34C3 ChaosWest - Hacking Worldviews Hacking Abstractions

How will you get the Q&A? Yeah, of course. Good luck. Thank you. Hello everybody at Chaos West Stage. I'm announcing the next... it's a talk, isn't it? Yeah. The talk. I'll try and go quick so there'll be a kind of questions. With a Q&A. It's gonna be in English and it's gonna be highly philosophical. So I hope you all had your two or three cup of coffees already. Please, take the stage. Thank you. Hi. So my name is Anders and I am a psychologist. So I do computers also, but it's not my main thing. So I was trying to think of how I could participate in this hackathon in more of a hacking context. And I realized that I actually do a lot of hacking already

because I study teaching and psychology. So teachers already hack their students. In the word pedagogy, how we teach, how we handle children, how we teach people. What is actually happening when you're teaching is you have to convince someone of a new perspective and you have to upgrade their perspective somehow. And what accounts as correct or counts as an improvement in someone's perspective is very contentious and that's what people argue about when they're talking about curriculum. But in terms of actually how to do it and why you might want to do that, I'm going to look at how you might hack an existing worldview in order to compromise it and then you can install a new upgraded worldview. So my research professor in my graduate program was

Ran Spiro and he, one of his major theories is cognitive flexibility theory and you know in science you have this two by two design usually and so the two variables here are usually ill-structured versus well-structured domains of knowledge. So we're talking about how people learn in either complex areas of knowledge or simple simple structural areas of knowledge and what his research found is that if the knowledge is well structured well structured forms of well structured forms of teaching that knowledge are highly effective if it's simple knowledge you can teach it in a simple way and the students will learn it correctly and remember it but if the knowledge is complex and messy and it's about making interpretations or it's

about judging maybe complex medical information and making a judgment call, then you can't train the knowledge in a simple way because it's not simple knowledge. If you do that, people will end up with misconceptions which damage their ability to access and use the knowledge. If the knowledge is ill-structured, you have to criss-cross the knowledge landscape back and forth and get it from many different angles, many different slices of the same knowledge. it will actually allow the person to form a robust cognitive network in their learning. So just like our ego has defense mechanisms, the partial inaccurate misconceptions or any type of knowledge has some resistance to being changed because otherwise our minds would be changing a

little too much. And my professor also theorized this as knowledge shields. So these are basically just defense mechanisms that kick in, you might misinterpret something, you might recall something that is related but not actually the same thing. There's a whole list of like 30 different specific knowledge shields that he identified. So this is another piece of the puzzle is if we are trying to learn something new and it triggers our existing knowledge in the networks of our brain, then those networks come online and start to process whatever information they've already learned in the past. So it might block taking in new information or updating those networks. So what this sets up is that we're actually, so this is like, you know, using the contemporary vaguely network model of the human brain

and how it stores knowledge as some type of imminent network or representation. So what this sets up is basically you have a network and you're trying to basically attack this territory and invade and infiltrate and colonize this territory with some type of new structure, some new pattern or network. So there's a whole bunch of different strategies you can use in that. You can surround, you can divide, you can you can weaken the foundation, you can come in from different angles and see different ways to infiltrate someone's existing learning and compromise it basically. How are you going to convince someone or really convince their brain that this knowledge is outdated and needs to be up-to-date. So something really interesting that happens in one of the

most effective ways of changing perspective is using abductive reasoning. So you know inductive, deductive and then there's also abductive reasoning that abducts the current thought and it's the answer the question is what could link here? What could be about this?" And so then the way that works in the brain is probably through spreading activation and possibly even backwashing from the postsynaptic to the presynaptic neuron. So we're seeing like this is how creativity works in the brain is we're seeing gradual spreading of priming and the priming and activation of related material that surrounds literally the the schema in question. So we're moving out into a broader perspective or a broader worldview and this is a really effective way of changing a perspective

because when a worldview becomes abducted into a larger worldview it has a rich interconnection with the previous worldview and what you'll see is that in a genuine shift of perspective you have a shift in your subjectivity and your language so if you you know and we'll get we'll get to an example in a second but if you are seeing something in a certain way and then you have a shift in your perspective, you'll literally feel and see a kind of shift in how you're seeing the world around you. So your language will actually shift and your language will be reinterpreted. Your subjectivity, you'll have a pivot in your experience of the moment. And then I want to maybe get to this later, but it re-transversalizes the replicativity matrix,

which is basically like how your brain thinks diagonally, will become up to date. And that's hard to talk about because it's not a thing that makes any sense. So how to abduct people into new worldviews. And one of the most effective mechanisms is allegory. And the etymology of allegory, it comes from Agora, which is an open space like a public market. And then Alla, which means different or something else. like aleatoric is like environmental sounds or whatever. So allegory means we're speaking publicly about something, but really we're speaking about something else. And that's right in the word. I love etymology. So when you have a story and that story is actually tracing

the contours of an existing network, you are mapping out your existing knowledge network. and at the same time applying different possible upgrades that link together into a higher order framework that collectively abducts the old network and makes robust connections previously back to the original network so that the new network is more stable. Yeah, so yeah, like that. So, wait, what is this? Yeah, so that first, that was kind of the first half of this talk and that is hacking worldviews. So how do we infiltrate a worldview and then change it to something else? I wanna make a bit of a distinction here

between worldviews and abstractions. So abstractions being concepts about concepts, how we think about our thought and how we maneuver and utilize these concepts, which is itself determined arguably by another concept. So whereas hacking a worldview is a very practical approach where you're thinking, you know, I have this person and I want them to see things differently. And you know, this is, it sounds so evil, but really this is from an educational perspective. Education means to lead out. So we're trying to lead people out of smaller perspectives or more limited perspectives into larger and larger, more inclusive perspectives that can also include the opposite of whatever they're thinking, for example. So whereas that's very practical and applied,

hacking abstractions is very abstract and it's more cognitive. How can we redesign a concept such that it is a more effective or more sticky concept than an existing concept? So this gets into the nature of critique and I've studied a little bit of critical theory and post-structuralist critical theory. and I'll mention a few names later, but basically there is, oh yeah, yeah, this is a good distinction. So the distinction between two types of critical theory, there's normative critical theory, and there's provocative critical theory. Normative critical theory attempts to establish a

recommendation of what we should do, basically. It'll say, this thing about society is broken, and we need to fix it and maybe we should do this. Provocative critical theory does not try to lay out a solution or a new worldview, a new scheme, or a new abstraction of what's going on. Instead, provocative critical theory utilizes language to provoke the reader and shift all types of different things potentially in how they're reading and how they're seeing the world. So provocative critical theory can be really fun to read and it can be very poetic. Lynn Fendler and Foucault are two provocative, critical theorists. So I'm just throwing out, I'm just making a string of concepts here. So enantiodromia is a Greek word.

And this means when something becomes so extreme that it becomes its own opposite or evokes its own opposite. And so this is what you see. If you take an abstraction, if you take an idea to its maximum point and you fully articulate the idea, I think that's the next slide. Then just merely articulating the full idea in all of its detail and pointing out the different aspects and facets of the idea, what it does is it actually is activating that entire knowledge network at the same time. So it's getting that all up into the air and floating around at the same time. And so then it's a lot looser because you know when we remember something, we re-save the memory so it can be changed every time we remember something. So it activates the knowledge and it articulates the knowledge.

So this is literally critique, is moving something back within its appropriate boundaries or its actual boundaries. So taking this neural network and using language to define kind of edges and boundaries and territories has an effect of moving that knowledge into a more whole version of itself by making these larger, higher order links and also linking it all together more wholly as a single unit of knowledge. And what happens when you do this is it tends to explode the network. It tends to create a little implosion, just like a koan would do, using, if you've read, you know, Hofstadter, like a koan using like a strange loop to create an inversion of subjectivity and a feeling of vertigo.

So Nagarjuna's fourfold method of negation, this is a famous Buddhist who, his method of negation, you may have heard it before, where it's this and also not this, and also this and not this, and also neither this nor that. So it's all four possible combinations of combining things at their either extreme. So you have both, you have neither, you have one and the other. And doing that kind of logic, you can get a sense of the non-dual logic that comes after the implosion. So there's a kind of a cognitive short circuit and you pop out of your current frame of reference and then from

that new perspective you have a moment of emptiness where you see that your previous concept, your previous abstraction was not what reality is but is actually something that you just popped out of because the frame interfered with itself. Yeah, so what we're talking about here as upgrading abstractions or hacking abstractions is where is the critical point? How can I, how can I, you know, how can I create an upgrade to this abstraction that's actually the taking away of something? How can I, how can I create something that includes this but actually speaks through it and across the previous abstraction so

that that new language doesn't deny the previous language but actually expands it inclusively. And how can I do that without adding any concepts to my students mind? How can I teach without actually teaching? Because I want them to be coming up with their own ideas. I just want to maybe say, you know, maybe that idea is going a little bit too far. Like let's, we can step that idea back and then immediately there's more room for other ideas to grow and to come together in new combinations. So this is a minimalist approach to teaching where I'm trying to not have a curriculum. You know, there still is a curriculum. This kind of thing is a curriculum, but there's no content. Rather, it's more the apathetic approach that Nagarjuna and Zen Buddhists in general take

where they tend to not provide solutions. They tend to be more like provocative critical theorists and they tend to, instead of solutions, just do strange gestures, provocative gestures that interfere with the existing modes of knowledge and the ability of those existing networks of knowledge to continue making sense. then they stop making sense and the student has to actually think about what might make sense. Oh, yeah. So what we can do now is we get into like the meme wars and all of that is predicated on how memes spread and how ideas spread. So if we can design effective abstractions that go beyond the current abstractions, we can reliably design memes that will be more likely to propagate and more likely to have

certain effects that we can see in advance and more likely to have positive effects because we'll be designing better ideas for people. So, let's see. Yeah, so if linking this back to the first part, if we tell a story and that story is linking together different concepts and those different concepts are increasingly articulating a larger concept network that is itself through that articulation becoming deprecated. What we have is the emergence of nonsense instead of concepts or like non-seps. We'd have this new type of thing

emerging between the layers of the other concepts and there's this emergent phenomenon of non-description where the actual the de-territorialization of this knowledge network starts to become a more dominant force than the territorialization of the network. So the wisdom starts to overpower the knowledge you could say. The flexibility of the thinker starts to overpower the hold that the individual concepts have on the thinker and that happens through this creation of rich complex connections that crisscross the knowledge landscape in many different ways and specifically if we're talking about a single event often some kind of abductive process that includes and involves the current mind in the new perspective. So I'm gonna stand over here

for this. There's possible applications and some of these are literally applications. I'm working on a contemplative mobile app that's just something really basic where you can type in words and they float around and so you can contemplate. You can allow your mind to wander but you can go back to the screen and remind yourself what you were thinking about. So this lets you, you know, prime yourself and explore your own associations. Also, in general, like this was something my professor was working on and he was doing these video mashups. Like he took some really famous movies and he cut them up into little tiny pieces and then tag to them really robustly. And then you could actually search for a tag, like if you were writing an essay or something,

and you could very quickly watch all of the clips with that one tag. So you would be re-watching the movie thematically in these little crisscrossing back and forth. But something that's even more exciting, I think, here is if we could algorithmize this type of protocol upgrade as a protocol, then you could make self upgrading protocols or you could make like a storytelling AI. It wouldn't be an AI though because it would have basically a simple math of how to apply different transformations in different contexts in order to produce certain types of concepts that would involve certain types of readers. So you could you could create a therapeutic storytelling bot

basically, that would actually understand in a sense where you were coming from because it would take in some type of background information and then process that according to what it thinks your current concepts are and how it thinks it might be able to sabotage those by exposing them to each other at specific points. So this gets into what would that logic look like if a computer could process this type of storytelling logic. So it's not a limited alphabet. It's not an alphabet of 26 letters. If you split the seasons or a narrative with rising action, falling action, rising action, and then a climax and a denouement, then you get something where you can

map all types of cycles onto one wheel. But how many slices should we cut this pie into? The English alphabet has 26 letters, the Norse alphabet has 24 letters, Hebrew alphabet has 22 letters, there are four seasons, seven classical planets, eight or nine actual planets, so how many different alphabets do we need? How can we go from an alphabet that has four letters to an alphabet that has five letters? You know if we have, you know, conflict resolution, conflict resolution, conflict, that's not a complete story, so we would need to have conclusion and then have another conflict or a stage that's what a conflict and a resolution in one step if we only have four or five steps. So this

is super abstract but I'm trying to talk about how you could build an abstraction processing engine that would be able to reliably build executable wetware upgrades for people so they could see things in a new way and have a reliably more open abstraction. So if you're interested in these topics, here are a few things you can read or look at. I wrote an essay that's on my website, Infiltrating Misconceptions, where I talk a little more about how the neural networks might work for this. There's also The Glass Speed Game by Herman Hesse, he won the Nobel Prize for that. Santos Bonacci. If you thought the previous slide here was interesting, then Santos Bonacci would delight

you. He just rambles for like an hour and a half about that. And also CCRU.net is amazing. The Pneumogram, I'm writing a book on the Pneumogram. It's on my website also. And then The Politics of Aesthetics by Jacques Ranciere. He talks about the distribution of the sensible and a redistribution of the sensible where we suddenly perceive a new way things could be or a new boundaries about what we're able to actually speak about. And here's my contact information. My website is myname.com. I'm working with Holo on Holochain and that's a fully distributed peer-to-peer application ecosystem. So it's like micro blockchains for that. And we're running an Indiegogo campaign that's going really well.

And I also have internetschoolofmagic.com. I'm one of the many co-founders of the Internet School of Magic. And so you can read classes about, unfortunately, not the transversal replicativity matrix, but you can read classes about some of the letters of that alphabet. Yeah, thank you. Thank you very much, Andrews. We now have a Q&A. Is that cool? Yeah, that would be great. If anyone wants to ask questions or raise discussion points, please. I can talk about the transversal replicativity matrix. Maybe. So you said this sentence, how can I teach without teaching? Why is this of motivation for teachers?

That's a really good question. I guess I have two answers to that question. And they're both about the way curriculum works, where right now teachers are used to convey someone else's material, someone else's content, you know, in most, almost all school systems. They're just basically using the teacher as this, as this puppet, and they should not be used that way to replicate someone else's ideology. And so first of all, it's if we were going to do a better system, if we were going to allow teachers to actually teach as human beings primarily, what would that look like and how can we learn that skill? And then second is given the way things are, how can we teach without teaching? How can we teach this curriculum but

actually be teaching something completely different? While we're ostensibly teaching about history, we're actually teaching, you know, something, well history, you know, that's just an example because it's often taught in a boring way, but what if when we're teaching about history that would normally be boring, it's actually really interesting and or teaching something entirely different about how to think and how to be in general, which is what good history teaching should be like. Thank you. The other question is, I'm interested in learning. Can you give me some hints how I can learn stuff that I'm interested in very quickly? Can I hack myself? Yeah, my research professor's favorite method for that was to use Google to explore the

deep web. So looking for specific, really specific keywords, really specific articles and very rapidly going through a lot of different questions that you have. So asking yourself questions and pursuing your line of questioning rather than pursuing the reading and the links that you find. So always going back to Google right away and asking the next question that occurs to you. And you can very rapidly iterate your learning that way. That's one idea. That's cool. Any more questions from the audience? What do you think would be the best ways to measure the effectiveness of your interventions on worldviews? That's a very empirically framed question.

And I like to deconstruct that type of worldview. But let me answer your question. Because it's a good question too. I started designing a, I called it SUP, because it was like a subtle user probe. And so it's a little, it's like the Cortana thing they have in Windows now, the little box where you type stuff in. But this box would be just for typing in what you're doing. So it would be a personal way of logging your activities. And it would just like blink a light once in a while, just very politely, and you could be reminded to type in what you were doing. And then over time, you would be able to use this to kind of connect different contexts together

and see how your connections of contexts are changing over time. So this is not an easy thing to measure. That's a really hypothetical, subtle experiment you could do that might have interesting data or it might not. Another example that's similar to that is you could, for example, look at, you could compare multiple users' search history. So you could compare, just say two users, search history, and the user that's more likely to be using more higher order abstractions, you'll see more, even if you just did a simple statistical, correlational measure of how related their search terms are over time, the person who's using higher order abstractions will more likely have higher correlation

amongst their questions that they're asking Google or whatever search engine over time, for example. But yeah, it's really hard to measure. And I think also effectiveness is close to the root of like the issues in public education systems where if you can't measure it, then we can't fund it and it doesn't exist. So yeah, and I was in a PhD program for educational psychology and educational technology. And I realized, well, I won't say what I think about it, but that field is actually doing the testing. It's creating the tests, it's designing the science, it's designing the evidence that is then given to politicians to justify standardized curriculum and centralized education and testing systems. So there are particular people who are funded, who are well-funded and they're creating this paradigm and they're using that language.

So I try to differentiate the language from the people who are creating the problem. Great question, thank you. Did I understand you correctly, that I should train my Google search machine to search better, find better when I ask questions? No, I was saying that if you looked at someone's search history, they would probably, you know, in theory they are probably using more higher order abstractions if they have more correlation in their search terms over time, whereas if they're just scattered, if they're asking a bunch of random questions, they're probably not working with a higher order abstraction. So if we have more questions then where can we find you? Where do you sit? Oh I don't I don't have a table I'm this is the first time I've been here so

I'll be wandering around but my number is 3314 and if you contact me on Telegram I'll respond same day at least. Andrews amazing thank you for this talk. Thank you so much. Thank you. Thank you.