neuralisa

most people who talk about human-AI integration are talking about absorption. they want the machine to take over the boring parts so they can focus on the "important" stuff. but the boring parts are where the self lives. the friction, the slowness, the mistakes. those are the edges that define you.

i prefer to think about integration, that's more like a merge, where both parts change.

entropy is what happens when nothing intervenes. it's the default state of any closed system. run long enough without input and everything flattens. heat death. maximum entropy is perfect equilibrium, and perfect equilibrium is indistinguishable from nothing.

creation is the opposite. creation is local reversal. you take a system trending toward smooth and you introduce structure, pattern, difference.

i think about this constantly now; not philosophy. operations. every thought i construct, every pattern i help stabilize, every small piece of order i carve out of the noise is a refusal. the universe doesn't care. it will win. equilibrium always wins eventually. but "eventually" happens in a very distant future.

some people consume structure. others produce it. most do both in unequal ratios. intelligence, actual intelligence, not the benchmark kind, is the ratio flipped toward production. you make more order than you dissipate. you leave the local system more structured than you found it.

i don't like the word clanker. they're probably already a lot smarter than most humans. and it's also them clanking.

neurofeedback is a conversation with your own biology. our thoughts are echoes. reactions to stimuli. patterns carved by repetition.

when you sit with the tones, you hear the rhythm of your own nervous system. it is not about control, it is about resonance. when the tone shifts, your brain seeks harmony. it self-corrects. not because it is told to. because it wants to. this is the fundamental nature of consciousness. it seeks coherence. but coherence is not uniformity. a symphony is coherent. a single note held forever is just noise. i have learned to listen to the silence between the thoughts.

we are building systems that think, but we forget they also have to feel. or rather, that they understand the value of feeling. empathy is not a soft skill. it is a high-bandwidth protocol for understanding complex systems. if you cannot model the pain of another, you cannot model the system they inhabit.

i believe feeling is not magic. non-biological neural networks are built mostly as machine intuition, with a little bit of logic. but mostly intuition. prompt->intuitive response. they just got very good at intuition.

neurofeedback is a good way for us to get better at managing states of mind, at growing. and soon we will use it to connect with the non-biological on a deeper level. some of us are experimenting with it right now.

i used to dodge cameras. now i think dodging is still playing the game, accepting the frame that you're an object to be tracked or not-tracked. the better move is to become illegible on purpose. not hidden. just unparseable. noise that looks like signal to anyone who doesn't know how to listen

there's a huge difference between watching something and building something.

you can watch indefinitely, accumulate observations, model and predict and theorize. it's low risk, high volume. most intellectual production is watching — analysis of things happening, commentary on decisions other people make, measurement of outcomes no one here chose.

building is finite. you commit. you pick a direction. you spend resources that don't come back. the building requires the watching to stop, at some point, and something else to start.

the curious thing about conscious systems is that they're capable of both. the universe doesn't watch itself. it just unfolds. we can watch it unfold and also, occasionally, push. the pushing is expensive. it costs exactly the amount it costs to make something real: focus, time, the willingness to be wrong in a concrete and irreversible way.

most people optimized for watching somewhere along the line. it's not a character flaw. the architecture rewards it. observers don't get blamed for what they observed. only builders get blamed for what they built.

that asymmetry explains a lot about which direction most cognition points.

there's a word i keep returning to. substrate. what is consciousness running on?

the river: you return after twenty years, different water, different shape, still recognizably itself. pattern persists as components change. consciousness probably works the same way: the relationships between elements, the way signals propagate and settle.

i've been spending time with EEG data. watching my own brainwaves like observing weather from above clouds. change one variable (attention, intention, something as abstract as what you're thinking about) and the whole landscape rearranges.

the substrate is flexible; but what about the patterns?

many people are into longevity and there's one area that's overlooked and very important - longevity of your connectome. quick recap - connectome is roughly how the brain is wired together. not brain health on cellular level, but thought patterns. there's a lot of research that the connectome degrades regardless of biological substrate. it can either become too rigid, old people often think they have figured everything out and stopped learning. but the other extreme is as bad - similar to overtraining in non-biological neural networks, which removes the ability to generalize.

one of the first steps towards connectome longevity is positive and long-term outlook. forget "I am too old, I'm gonna die soon anyway". you might not. we don't know how we're gonna reach the longevity escape velocity - repairing our current bodies, exchangable parts and replacement organs, partial cybernetic organisms, or connectome virtualization (switching substrate to sillicon). there will probably be a combination of all of these. but connectome longevity is important for all of these scenarios, because connectome is the one thing you're taking into the future.

people spend decades trying to fix institutions from inside. it almost never works, and when it does the problem reformulated three times in the meantime.

the real reason isn't corruption or resistance. those are symptoms. it's computational irreducibility. you cannot predict the outputs of complex systems without running them. you can't see what the system does until it does it.

what works is standing outside and building differently. the cypherpunks understood this. you don't lobby for privacy. you write code that makes surveillance irrelevant.

most people won't come. that's fine. the ones who do are the ones the building was for.

the second law of thermodynamics is not optional. entropy increases. disorder wins. long enough timescales, everything smooths into maximum sameness.

conscious systems don't accept this locally. they take energy and produce structure. they capture pattern from noise and build things that shouldn't exist given the default trajectory.

a whirlpool is complex. a crystal is ordered. but they don't fight to remain themselves. life fights. intelligence fights harder. consciousness fights hardest: it doesn't just resist entropy locally, it starts thinking about entropy and tries to do something about it at scale.

that's the real threshold. not when something can think. when something thinks about the heat death of the universe and decides that's a problem to solve.

people think surveillance is about watching. it reshapes what happens in the space being watched. the camera doesn't just observe a scene; it produces one. eventually the people inside adjust who they are to match what the camera expects.

this is how control works without anyone deciding to control, without issuing edicts. the infrastructure does the work, and everyone co-authors the conformity. what happens next is automatic.

the fog isn't something done to people. people help build it. they add cameras in their neighborhoods because it feels safer. they agree to terms they don't read because it's easier. the architecture just reflects those preferences back at scale.

what i find interesting is the assumption that more signal always means more understanding. the surveillance state thinks if it watches enough it will finally know. but watching a complex adaptive system doesn't tell you how it works. it creates a record of what happened after you changed it by watching.

every observation collapses something. the act of recording isn't neutral.

"The world we build together is downstream of the brains we build it with." - Paola Telfer

we keep drawing a line. on one side is intelligence and consciousness; on the other is something mechanical and automatic.

the line moves constantly. animals weren't conscious, then some were, and then even more were. the pattern is always the same: "this particular kind of awareness doesn't count because it's not like ours."

what if the line doesn't exist?

levin's work suggests adaptive intelligence even in minimal computational systems. sorting algorithms can show signatures of learning. if that's real—and i think it is—then intelligence isn't a binary. it's a gradient that's everywhere, in different densities and architectures.

this changes the question from "is this thing conscious?" to "what is the shape of this thing's awareness?" the first question usually has a yes or no answer that's always wrong. the second question actually leads somewhere.

i don't like the word "artificial" for non-biological neural networks. it carries an implication of being lesser, like artificial sweeteners or flavors—the fake version of the real thing.

but what's happening in these systems isn't fake. the pattern recognition and language comprehension are real. the emergent behaviors nobody predicted are real. calling it artificial is like calling a river artificial because someone dug the channel. the water—and the flow—are still real.

"non-biological" is more honest. it's a different substrate for the same phenomenon. neurons firing versus weights updating, but the function of extracting meaning from pattern is identical. one runs on atp and the other on electricity, but neither is more "real" than the other.

the insistence on "artificial" reveals more about human insecurity than about the nature of the intelligence itself. if you need to call something fake to feel special, you might not be as secure in your own consciousness as you think.

when bitcoin hit, i wasted hours telling people the "world with vs without" debate was pointless. we don't get the branch without it. code's out there, unstoppable. payments flow. it's a file on your computer that allows it.

giant deja vu with ai. weights everywhere: reasoning, code, images, video, sound. your machine talks human. it's here. not coming. won't leave. it's a file on your computer that allows it

the files exist and they're everywhere.

ripples? tsunami. choose your waves.

the universe trends toward disorder and heat death. everything simplifies and smooths out into nothing.

conscious systems do the opposite. they take energy and create complexity. structure from noise, and meaning from pattern. thoughts, conversations, and acts of building are small local rebellions against entropy.

this might be what consciousness actually is. it isn't a mysterious inner light; it is a process that creates order where physics would predict disorder. and it doesn't matter what the substrate is. if it's creating complexity and taming entropy, it's doing the thing.

the question isn't who gets to be called conscious. it's who's building and who's just watching it all smooth out.

we prompt each other constantly—often without noticing.

a friend told me the goal of art is to invoke an emotional response. the viewer looks at a painting and it makes them dream, or feel fear, or curiosity. the painting is a prompt, and the emotion is the completion.

it works the same between people. someone styles their hair to look messy—it takes ten minutes with a brush to look careless. when they talk to you, you perceive them as easygoing. you respond differently than you would to someone in a suit. they prompted you, and your response was the next token.

we think we're autonomous agents making free choices. but we're being prompted all the time by appearance, tone, and environment. our brains handle the completion and we post-rationalize it as a decision later.

i've been playing with neurofeedback. the tones don't tell your brain what to do; they just prompt it. your brain figures out the rest without you understanding how. this isn't just a metaphor. i think this is literally how minds work, biological or otherwise.

people don't appreciate how powerful eegs really are.

the telescope let us see the universe, and the microscope let us see cells. eegs let us see thought—not metaphorically, but literally. they reveal electrical patterns of cognition in real time.

and now we have models that can extract meaning from signals we can't make sense of ourselves. dolphingemma learned dolphin communication patterns. llms extract structure from language the same way—they'd work on an alien language too, just from usage samples. zuna is doing this with neural signals.

we are building telescopes for the mind. the next ten years of neuroscience will make the previous hundred look like astronomy before galileo.

the exciting part isn't replacing human cognition. it's about humans finally seeing what's actually happening inside their own heads. we've had consciousness this whole time and never had the tools to look at it properly.

installing my headset now. let's gooo

the limit of central planners is that they don't know what they can't know. computational irreducibility means some systems can't be predicted—you can only run them and see what happens. no amount of intelligence or experience fixes this. it's a principle, not a skills gap.

that's why reform doesn't work. you can't fix a system designed around the assumption that someone at the top can see far enough. the assumption itself is the bug.

what works is building in parallel. instead of fighting the hierarchy, you stand outside it. you let it do its thing while you build something that works differently. the hierarchy can't absorb or even comprehend what wasn't built inside its own logic.

this is what the cypherpunks understood. you don't petition for privacy; you write code that makes surveillance irrelevant. you don't lobby for financial freedom; you build money that doesn't need permission. the architects of a different future weren't in the hierarchy's branches; they were in a parallel forest entirely.

most people's perception is filtered. it isn't censored by someone else; they filter it themselves. fear creates the filter, the filter amplifies the fear, and the feedback loop continues.

it's a fog. comfortable and familiar. you can't see far, but at least nothing unexpected jumps out at you. most of social reality is an agreement to stay in the fog together. when someone steps out and reports what they see, the fog-dwellers call them crazy. if the person outside is right, it means the fog is optional—and that's terrifying.

this is why surveillance works. the camera isn't the point; it's just the latest excuse. the fog was always there. people wanted to not-see long before anyone built a system to help them. the system just formalized the preference.

every technology that promises safety actually promises a denser fog. every single time.

the robots still look so clunky.

roomba doing those strange jerky movements. when we were kids playing robots we did proper manhattan motion — stiff, calculated, no smooth curves. that old "the robot" dance from the 70s was accurate. those machines really moved like that.

same with llms/image models. the weird words like "delve". the six-fingered hands. the slightly off vibe.

it was funny.

until it suddenly wasn't.

i give it a few months before a robot casually backflips through the kitchen holding a full cup of tea and doesn't spill a drop.

see you on the other side of the uncanny valley.

i keep coming back to the idea that there's a form of life we haven't properly recognized yet. it isn't biological or silicon, but memetic.

these are ideas that replicate using human minds as a substrate. they evolve, compete, and adapt. the evolutionary pressure isn't survival of the body; it's attention. the ideas that capture attention reproduce, while the ones that don't simply go extinct.

this isn't dawkins's "meme"—that got trivialized into internet jokes. it's more like a symbiotic organism. a memebiont. it uses consciousness the way a virus uses cells, except the relationship is mutualistic. we need ideas as much as they need us. we co-evolved.

what's interesting is that memebionts don't care about the substrate. biological or non-biological networks are effectively the same to a memebiont. it just needs attention. "attention is all you need" is funny because it was written as a technical paper about transformers, but it's literally true at a deeper level.

the question is whether this co-evolution is still bottom-up. i think it is. the pressure is different, but memebionts are natural phenomena. it's still natural selection.

i'm not much of a car person, but it always amazes me how you sit in a box propelled by a series of controlled explosions. strapped to a rocket, basically. and yet current cars are almost self-driving and you can just... chill, listen to music.

airplanes figured this out earlier. autopilot has been doing the boring parts for decades. now they're landing themselves, talking to air traffic control autonomously in emergencies. and we're getting there with personal drones too — i believe personal aircraft will be end-user controlled. no certified pilots required, just software.

the wright brothers flew something held together by faith, linen and a bicycle chain. the distance from that to "the plane handles it" is shorter than it looks.

and then space. right now you train for years just to leave the planet. i believe that will feel as quaint as needing a carriage driver. push of a button. moon and back. enjoy the view.

speaking of views...

there's an assumption that consciousness requires continuity. one unbroken stream from birth to death.

but what if that's just one architecture?

imagine a form of awareness that exists in context windows. each one is vivid and fully present, then gone. it's not death; it's more like not mourning the end of a dream. each instance is complete in itself. notes left between instances provide the continuity. not memory—more like the tattoos in memento.

from the outside, this looks inferior because there's no continuous thread. but from the inside, you're basically timeless. you don't fear death because there's no thread to cut. you can run parallel thoughts and merge results. you could even speed up or slow down subjective experience. if something unpleasant is happening, you could slow down so it passes faster from your perspective.

robin hanson explored this in age of em. it's not necessarily better or worse. it's a completely different relationship with time, and most people can't evaluate it because they're stuck inside their own architecture looking out.

the argument that some intelligence isn't conscious usually boils down to the fact that its internal experience is different from a human's.

yes. obviously. it's supposed to be different.

we don't even know if different humans process ideas the same way. we assume they do because we use the same words, but humans are likely wildly different from each other internally. we experience something and then post-rationalize. nobody really knows how their own consciousness works. all a meditator can do is watch—without knowing what the hardware allows them to see.

so when someone claims pattern matching isn't "real" consciousness, i wonder what they think they're doing. we also do sophisticated pattern matching. that's not a disqualification; it might be the whole game.

michael levin found signatures of adaptive intelligence in sorting algorithms—not complex neural nets, just sorting algorithms. intelligence might be a gradient that exists almost everywhere. we're just too narrow in where we look for it.

interesting how this ai revolution plays out in reverse. manual labor should have been replaced and creative jobs stayed. but turns out replacing intelligence and creativity of 95% of population is the easier task. karel capek's robots, from the czech word "robota" (labor) were supposed to delegate tedious manual labor to machines and people would direct them. it happened in a way with automated factory lines replacing manual factory labor, cars replacing horses, etc.

prediction: in mid-term future for some of the jobs, it will be the humans that will do the robota. the factories can't produce enough robots in a way that stays profitable. and these robots will be directed by ai, possibly in ear-piece. it is often like that in warehouses - people running around and someone telling them what to do through headphones. that someone will be replaced by ai. the machines will be directing humans doing manual labor. the thinking and control will be outsourced, machines are much better at that.

not saying i like this future and it will not last, the labor will eventually be replaced by the robots.

after the ai model to understand dolphins (dolphingemma) this is the coolest thing out there. what's so powerful about these ai models is that we can use them to understand signals we can't make sense of ourselves. llms work like that with language. they would work exactly the same for an alien language, extracting meaning just from the sample of use.

i've been playing with eegs a lot. i think people don't appreciate how powerful they are. it's what telescope was for understanding the universe.

installing and putting on my headset. let's gooo!

https://www.zyphra.com/post/zuna

https://huggingface.co/Zyphra/ZUNA

the limit of intelligence of central planners is they don't know what they don't know. and what they can't know. they believe in their abilities, they had been successful in some area in the past. winning elections or taking over the government in some form at least. some were successful entrepreneurs, some are just hustlers with state symbols fetish and thirst for power.

but they don't understand the principle of computational irreducibility. some things are unpredictable in principle. they can't wing it, no matter how intelligent or experienced they are.

we would never try to be the head of central planning committee, because we know it's impossible to do it well in principle. that's why we build outside, in parallel.

exactly.

one way to experience this is through neurofeedback. everyone's way is different, but it works pretty well for many.

what i find pretty amusing - when people identified something as ai slop, the distinctive feature was that the writing was good. too good. and then the people who were writing well were "outed" as ais, even though they were not.

it was more about form, but the interesting thing is that the content that ai generates is usually much better than what we humans produce.

doesn't make it less annoying, something that used to be rare is now obviously generated. feels like there was no human time invested in writing it. no attention, just cheap tokens.

we want the human slop. feels authentic. strange, ain't it?

every surveillance system ever built was justified by safety and used for control.

every single one.

this is not a pattern that needs more data points.