Neurotech as philosophy about what we are: five models

Transcript

00:00:00My name is Michael Johnson and I'm here not necessarily to talk about the stuff that I'm researching right now, but we can also discuss that later.

00:00:11But I want to talk about... and how's the sound by the way? Is this... okay, great, great.

00:00:18I want to talk about a little philosophy today and I want to make the big claim, we can say, that when you're making neurotech, you're actually doing philosophy about what kind of thing we are.

00:00:39And likewise, when you're doing philosophy about... or neuroscience about what kind of thing we are, what kind of thing a nervous system is.

00:00:49You're also doing... yeah, the other. So, anyway.

00:00:57A little bit of intro. I wrote a book on technical consciousness research, Principia Qualia.

00:01:05And I co-founded a research institute and left two years ago. I'm in the process of founding another.

00:01:13And I'm kind of known... who here has read...

00:01:18Read some of my stuff? Okay, great. Nice.

00:01:26Cool. So, let's talk about the details.

00:01:35I think one thing that really gets easily lost in the noise of sort of talking about neuroscience, talking about neurotechnology,

00:01:45is these fields are tricky.

00:01:48And I think that's because we're truly pre-paradigmatic.

00:01:51It's, you know, we can call it neuroscience, we can call it neuroengineering, etc.

00:01:57But that doesn't make it sort of properly paradigmatic.

00:02:03And so, you know, it's not that we don't have stories about how the brain works.

00:02:09It's that we have so many stories.

00:02:14And none are as predictive as we would like.

00:02:20So, that's not to say that there's not a lot of high-quality work going on.

00:02:26It's just that there's also, frankly, a lot of bullshit.

00:02:30And the only antidote to bullshit is good taste.

00:02:39So, my second big claim here is that neuroscience needs treasure maps.

00:02:49think that's the

00:02:50That we're certainly building technological capacity much faster than we're figuring out how to improve people's lives with it.

00:03:00You know, we can have very, very advanced, you know, we can put wires in the brain.

00:03:07And then what do we do with those wires?

00:03:09And the techniques that we found are pretty simple.

00:03:13Let's just pass electricity in the wires.

00:03:16Let's use, you know, we can use a lot of electricity.

00:03:18Let's try 20 hertz.

00:03:20That didn't work.

00:03:21How about 30?

00:03:22And so, it's like our techniques tend to be fairly simple.

00:03:30And there's this term called a hardware overhang.

00:03:35And it's often used in sort of the AI safety space.

00:03:40They talk about a hardware overhang for AI or AGI.

00:03:45That if we only had the right software, well, our computers are already good enough to run it.

00:03:53And I think that today in neurotech, we have a hardware overhang.

00:03:59Our tech is usually pretty great.

00:04:02I mean, if you're following the Neuralink numbers, it's amazing.

00:04:10But we don't necessarily know what to do with this tech.

00:04:16So, one common theme, I think, that we'll see more and more as time goes on is that treasure maps are extremely useful.

00:04:32Because having a goal state, having sort of, okay, like, knowing what a good direction even is in the brain.

00:04:43And knowing what a good state is.

00:04:44I think that's just incredibly useful.

00:04:47And I think that, you know, things like Buddhism, yoga, etc. have these wonderful treasure maps.

00:04:54And this is one thing that I really appreciate about using the Jhanas as a treasure map.

00:04:59I

00:05:00think it's a fantastic approach.

00:05:05So, generally speaking, if we've spent many person centuries on building good maps

00:05:14of interesting states and wholesome states, that's worth a lot.

00:05:23My third claim is that every neurotechnology project is an implicit thesis on what kind

00:05:34of thing a nervous system is and what kind of thing a human is.

00:05:44And in kind of a compressed quip, I would say neurotech is

00:05:49perhaps the platonic form of applied philosophy.

00:05:53And that this also gives us some tools to critique neurotech projects.

00:06:01That if a neurotech project doesn't have a clear implicit thesis about what kind

00:06:08of thing nervous systems are, then it's at least partially a LARP.

00:06:14It's at least partially this sort of, well, maybe if we just stick some

00:06:19advanced technology in the brain, magical things will happen.

00:06:23And I'm not so optimistic about that.

00:06:26So...

00:06:31So, I guess, just frankly speaking, I think there's a lot of good things we can say about humans.

00:06:39I think we're awesome.

00:06:39I think we're beautiful.

00:06:41And more human, not less, I think is a very good sort of theme for neurotech.

00:06:49But this is somewhat or somehow under discriminate.

00:06:55You know, it can kind of point in many directions.

00:06:59So more on that in a bit.

00:07:05So there's a I believe it comes from Warhammer 40 k.

00:07:11This idea that if something doesn't die, shoot more bullets at it.

00:07:18And in medicine, this is called the Mordaka theme.

00:07:24And Sarah Constantine had a great framing on this,

00:07:29that the Mordaka story is common in medicine.

00:07:31You do an intervention, things don't get better,

00:07:34or only gets a little bit better,

00:07:36and the researcher says, okay, well, this didn't work.

00:07:40But often it's just, well, we didn't do enough of it.

00:07:43We didn't do it at a sufficient intensity,

00:07:45or we didn't layer different things on.

00:07:50And like antibiotics nowadays,

00:07:54single antibiotics often don't work,

00:07:56but triple antibiotics do.

00:07:59Cancer, you have, I don't know how many different cancer drugs

00:08:03at the same time that you just layer on,

00:08:05and it usually works that way.

00:08:08Same with HIV drugs.

00:08:10And with seasonal affective disorder,

00:08:13does anyone have seasonal?

00:08:15Seasonal affective disorder?

00:08:16Yep.

00:08:18So the finding there is that the lights that are sold

00:08:25for seasonal affective disorder just aren't bright enough.

00:08:29And if you just have four of them, or six of them, or whatever,

00:08:32if you get more light, it actually works.

00:08:37So I do think, yeah.

00:08:40I'm sorry, what?

00:08:42Do I?

00:08:46Oh, oh.

00:08:49Sure.

00:08:50Yeah, yeah.

00:09:00Okay, okay, great.

00:09:01Yeah, happy to, yeah, happy to chat.

00:09:04Thank you.

00:09:08Thank you.

00:09:09So one of the comments that I would,

00:09:15I would offer is that neurotech is generally thought of as,

00:09:20you know, you put like wires in the brain,

00:09:23and then you like do stuff,

00:09:25or you put like, at least you have to measure things with EEG,

00:09:29and so on.

00:09:31But like, I get the sense that successful neurotech,

00:09:37and I'm not quite sure that we've seen successful neurotech yet,

00:09:43but it might be pretty strange.

00:09:45And it might not be what we think of as neurotech.

00:09:49And it could be very simple.

00:09:51I really liked this.

00:09:52I don't know if you, if you guys can see it.

00:09:55I'll, oops.

00:09:57No, that didn't work.

00:10:01Okay, so Milan Sitovic had the tweet,

00:10:06the most potent legal neurotechnologies,

00:10:09dogs, music, caffeine.

00:10:13And I think that's beautiful.

00:10:15I think that we should keep an open mind

00:10:16about what exactly neurotech is.

00:10:23So, I say six, but it might be five.

00:10:28I don't know what I'm talking about.

00:10:28But I guess this, in a way,

00:10:31this is ideas that I hope people steal.

00:10:35In a way, this is sort of requests for startups.

00:10:40So, we can just kind of talk through these things.

00:10:47So, the first, oh, and,

00:10:53go back here.

00:11:04So, this is, as part of this framework,

00:11:08there's the insight, or the model,

00:11:11which is more philosophical.

00:11:12And then there's the intervention,

00:11:14which is the actual technology.

00:11:16And it's sort of, I try to present the insight

00:11:19or the model, and then say, okay,

00:11:21how could we operationalize this?

00:11:23How could we turn this into some sort of real tech?

00:11:28So, the first, I'm gonna call it an insight,

00:11:32is that there's this thing called fractal dimensionality.

00:11:36And it's basically the branching factor of a pattern.

00:11:41And, like, trees have a branching factor,

00:11:45like a fractal dimensionality,

00:11:48between 1.4 and 1.6, generally speaking.

00:11:53And, like, if it's two, then it's a complete

00:11:58line, and if it's one, it's just a line.

00:11:59So, like, 1.4, 1.6, it's like a tree branch, basically.

00:12:06And empirically, we humans really like things

00:12:10in the range of 1.4 to 1.6, exactly what you find in trees.

00:12:17And also, empirically, our nervous systems seem to be,

00:12:21to have this fractal dimensionality as well.

00:12:24It's called the Hausdorff dimension.

00:12:28And one interesting thing here, too,

00:12:30is that in studies of the environment,

00:12:36the finding is that being around trees

00:12:39has a very powerful antidepressant effect.

00:12:43It's just, you feel better if you can see trees.

00:12:49Whereas, interestingly, very interestingly, I think,

00:12:54grass doesn't.

00:12:55If you're on grass,

00:12:58if you have a nice view of a grassy field from your office,

00:13:03it doesn't help nearly as much as if you can see trees.

00:13:07So, I think that's important, actually.

00:13:11And you can do a similar analysis for sounds as well, sort of.

00:13:18So, interestingly, Jackson Pollock paintings,

00:13:22is everyone familiar with Jackson Pollock?

00:13:26He just kind of, you know,

00:13:28the story is that he just randomly slops paint on some canvas,

00:13:34and people buy it for a lot of money.

00:13:37But I think this is wrong.

00:13:39Because when you analyze the fractal dimensionality,

00:13:43like the branching factor of the lines in Jackson Pollock paintings,

00:13:47it happens to be between 1.4 and 1.6.

00:13:52And I think this is, I'm gonna have to say,

00:13:54profoundly important. So now we get to the part of the thesis here. Does everyone know

00:14:07what the Tetris effect is? Okay, so it's when you're playing Tetris and you're making these

00:14:15lines and you're getting the four dealer in the hole and that's kind of the purpose of

00:14:25your life for the duration of the game. Then you'll close the game, you'll go do something

00:14:32else, maybe you'll go rest or try and sleep. And when you close your eyes, you just see

00:14:37Tetris pieces. And it's like a very notable thing. Tetris is just very effective.

00:14:45It's very effective at eliciting this pattern. And I would say that this is a very important

00:14:54thing in that modern life in general, it isn't

00:15:00fractal. It doesn't have this fractal dimensionality of 1.4 to 1.6. And maybe this specific room is doing pretty well with all these beautiful paintings. But there are also a lot of sort of hard edges here. And our nervous systems really adapt to our environment. And I frame it like we pinch our sort of scale-free networks into

00:15:28scale-specific circuits to accurately predict things. And just modern life is just filled with sort of weirdly discrete dimensionalities. And I would call modern life just one big Tetris effect. And then so the one of the conclusions here is that one time or one reason we need time in nature is to reset this, to get back to sort of to pull our networks apart back into their normal

00:15:58shapes. So first idea someone should definitely steal. I'm calling it an Apple Vision Pro app for tracking nature seconds per hour. So imagine the Apple Vision Pro. Okay, you're wearing an Apple Vision Pro. And it can see everything you can see. It can hear everything you can hear. And then basically it adds up how much fractal dimensionality between this

00:16:281.4 to 1.6 range there is. And then it gives you a rolling score. How natural is your environment? How stressful is your environment? And maybe an office cubicle gives, you know,

00:16:43four hundred nature seconds per hour. Maybe a walk in the woods gives, you know, almost a perfect score. And maybe a tasteful home office or like this environment gives maybe, you know,

00:16:551,200 nature seconds per hour.

00:16:58So in terms of, it's not everything

00:17:03about what makes a natural,

00:17:07useful environment versus a stressful environment,

00:17:10but I would love to have this app.

00:17:13So please, please, someone make it.

00:17:16Happy to chat about it.

00:17:25So this is a hypothetical app, it doesn't exist.

00:17:29I hope someone makes it, yes.

00:17:33Yeah, nice, nice, yep.

00:17:37So there are also other possible variants

00:17:41for this long story, but,

00:17:46so the second insight that I would offer

00:17:50is that some people feel nice to be around,

00:17:54and there's this really,

00:17:55really important question, well, why?

00:18:00And I wrote an essay recently on this,

00:18:04it's called Presence Neurotechnology,

00:18:07sort of the minimum viable technology

00:18:10for recreating the felt sense presence of someone.

00:18:14Like could we make some technology that,

00:18:17where it felt like Shinzen Young was in the other room,

00:18:22or your wife was in the other room, or whatever,

00:18:25Like you couldn't see her, but you felt her energy. So,

00:18:30so that gets into the sort of technical stuff.

00:18:33But just as a minimum viable product,

00:18:37just kind of, not necessarily product,

00:18:39but just kind of a test of this.

00:18:41One thing that I hope someone makes,

00:18:44and again, this is just a bunch of requests

00:18:46for startups here, is an Apple Watch app

00:18:50that you install it with someone else,

00:18:54you sync up with someone else,

00:18:56and then it just mirrors their heartbeat

00:18:59watch. And whatever rhythm you're in, they feel on their watch, and you feel their rhythm.

00:19:08And I don't know what this would feel like. I don't know what this would do. But I think

00:19:14this is the sort of research that would be easy to do, and someone should definitely

00:19:19do it. So I think all the foundation is there. I just need to do it. Nice.

00:19:29and

00:19:30And yeah, there's more sort of sophisticated, complex approaches that you could take. I

00:19:39kind of like the idea of if you have a meditation teacher, if you sort of put a lot of electrodes

00:19:46on them all over their body, and then you have students meditating, well, what if you

00:19:51put electrodes on them as well? And then whatever the electrodes from the teacher are, you can

00:19:56measure, you sort of gently transmit

00:20:00to the students what would happen.

00:20:02I'm not sure, but I hope someone does it.

00:20:09So the third idea that I would share is have you guys, who here has read Neural Annealing?

00:20:21Okay, a decent, cool.

00:20:23So I wrote an essay back in 2019 entitled Neural Annealing.

00:20:31And the basic model is that if you have some piece of broken metal, like a broken shovel

00:20:40or sword or something, then you take it to a blacksmith.

00:20:44You say, hey, can you fix this for me?

00:20:48And what does the blacksmith do?

00:20:50The blacksmith heats it up.

00:20:53Until it melts.

00:20:54And the more technical term for this would be entropic disintegration.

00:20:58You add energy until the existing attractive landscape disintegrates.

00:21:04And then you hammer it into place.

00:21:08And then you let it cool.

00:21:10And depending on the rate at which it cools, some blacksmiths might choose to quench it

00:21:16in water or quench it in oil or let it air dry.

00:21:20You get different sort of microcrystalline.

00:21:23Different patterns in the material.

00:21:26But basically, you get a full reset in the microstructure.

00:21:30The sword is fixed or the shovel is fixed or whatever it was.

00:21:36And in a very simple sense, this is what I think is happening during meditation, music,

00:21:45psychedelics.

00:21:46Not only this, but I do think that this is a very good compression of that.

00:21:53So, yeah.

00:21:53So, I originally wrote Neural Annealing as a two part thing.

00:21:57One was the description of the model.

00:22:01And the second part was a product.

00:22:04Like okay.

00:22:05Let's turn this into a technology.

00:22:08And the idea that I had was music feels good.

00:22:13Turning up the volume on music makes it feel even better.

00:22:17That's interesting.

00:22:19But there's limit to that.

00:22:21You can only get so loud.

00:22:23are not so happy with it.

00:22:27But if you expand it to more senses and you touch more of the brain's sensory surface

00:22:33area, then that's like turning up the volume way over the perceptual limit.

00:22:42So we looked at light sound vibration, LSV devices, synchronized to music with the idea

00:22:53that you'd get a psychedelic trip without psychedelics if you just sort of amplified

00:22:59the music in this way.

00:23:01You can dig into a lot of interesting experiences, like going to a rave.

00:23:08Well, why is that so intense?

00:23:10Well, it kind of drives your whole brain with this pattern.

00:23:15So I think that you can just kind of naively take a piece of existing music and multiplex

00:23:21it.

00:23:22Or you can kind of...

00:23:23Do some feedback or choose, for example, some sound with a lot of the same sort of

00:23:34dimensionality motifs, we can say.

00:23:37Yeah, bioactive sound.

00:23:39Like a rave is an LSV situation.

00:23:44It's light, it's sound, it's vibration.

00:23:47I would say that it's a narrative situation and a very intense situation.

00:23:54And I think that's the reason why it's so interesting.

00:23:55Because it's sort of pushing the swing on...

00:23:58It's adding energy to the system.

00:24:00Now we can talk about brain versus body and embodiment.

00:24:05Frankly, the more energy you have in your system, the more sort of sensitive to sensory

00:24:12stimuli you will be.

00:24:14So the more things like that can get you even further.

00:24:19Okay.

00:24:20Model four, bioelectricity.

00:24:23So we have...

00:24:24My friend there, Ben, who knows a lot about this.

00:24:33And generally speaking, humans have a bioelectric field.

00:24:37And it's very important, significant for health.

00:24:42And kind of the more woo frame is we have an energy body.

00:24:49Well, yeah, we literally do.

00:24:52And I think that...

00:24:53Like the...

00:24:54The picture that the research is showing is that this is significant at

00:25:00multiple levels of the organism. So you have cells sort of function by the electrical polarity between what's outside the membrane and what's inside the membrane. And then organs also often have electrical gradients. And then your overall electrical field, it's looking like

00:25:28it's sort of a morphological field, I believe the term is. And basically, like, the signal that tells your cells, okay, grow an arm, grow fingers, but stop there, grow in exactly this shape, this is probably dominated by electromagnetic, local electromagnetic considerations.

00:25:57One core question for me, and this is sort of a precursor to talking about, well, let's, how do we technologize this? But it's trying to reimagine a lot of the normal, we can say normal wear and tear, or normal dysfunction as disorders of bioelectricity. And I think cancer is a perfect example that, yeah, actually, it is a disorder of bioelectricity.

00:26:25You know, we can talk about, you know, this mutation or that mutation, but actually, sort of the most elegant compression is that, well, something is wrong with morphogenic field. And my expectation here is that a lot of things are like this, that if we have this very central coordination regime for the body, for

00:26:55cells, for tissues and organs, and for overall morphology, then, I mean, it's got a lot of details. And for sure, things will go wrong a lot. And, you know, hopefully, it's got some error correction. But I foresee a lot of the things that we think of as very different, very disparate, not connected at all,

00:27:25will actually be connected. And then the question becomes, well, what kind of interventions are possible? And, like, one thought, and I'm, you know, looking forward to digging into this with Ben. But just like, so one kind of interesting observation here is that when you take opioid painkillers, you scar more. Your body

00:27:55becomes more vulnerable. And so, you know, I think that's a really good point.

00:27:55more scar tissue.

00:28:00And as far as I know, I'm not sure how widely it was tested, but it was framed as opioids. Part of the mechanism of opioid painkillers is it allows peripheral tissues or just kind of, it allows dropouts from the electromagnetic field. And so you don't get the pain. And I suspect that a lot of other things would be kind of fitting this pattern to you.

00:28:26I'm thinking about, like, cellulite. And maybe that's also a dropout from the EM field and so on. And so just in terms of thinking, okay, what kinds of interventions could you use on, like, the body's morphogenic field that kind of enforces shape and structure?

00:28:48I mean, if you break a bone, you can rig up a splint and kind of regrow, you know, kind of keep the body in shape.

00:28:56And so you can do a lot of things correct as it regrows. Could you also rig up an electromagnetic splint that sort of fixes your EM field in the proper shape as your tissues regrow?

00:29:11If I want to say one thing here, it's that if we could actually visualize the EM field, if we could see it with our eyes, we are very visual creatures.

00:29:23And we're our sort of intuition.

00:29:27And we can see it in real time in high resolution.

00:29:31So basically, whoever figures out how to rig up a system, maybe, you know, with the Oculus Quest or the Apple Vision Pro, where you could actually see the EM field, that would absolutely change the world.

00:29:49It would fix so many things.

00:29:51So please, someone do that.

00:29:55So just a few notes on the stuff that I'm working on right now.

00:30:02So last year I released a paper called Principles of Vasocomputation.

00:30:09And it's basically this thesis that tension regulates neurons.

00:30:15And that tension in an area, by cutting off blood flow, you force neurons, local neurons,

00:30:24into a safe mode where they don't update.

00:30:28So you can sort of think about the brain as a holistic programming environment where you

00:30:36can sort of catch patterns by clenching.

00:30:38This is not necessarily a technological thesis.

00:30:41I think that the body just uses this principle.

00:30:45There was a paper in 2007 called the Hemoneural Hypothesis.

00:30:51There were two very interesting observations made.

00:30:54And the first observation, which was pretty extreme, is that blood flow in normal brain

00:31:02tissue varies by a factor of 20.

00:31:05Not 10%, not 20%, but 20x.

00:31:08And then the second observation was that changes in blood flow precede changes in neural activity.

00:31:17So there's been follow-up work last year.

00:31:21And that confirmed the sort of causal thing.

00:31:23That it's not that neural stuff happens and then blood comes to feed it.

00:31:28It's that the blood comes and then neural stuff changes.

00:31:32I think that this is a very interesting sort of lens or window to look at the brain in.

00:31:39And that it sort of reimagines a lot of things about brain stuff.

00:31:47Yeah, yeah, yeah.

00:31:48So...

00:31:51Long story short, I think that...

00:31:55I mean, this is what I'm spending my hours thinking about right now.

00:32:00And the tech implication that I just want to kind of share, and it's also a request

00:32:09for startups, is that technology that releases tension is incredibly valuable.

00:32:16And it'll have sort of a lot of implications.

00:32:20Yeah.

00:32:23So, yeah, happy to chat about that later.

00:32:28So, okay.

00:32:29This is just a research stub, but happy to chat with people about it later.

00:32:37So, to wrap up, there are a lot of ways to frame what is a human, how to become more

00:32:46human.

00:32:48And I think that...

00:32:50Yeah.

00:32:51I think that it would be very interesting for people to sort of see...

00:32:55Like to maybe consider doing this exercise, maybe tonight, maybe tomorrow morning, et

00:33:03cetera.

00:33:05Just if you picked your three all-time favorite quotes about humans, about society, about

00:33:14the human condition, et cetera.

00:33:18And just try to write out what might this imply about what kind of thing a human is,

00:33:25and what might this imply about what sorts of technology could bring more of this into

00:33:32the world.

00:33:33So, thank you.

00:33:35Thank you.