” Not a vibe, not a guess. An actual answer. Here’s mine.