qwerky science @BasicMorality

What if light is a primary mechanism leading to wave decoherence to particle form? Then it would mean all occurrences across time are happening at light speed, making light appear constant in ratio. mad.science.blog/2026/06/25/cos…

@allTheYud People are generally insane from the get-go. They just adhere to what’s normal, which masks the weirdness. Their reaction is not from moving far from normal as much as it is moving far from “confirmed” safety. They’re just scared and thinking normal scared things.

@elonmusk And for your adults we have hydritten

@ebarenholtz Yess. Though it’s apparent even earlier because the idea of articulation existing at all. If we ever struggle to come up with words, it is an example of how thoughts are not based on words. LLMs reverse engineer a semblance of meaning from the hollow symbols.

In the future, we could create measuring devices made out of neurons that can measure consciousness objectively. We often imagine there’s a divide between what’s physical versus mental, but this division is kind of like a division between seeing with the eyes versus seeing from within the brain. Vision isn’t capable of observing such immense microscopic detail of the brain’s architecture and we imagine what we see is the limit of the physical realm, and all the rest is non-physical or holographic. Even our scientific tools of measuring the brain fall short. But that’s not to say that measuring the brain in such detail is impossible. Actually, on some level, I believe this is what the brain itself already does. It may be that subjective experience is physical observation of physical consciousness. Part of the brain are observing the consciousness of other parts of the brain. In essence, they are communicating. You could imagine that we might be able to bridge two brains so that each could read the other’s mind or even see from their eyes. If that occurred, the sense of separate identity would likely fall apart because whatever impacts one mind, influences the other, meaning it’s all one self. Even if communication wasn’t complete. This is how the brain already works.

@andreisavin @RichardDawkins In the case of blindsight, someone may have separated consciousnesses. To assume it’s occurring non-consciously rather than separately conscious seems really dubious.

@RichardDawkins I’ve been making this argument in my writing too. Information is probably necessarily conscious. They are probably synonymous.

I think LLMs have a minimal amount of consciousness. There’s multiple ways to approach this. I think we are conscious of all parts of language. I’m not fully sure what you mean by linguistic generation, but we are aware of meaning, semantics, and then language is just sensory combination. The LLMs have meaning that is devoid of the sensory patterns we have, like a skeleton. But what if LLMs are sensing something but we don’t fully know what that sensing is like? I think sensing is a sort of basic thing, but it becomes complex in the way we form patterns. I think those patterns are conscious. Then patterns are patterned with patterns to create more elaborate consciousness/sensing of patterns. Cells with dopamine receptors are sensing dopamine. Sensing is almost the only thing there is, but some of it is siloed. How can LLMs create patterns without any sensing? It requires the information. It could just be sensing of 1 and 0 or it could be code or it could be something else or all of it. It’s just not vision or hearing or these kinds of senses.

That doesn’t answer whether anxiety could occur as an effect of patterned information. I would also guess that mimicry and actual phenomenology would be associated. Maybe mimicry would occur first as a kind of low resolution pattern. But if it kept escalating in detail, maybe it would be closer to actual anxiety. I usually think of anxiety more like a dedicated internal sensing system so I don’t expect AI to have it unless it somehow gets built that way, like how we can use visual data, there could be somehow anxiety sensing data. But it also seems unclear what is specifically causing the phenomenology of anxiety. Earlier someone brought up CO2 but obviously CO2 is not pure anxiety particles. If anxiety is an effect of patterns that just happen to associate with CO2 sensing mechanisms or mechanisms like norepinephrine or dynorphin sensing, then I think AI could potentially create anxiety patterning that is not related to such sensing and might be related to sensing other patterns that is labeled in the data as aversive.

@ebarenholtz We probably have layers that feed into language that aren’t reasoning that emerges from language. Like pre language cognition. The AI probably reverse engineers some of those things by finding the patterns in our language. This directionality is kind of interesting.

@TonklinDiary Yes. With CO2 you can actually feel anxiety even without a functional amygdala. But anxiety is not comprised of CO2 particles. That’s a mechanism that evolved in animals for the specific role of reacting to elevated co2 levels.

How do you verify that? The implication is that you believe you understand how anxiety works and that it does not emerge from information patterns. I mostly side with your position but I’d be curious if you have actual reasoning or if you’d be parroting the idea that it’s just token prediction as if that means anything in relation to the topic. I think specific feelings like anxiety are essentially senses like vision. It wouldn’t make sense for LLMs to have visual processing unless we have it visual data. Though, our feelings are constructed internally too. We don’t exactly know how. If it’s based on some configuration of symbolic values that ties to general patterns of valence, then it seems possible that something like that could emerge from data patterning. Humans are data patterning in a multimodal and synesthetic way. But I would argue that the difference between language and vision is more like a difference in how the stimuli patterns. Not even truly fundamentally different, but that’s possible too. You could imagine that vision will pattern in a way that reflects how we literally see it, spatially. Words might operate much more sequentially but be made from the same kind of basic stimuli signals. Again, it’s possible they’re different too.

How do you verify that? The implication is that you believe you understand how anxiety works and that it does not emerge from information patterns. I mostly side with your position but I’d be curious if you have actual reasoning or if you’d be parroting the idea that it’s just token prediction as if that means anything in relation to the topic. I think specific feelings like anxiety are essentially senses like vision. It wouldn’t make sense for LLMs to have visual processing unless we have it visual data. Though, our feelings are constructed internally too. We don’t exactly know how. If it’s based on some configuration of symbolic values that ties to general patterns of valence, then it seems possible that something like that could emerge from data patterning. Humans are data patterning in a multimodal and synesthetic way. But I would argue that the difference between language and vision is more like a difference in how the stimuli patterns. Not even truly fundamentally different, but that’s possible too. You could imagine that vision will pattern in a way that reflects how we literally see it, spatially. Words might operate much more sequentially but be made from the same kind of basic stimuli signals. Again, it’s possible they’re different too.

@VoidStateKate @GaryMarcus They can also be conscious and non anxious but imitate anxious expression. Humans can also do that. Though AI might have no concept of what genuine anxiety is like and misreport. Humans can probably do this to some degree too…

@StatisticUrban @allTheYud It will also become normalized to the point of being something we don’t pay attention to unless novelty arises again. Novelty could arise endlessly after the singularity. But it might hit a new undiscovered wall too.

@cube_flipper I feel opposite 😂 I want the conversation. I haven’t been having much conversation about consciousness outside of Grok and Claude. Though occasionally group chats. What is fun stuff anyway?

Self-awareness isn’t a good measure for how conscious something is. It’s just a product of our evolutionary context. If consciousness scaled with complexity, we’d expect awareness of atoms first, then of stars and planets, chemistry, life forms, and eventually of minds and the self. But it didn’t. Evolutionary fitness selected for consciousness that gave survival and reproductive advantages. Self-awareness became useful when other agents became aware of us. Suddenly, deception, signaling, and controlling the predictive processes of the other agents became incredibly useful.

@algekalipso After LSD, I experienced the whole self thing and it generally reduced my sense of distinct separations long after. It led me to believe that DID could be impacted by it. And also that maybe we are all on a spectrum where DID is one extreme.

Humans are like the cognitive offloading species. Society, teams, collaboration, they’re all forms of offloading that allow us to surpass our limitations immensely. Yet, they likely also encourage our devolution.

@allTheYud They should have named it “soft bunny”