 Hello and welcome everyone to ACTIMF lab. This is ACTIMF live stream number 21.1. And we're here with John Boyk and several other colleagues looking forward to a discussion today on this science driven societal transformation series and project that we've been exploring and had several background sessions.01 through .04 with John. And in the discussion today and next week we're going to be looking to draw out some threads and include some new perspectives and hear from the questions that people may have. And for these discussions, we're going to be using the same jam boards that we were using for the discussions with just John and I. And we're going to jump around the jam boards. We're going to hear questions from anyone who wants to ask live or in the YouTube chat. And we'll have fun. So we'll just start with introductions and then get a quick little recap and off we go. So I'm Daniel. I'm a researcher in California and I'll pass it first to Alex. Thanks. Hi everyone. I'm Alex. I'm a researcher in systems management school in Moscow, Russia, and I pass it to Stephen. Thank you. Hi, I'm Stephen. I'm based in Toronto. I do a lot of work with participatory theatre and community development and I'm researching how social topographies can be used across scales of community psychology and other development practices. And I'm going to pass it over to Lee. Thanks, Stephen. I'm Lee. I'm a researcher based at the University of York and I'm studying systems transformation. And my supervisor is, is Jo and Faisy, who I think you draw on a little bit, John, in a couple of these pages. I was really, really happy to find someone else looking at systems transformation and active inference at the same time. But although I think you're a lot further down the line than I am in thinking about this, I really look forward to this session. Cool. Well, welcome, John. Maybe you could give an introduction and start things off. Sure. Thanks for having me and thanks everybody for joining us. My name is John Boyk. I'm a courtesy faculty at Oregon State University. And the purpose of these talks is to discuss the series of papers that I just recently published in the Journal of Sustainability. The title is science-driven societal transformation parts one, two, and three. And if you would go, we're going to do a quick recap of some of the material that we've discussed before. We'll jump around a little bit just to catch some main highlights. And of course, feel free to jump in with any questions as we go. So if you could turn to Jamboard 1, slide 1. Yeah, you got it. Okay, good. So this just is the pictures of the different papers. The links are done on the lower left. So in case anyone wants to check out the papers, please do. They're all open source, freely available. And we're going to focus today. We'll jump around a little bit and capture the main topics, but we're going to focus on part three design today. All right, slide two, please. And as we go through this, if there's more questions, if someone wants more materials, all of the links are available, all the links to all the papers and some other materials are available at my website for this project, principlesocietiesproject.org. And I might mention there's also a interactive model there that you can, the leader frame, which we'll talk about. So you can see how the economic simulation goes. Slide five, or slide four, excuse me. Yes. So what is the purpose of this? What is the problem that this series is trying to solve? The problem that faces society or really human civilization is a host of very large so far intractable problems, including climate change, biodiversity laws, poverty, income and wealth inequality, and soil depletion, groundwater depletion, pollution. I mean, it's a long list of very serious problems that we face. So serious, in fact, that it's not clear at this point how civilization will fare in the coming decades, even in the coming years. I mean, the world is armed right now, the world is armed to its teeth with nuclear weapons. So, you know, any moment things could go awry or things could just deteriorate rapidly as climate change and biodiversity laws gain speed. So hopefully we will solve our problems, hopefully as societies. Hopefully we will improve quality of life and quality of the environment. But that is what we're shooting for in the series is pathways to do just that. I would like to point out that this series is focused on transformation in the sense of de novo design of fundamentally new systems, societal systems, which I talk about six, I believe. So economic systems, legal systems, financial systems, governance systems, things like that. So the problem is the major constellation of problems that we face globally and the proposed solution is to view society as a cognitive organism. And just like any organism, it has an intrinsic purpose of achieving and maintaining vitality. So how does it do that? How does any organism survive? It develops, it learns, it acts, it learns from its actions, and it adapts to new conditions as necessary. This series is actually an R&D proposal. It's aimed at the global scientific body and my hope is to interest the scientific world in this problem of how do you transform? How could we build a better society? If we were to create any conceivable way to organize society, any conceivable economic systems or governance systems or legal systems, educational systems, what would be the best in terms of quality of life and quality of the environment and long-term sustainability? I think that's a question that science is now capable of addressing. Perhaps this is the first time in history that science had the tools and the skill and the background and the computational power to address these kinds of very, very large questions. But I think that we can do it now. The Pope R&D program is described as a partnership between the global scientific community and local communities. So the idea here is to build, create, and test new systems at the local level in cooperation with local volunteer communities. And I talk about implementing those systems and implementing field tests via a club model, which we'll discuss a little bit more. Okay, so that's a very quick overview. Maybe on to slide five. A first, Stephen, with a question, and then anyone else who's raising their hands. Sure. Just keep joking at the hands. Yeah, go ahead, Stephen. Thank you. Yeah, just one question. Unless you mentioned architecture. So you've got the idea of architecture and you've got these different things that we have got in the constellation. And I wonder if you could speak a little bit more to the idea of architecture and how that relates to architectures in the brain or maybe distributed systems of the brain. The brain is what I'm thinking of when I use the word architecture. Well, a little bit. I mean, people talk about computing frameworks and computing architectures too. But I really had in mind the way the brain functions. And maybe I should make that even more broad because if climate change and biodiversity loss and these other major problems that we face are symptoms of a cognitive dysfunction at the societal level. I mean, that's what I would call it. The societal dysfunction is a cognitive dysfunction because we've known about these problems for decades. With climate change, we've known about it for more than 50 years. And yet we failed to adequately address the issues. So I would say that all of these major problems are symptoms of a cognitive problem or symptoms of a cognitive dysfunction. And then the question becomes, OK, well, if we're not cognating correctly, where do we look for better models? Where do we look for ideas of how we might cognate better? And I would say that in the world of biology, including the human brain, there's just a million or a large number of examples that we can learn from. How does nature compute? Nature has been at this for a very, very long time. And how does it work in nature? And maybe we can align ourselves with the same program, the same process that's going on all around us. And if we do, then I would say there's a chance of building systems that are much more effective at achieving the intrinsic goal, which is to achieve and maintain vitality far into the future. Yes, Stephen. And then Lee. And I was just going to sort of building on that because with active influence, one thing that sometimes comes up in the, you know, is that rather than the brain being like a computer per se, that it's a distributed and actually now maybe even not even functionally specialized in certain regions. But you know, it's more of a distributed system of which, you know, we're not exactly sure exactly how that works. The idea of it being a dynamical sort of prediction engine. So I'm wondering how that could, how much might that inform where you go with this? Oh, yeah, totally because we're using the word brain. But even when I'm thinking brain, I'm thinking whole body and bodies connected to each other, other bodies in society. In part one of the series, the title is called World View, and I try to set up a context for all of this. And as part of that context, I point out that everything is intelligent, like every piece of life is connected to every other piece of life, you know, either strongly or weakly. And the whole of life is processing information. I mean, that's essentially what life is doing is processing information, reducing uncertainty about the past or about the future from the past. So, yes, absolutely. You know, we have the four E's, for example. And we have distributed, we have, like in part three, I use the analogy of between neurons and individuals in a society. And so how do neurons associate with other and disassociate with other? There's kind of a complement as to how people associate and disassociate with others in the process of societal cognition. And so we have connections in the brain. We have connections to our bodies, connections to our emotions, connections to our feet and fingertips, and connections to everything that is happening around us in our media environments. Connections to each other, and all of that plays a role in cognition. So I'm not, yeah, so I'm talking about cognition in the large sense, not in the narrow sense. And I'm certainly not talking about a computer model that tells everybody what to do. That's not at all what this is about. Thanks, John, Lee, and then anyone else who raises their hands? More of an observation, really. It's just when you were saying that you were drawing on the natural biological systems for inspiration, it kind of occurred to me that then what you're doing is a kind of a biomimicry of cognition. But building on Stephen's point, it's not in the older sense of the mind is computed but much more in the newer sense is biology at all levels being able to cognate. Right, absolutely. I think I read by Buckminster Fuller, he said something one day, he said, I think I'm a process, not a thing. I'm paraphrasing there, but that's right. It's like all of life is a process, an intelligent process. Cool. And those, I'll say it again, but those ideas really are new to science and what the mainstream science, new to mainstream science in just the last two decades or something like that. Fairly recent concepts of complex systems, cognitive systems, the interconnectedness of all things, the intelligence of all things, trees communicating with other trees through soil fungus. All of these, many of these ideas are actually quite recent in the science world. And a funny note on that is that mainstream science is new in the social cognition world. The professionalization of science is a new topic and a new genre. And I really like some of these points about distributed systems. We can think about distributed systems within one level of analysis or one system of interest like within the ant colony, it's a distributed system or within the brain, it's a distributed system. But once you start thinking about brains, John, you mentioned the 4E, the extended, inactive, embedded and cultured. It could be more E adjectives, but that brings out the richness and says, yes, it's distributed system amongst the neurons in the brain. And then the brain is interfacing with the body and the body with its niche and the tools and the social affordances. And it's starting to sound like a lot of topics that we've talked about on active live streams. So, so awesome to see how we're taking some threads from the past and weaving them together and I hope that we can draw this out how a multi-scale distributed systems perspective on cognition coupled with a social transformation agenda and preference vectors representing policy can get combined through active inference into something that's applied. Right, right. How to act, how to make decisions under uncertainty. That's what all of this is about. As a counterpoint to some of that or as a different view on some of what we just talked about, imagine if we can look at how our society treats individuals. I mean, we can look at extreme inequalities of wealth and income, different levels of access to healthcare, like there's all these social issues that we face right now. Imagine if your brain acted that way somehow with different neurons where some neurons were just simply starved, some neurons were excluded from the computational process, some neurons were given very few resources just barely enough to survive and others had so many resources that they were like wasting their resources because they had so much. I mean, imagine how well the brain or how poorly the brain might function if that's the way neurons interacted, if that was the community of neurons was anything like the community of individuals in our society, we wouldn't be able to make hardly any good decisions, we'd be a mess. And we are a mess. It reminds me of Dennett's fame in the brain model and the firing rates of neurons are not identical. Some are firing many times per second, others fire not even in a day perhaps. So there's a lot of distributions in these complex systems and how do we think about the whole distribution as an outcome of a complex process rather than just trying to identify some sort of an outlier or some sort of single phenomena, it's not the leverage point that system interventions will be effective at. So, Stephen, and then any other raised hands? This also ties in I think nicely with the idea of post-normal science the idea that in the post-normal science world where in high levels of uncertainty you have to make decisions, so COVID is a classic case, right? They had to make choices in January, February, March, April, even now and you don't have the data, right? You don't know. And the same way with active inference, the brain and different parts of the body are making action policy selections as is going to inform action and that's not quite the normal realm of where science is. This is where we're going to be going as we hit climate change and all sorts of other complex questions. So that also is quite an interesting tie-in because, Christon, I don't know if you know, but he did some modeling on COVID-19 as well. He's gotten into that. Yeah, John, maybe you could unpack that in the context of second order science, which you discussed in the papers? Yeah, or maybe I can just pass it to Lee and he can say a few words. Yeah, he's the expert here. No? No, I'm happy to. So the concept of second order science really is just that the idea that the scientists can't be separated from the systems that they are studying. So really where that's kind of led to in systems transformation research is that the only way that we can really understand systems is by engaging with them and learning how to change them ourselves. So we can't kind of stand back and just try and focus on epistemic knowledge alone really. And that also implies this kind of second order understanding of ourselves and how our own cognitive systems work as part of the systems that we are engaging with. Yeah, I hope that was a bit... Value-driven science where the scientist is part of the experiment with the stakeholders. Absolutely. Lee, while we have you here, that paper by Faizi was authored by a very large number of people. Can you just say if you know the story behind that? Well, a little bit. I think it was authored before I started studying with him, but there's a transformation community really, I guess, of researchers. And I think it probably came out to one of the conferences that's been going on. So it's sort of the product of a discussion between many researchers, which is the way that they kind of prefer to write really. Moving away from this idea of the single genius to let's kind of together and figure this stuff out together. Right. I referenced Faizi's paper several times in my part number two. So readers can see how that fits into the bigger picture here. And it's funny that that idea that the product of research is collaborative. It's a colony phenotype. It's a group phenotype. How well this paper is and how many people cite it and how much impact it have, it's not going to be reduced back down to a single person. And so let's take that. And then just like you brought up, Lee, the implications of learn by doing is that we have to infer by acting, basically. So then what does it mean for the scientist who's all in on that to be studying climate change on a coast or the conservation of a forest or sociology or anthropology? Very interesting questions that I hope a lot of different listeners can see resonating with their own situation. So Steven and then any other raised hands. I think just also building on the idea of this distributed input is and in many cases it may end up a lot of the may end up that a lot of the work is not in science anymore. It's like science informed development or science informed or even science informing a model and the model informs practices which inform the way that communities can get involved. And now the real agent is actually the community and we're serving them. So this interesting shift between where science fits in and can change. I think that's quite interesting as well. I could have titled the entire series Science Influenced, the transformation because that is the concept I was going for. I called it science driven but my idea there was similar. And also the historical perspective, which includes many world knowledge traditions and hundreds and thousands of years and multiple continents, not just the story of Western academia, understanding that that has not been a steady state. There's no stationarity with peer review. Again, mainstream science, especially the second half of the 1900s and further, that's new. So science is in transformation and has always been in dialogue from 10,000 years ago. So now, given the agency that we have and who we are and our affordances, our preferences, what's the move? So we don't have any sort of a locked in phase to resist against. That's not to say there's not inertia, but it is to say that there's actually a lot of agility in the system. The question is, what are the leverage points where we can actually make that impact and how my active inference come into play? Maybe since we're on the topic, maybe we can just switch to a jump ahead in just a little bit. I can find it. You find the slide. Or you find the slide and then Lee, go for it. Go ahead, Lee. Just to pick up on your point, Daniel, really. I've kind of been hesitant to say it. In the active inference, this could serve in the light, but I see such a huge amount of emphasis on active inference as a means to create AI and to improve AI. I really spend a lot of time thinking about how can this knowledge be used to improve our own individual cognition and societal cognition, human cognition. Maybe I'm just missing some of the literature on that. I don't know if there is any point me towards it, but it really seems to be like the whole big assumption is we've got to put this stuff in AI because that's where all the progress is going to be, et cetera, et cetera. And I kind of wonder, is that really a wise goal? I don't know. I think active inference and several other paths, so not exclusive to active inference, it rethinks artificial part of artificial intelligence because it's part of our extended niche. And so, yes, there's been a lot of focus on applying active inference to machine learning and to reinforcement learning, those types of areas. And then what we're talking towards here is the integration of natural and artificial intelligence under a broader extended cognitive systems scope. So humans in the loop or real-time computers, just like every stoplight or every bank transaction is going to be, every tweet is going to be designing those interfaces. So that within each blanket, there's functional cognition not viewed only from the outside in, but as part of a broader system. And maybe we find out that we had good enough artificial intelligence, so-called algorithms in 2004, and that we just had to hook them up in a different way, just like connecting different parts of piping could give you an engine that's more or less efficient. Maybe it wasn't the case that we needed some sort of scaling on a larger dataset, but we needed a new way to combine subcomponents, which is what biology does as well. So, yeah, John. Just a few more thoughts on that. So the science world and particularly this cognitive framework or concept of what is cognition, I think can be useful both at the societal level and at the individual level. I mean, in the sense that if science can help direct the attention of the world to how intelligence happens all around us in nature, all around us every day, how we evolve to expand on the information processing, the capacity for information processing. So it helps. And it just helps an individual to understand a little better or have a different perspective on, what am I doing? Why am I here? What am I doing? Why am I? What am I? To a degree that these new developments in cognitive science can help to shed some light on questions like that. And at the same time, for society, like what are we doing here together? What is our purpose together? Right? And if you know your purpose, it's a lot easier to know where to go next, how to frame things in a way that is suitable for your continued well-being. And it's a bright line in some senses between first and second order science. In the first order science literature, there might be a lot of sort of engineering-type appeals to, well, we'll make the society more efficient or science-informed policy will just simply be better. And once we take that second order turn, we start to think we're going to have increased clarity about our values. It's not going to be an easy conversation. We might need to invent new ways of having the conversation or formalizing our values, but we need some sort of path that's going to bring us closer to clarity on our values rather than obscuring it with technical advances. So, Stephen? John, go ahead and then Stephen. I was just going to say, obscuring it with technical details, but also obscuring it with incompatibilities between what the science world is saying and doing and what, say, the governance world is saying and doing or the economic world is saying and doing. Obviously, if all of these systems are part of the cognitive architecture, you want to design them so that they are integrated with each other, so that the signals travel from one to the next, just as the signals in your body travel from one system to the next system. Yeah, Stephen? And I think this also ties into partly what this lab has been talking about quite a lot in trying to enable this possibility of making those transitions without losing the rigor of the modeling, because the modeling is seen as almost another way to get empirical data. It's not quite the same as science, but the challenge you have, as things unlike traditional science going towards engineering, you do have this ability to embrace effects and values, but the challenge is how to stop things getting prematurely reified. It's so easy things get reified and suddenly, effectively, you're just using another word for system boundary, and you're kind of really doing the same paradigm work with some different window dressing, right? So it's really, that's the challenge, I think, and I don't think anyone's quite managed to do that yet, but I think that's definitely... No, we're on the cusp. This is happening right now. This struggle of what does it mean and how do we do it? And the premature reification, as you said, Steven, it's basically, there's a perfect investing strategy. You know, if you had the time machine, you'd know exactly when to buy the dip, sell high. We don't have that. We're moving forward. We don't have access to the future. So there's two kinds of confidence errors. There would be being overly confident, so too early zoning in on a solution. That's the premature reification. And the other fallacy or the other side of that perfect strategy would be waiting too long. And so somewhere between waiting too long and acting too fast is that trade-off space that we have from so many different perspectives with not just finding the value of explore versus exploit that's gonna be best, but rethinking the explore and exploit outcomes as part of a process that might be doing something like balancing epistemic and pragmatic value, for example, in the free energy principle. So it's a really awesome conversation. And John, continue. And altering how you measure fitness relative to those two differences. Well, since we've been talking about science, the role of science in this, we jump ahead to slide two, or jailboard two slide 16, if you would. Yep. And while you've... Oh, okay. That's fast. And while we get that rolling, just one more thought on second order versus first order science. You know, phases, papers, great. Really, I think an important one. But it's new. I mean, it's new to the science world. The science world is very focused on what we would call first order science, where the scientist as a part is an observer of a system, not engaged with transformation of that system or achieving value driven goals in that system. And first order science is fantastic. We have learned a tremendous amount of knowledge, gained a tremendous amount of knowledge from first order science. So it is fantastic for what it is. And, you know, has to be continued. But in addition to that, we can also start to be thinking about what is second order science, how can the science world be engaged in a value driven effort towards some better future. And I suspect that there are lots of scientists who are frustrated, who are frustrated because they don't know how to... There's not an avenue for them to participate in really jumping in to make this world a better place. Intelligent people look at our world and they see the enormous stress that we're under in this moment and are fearful of the even larger stress that is likely about to be fallless. So I think that there might be a tremendous amount of frustration in the science world of how do we use our capacities to actually fundamentally make things better as opposed to just passing off a, you know, sort of a little electrolysis in it a bit here, but just passing off a recommendation to some policy makers who then do whatever they do. You know, the scientist is then out of the loop. So I think that this is the time for second order science to flourish. This is the opportunity for the science community to really get engaged in these deep and important problems that civilization faces right now. And in the series, I'm trying to propose an R&D program that is a path for the science world to get involved in these deeper questions and have a deeper impact on civilization. So this slide is kind of a summary from my paper of why should science, you know, why should science care about this and why should science get involved. And we can talk about any number of these, and I'll just sort of jump through them a little bit, just to highlight a few ideas. The very first one is just this idea of societal cognition. That's, you know, that's kind of new. There's great potential for scientific discovery just on that question of what does society cognate? How does it cognate? You know, in what ways, by what mechanisms, by what features, with what characteristics, I mean, there's just, you know, just as far as expanding the level of knowledge, this question is really interesting. The second one. So we're talking about designing societal systems, and obviously this is a really complex issue. These are not, this is not a simple, how do you build a better can opener, right? This is difficult and complex issues. And the science world has skills that can be really useful in addressing these questions and framing the movement towards how do we start? How do we test? How do we measure fitness? What is the goal? What is the purpose? What are we doing? You know, like the science world would have a lot to say about framing this entire question of what is societal transformation and how would we move forward with it? I might add too to this that, as I see it at least, one of the benefits of having the science world involved in this question of societal transformation is that the science world can provide sort of objective and consistent metrics and frameworks to help guide the process. That is, I turned around, if this process of societal transformation, this program of societal transformation was merely based on what a group, this group votes for this, let's do this, and this other group votes for let's do this, and it was just a popularity contest of whatever was most popular in the moment that's where transformation is going. I don't think the science world would have any much interest in that, like popularity contests or that kind of effort. Where is the knowledge and where is the science in that? On the other hand, I think it would be very interesting to the science world. If transformation was based on some kind of principled approach, like we can agree, to an extent at least, we can agree what fitness might mean for society, and some of that would be based on objective measures, some of it might be based on subjective measures, but intelligent people can get together and conceptualize what fitness for a society or for a societal system might look like. These are, of course, complicated problems, deep problems, but if we could do that as a community, if we could find some way to assess fitness in a reasonable way, a reasonably coherent way, a reasonably consistent way, then we have a path to move forward of different groups could try different ways to reach a high fitness, and all of those experiments would be adding to the body of knowledge of how do we do this so that fitness improves, right? I don't mean to say that coming up with fitness scores and fitness evaluations of new systems would be easy, but it is a topic that we can tackle, and it's because of our capacity to conceptualize fitness and formalize fitness in the sense of a formal assessment, I'm hoping that makes the whole project interesting to the science world, because now this is science, now this is expanding knowledge, this is using first principles and scientific principles to evaluate systems, and it's not a popularity contest, but this is adding to our understanding of how the world works, and how intelligence works, and how cognition works, and expanding it out to the societal level. I don't know, maybe there are some thoughts on that alone. Yep, Stephen, and then anyone else who raises their hand? Yeah, just a couple of questions around the idea of fitness landscapes and how that might be used, I sense you're sort of proposing that you could use these kind of fitness landscapes, I suppose you could even have descent on free energy landscapes to try and help inform policy selection, i.e. maybe which research streams you're going to go down, because at the moment, we're kind of either tied into what the market says or some other methods, because maybe you could sort of separate out, because I could see a couple of different ways you could look at that then. Is it like an extra field of practice to inform choosing the policy of which research field to go down and which avenues, or is it also within the field? Is it with the fields connecting to each other? I just wondered how that might be. You know, it's all of those, because everything that we do ought to be serving our purpose. Obviously, there's a great room for play and exploration and all that kind of stuff. I don't mean to make humans two-dimensional or something. Humans are complex organisms. They have complex needs, and all of those needs must be addressed. But the concept of fitness, I see that as permeating multiple levels, because yes, we would want to choose policies that help us. And then what does help mean? How does it look if it helps us? What is that all about? In some sense, if we just say, I want to help, we're already thinking I want to improve fitness somehow. But not just policy, but even structural design of what is a system? How does society make decisions? How do we communicate? What is our communication systems? How do we get input from everyone into this process of decision-making? You know, like so across the board, both in the structural design of systems and institutions, and even to some degree, social norms, right? Through education systems. Like what kinds of social norms help us? And then obviously too, then in policy, but then even in our individual lives. Like it helps, I think it might help individuals to understand that what we want is improved quality of life, lower uncertainty. How can I make decisions in my own life that bring me the things that I think might actually make me feel better, that actually might serve me on an even a superficial and deep level, right? Thanks, John. Lee, and then Stephen. Yeah, just some thoughts on the fitness metrics really. And this kind of goes to the point about the way that scientists need to engage with local communities, because there's quite a lot of work that's already starting in that area. And a lot of it is based around this idea of place. So scale is already being taken into account and this whole kind of sense of the emerging bioregional movement which just essentially means aligning your political and governance boundaries with the boundaries of ecosystems. And that then naturally has the effect of it partitioning in a way that's meaningful to local people. And then you can start to ask them, what is it that they value within their place, because they're already kind of connected to it in a way. And I mean, I suppose one thing for me is this idea of metrics really versus counterfactual simulation because once you start putting a metric on something, you have a number and then that's something that people can't really care about, whereas if they can imagine how their local river system, for example, wouldn't be polluted by sewage being released from the sewage works or run off from them, that then has the effect of engaging people more in the actions that we would need to collaboratively take to make these kind of things happen. Sure. Different people might react in different ways to some people might be more, you know, really appreciate the modeling aspect, other people might just appreciate the boots on the ground, local change aspect, but I would say there's room for all of that. And the bioregionalism concepts are great. I've been a fan for a long time. And who knows, maybe in the future we'll move more and more towards that so that communities are really localized in their environment and protecting their environment and all that kind of stuff. That's great. In some ways, science has to catch up with communities of change who have been working on this for a long time, these concepts. And in other ways, science is already, you know, pulling communities, you know, helping to say more, there's more, there's more. There's more we can do. There's a bigger picture we can serve here. So it'll be a give and take, I think, between communities. And this one more thing is the bioregionalism idea is great. And a lot of work also has been done and alternatives to GDP. So, you know, the metrics that people are talking about, about, you know, considering education and longevity and disease and all these other things as opposed to just a single GDP measure of societal quality. So there's lots of work being done on multiple fields. And to speak to the metrics from an active inference perspective, let's just say that we wanted a river that was unpolluted. We wanted no toxins in the river. So our preference in the active inference model would be that the river is clean. That'd be in our preference vector. Now traditional reinforcement learning would look at the level and it would reinforce policies that reduce pollution locally and then it would forget policies that didn't locally. Now the question is, let's just say you start decreasing it and all of a sudden it starts creeping back up, your reinforcement learner is going to be quite confused. So in active inference, we can say that our metric is actually our uncertainty around our generative model of the river's pollution. So then what actions can we take if our uncertainty starts to get wider? How can we increase our precision on our generative model of pollution and then of course, concordantly with our values that it be unpolluted, how can we act to reduce our uncertainty? And so it's the reinforcement, economic maximization mindset. Pick the metric. I want everyone to read. That's the metric. Reading. Well no, the preference is that everyone can read but the metric in active inference is our uncertainty around our observations and it's an extremely subtle at first but eventually very far reaching way to think about these systems and what fitness is because a lot of times fitness metrics people are going to be thinking right like GDP or right like something that we want to see but we can actually now tease apart our preferences from our affordances from our uncertainty in our generative model and several other features of active inference scale-free models and that is going to help us have a multi-stakeholder conversation in a new way. That's great. Great, great, great, great. I just add to that. I totally agree with what Daniel just mentioned there because it's not almost, it's not an active inference per se on its own but it's the ability of active inference to be used in this epistemic foraging way and I think the point that was made there about reinforcement learning which we kind of think, oh that's some machine learning method but actually reinforcement learning is the dominant approach to development. It's like it's the way humans have structured in a way their own thinking. It's not exactly reinforcement learning is such a bit but it's so constrained and it's so goal-driven and the only integrating way to overcome it and why that should then start saying there is no alternative is the market saying it can't deal with the complexity beyond a certain point so you have to offload it to this market which again, you get into that kind of just becomes whatever's the favorite of the day and this gives us a way to get into this nuance complexity gives us at least a chance to go somewhere other routes and plausibly deliver on that promise. I think that's really cool. Right, absolutely. To give a spatial metaphor on reinforcement learning versus active inference it's like if there's a mountain peak that we know that we want to get to the reinforcement learner would say okay the goal and the reward is the elevation and so it can get trapped on a local elevation peak but then it might not take a step backwards or down a few to get to the highest peak but again if we do a trajectory of policy inference on getting to the top of a mountain again of course as a preference that we have for ourselves in the future we can ask am I reducing my uncertainty about how I'm going to get there and the way that we're going to do it together with potentially a group of hikers who have different capacities instead of how are we going to get to the top of a little null and then just stake out so it's a different conversation style Steven and then anyone else Yeah and sort of building on that is when you have that come into play is the ability for the active inference approaches like you're talking about magnified greatly when you don't just focus on the one person or the two people who are like the agents of change but you say well what about the local people who have no idea about the goal they have no ability to elevate they have no ability to contribute to the reinforcement learning but they know the slopes they can tell me the history they can tell me where people fell to their death in the past or where there's you know all these and that then makes the place place based community based contributions much more plausible as why they enable the scientists and other people to do exactly what you're talking about Yeah Excellent just to point out so we're talking about reducing expected uncertainty about the future right think of all the ways that think of all the implications of that in say you know design of systems or design of institutions or actions of humans facing some problem so one question that immediately arises is do we have enough information are we collecting the right information at the right frequency to even understand what's going on to reduce our uncertainty second is how good are we anticipating the future can we do that with some accuracy do we know what's going to happen next do we have any idea what's going to happen next how good are our how good are our world view concepts that lead to anticipation of the future and how good are our computational models about what's going to happen next how good is our information sharing capacities how are we communicating with each other are we communicating the information between groups and between individuals such that we can understand what's going to happen next and understand what our uncertainty levels are you know I can go on but there's like it seems like a simple proposition at first and then you realize that proposition has enormous depth it implies all kinds of things about how a society might function and organize itself one thought on the wisdom of crowds and distributed computation there's a famous example from early in quantitative genetics where it was at a world's fair and a crowd was asked you know you put in a raffle ticket you can predict how much some cow weighs and then the average of the estimates was extremely accurate although many individuals were of course distantly off the mark and a lot of people take that just straightforwardly as existence of the wisdom of crowds but there's a few pieces of that scenario that really matter the first is that they may have experienced they may have been working in agricultural settings or have experienced with getting feedback on their predictions and secondly the estimate was a bit bounded it wasn't that the cow was one pound or 10,000 pounds there was an estimator and then it was still a cognitive diversity that contributed to that estimate so you couldn't just look at somebody who overestimates that what would get rid of that guy is wrong it's like that's counterbalanced by someone on the other side of that curve that's how we're getting an estimate so having a distributional in a community perspective on decision making helps us recognize and encourage and foster diversity of opinion and also structure a resilient community so it's something that will be really interesting to see how and where these kinds of memes and ideas pop up yeah absolutely valuing differences of opinion valuing differences of perspective was one thing that you had said there I just add something to that is in our polarized society today we there's liberals and there's conservatives at least in the US and they seem to be in two different universes almost right but liberal and conservative at their best is a little bit like exploration versus exploitation the necessary balance for any organism or for any decision making process to balance exploration of the unknown and exploitation of what is already known what is traditional what is our history what do we know works at their best liberal and conservatism are both important just as exploration and exploitation are both important and problem solving and they're relative just like left and right when you're spinning around and things are changing it's hard to say and that's why we can pull back to some real pillars of biological systems instead of trying to scaffold further and further into the political or logistical ether we can actually return to our grounding as a biological society so Lee I saw your hand raised or maybe I didn't Stephen and then anyone else who raised their hands that was quite interesting we made that point there about we looked to explore the unknown and I think that in addition to that because that's from the perspective of say the scientist or the person in the kind of the knowledge generating game that's their profession or whatever but there's also the exploring the known but not known that it's known the tacit the so this is where communities can come in they by realising that some of that unknown is participants in those environments know it just needs to be surfaced this is where some of the work with the environment side of the active inference the niche and these topographies that I'm quite interested in my side can come in because now they're exploring and revealing stuff and in a way which is networked and sort of not just isolated as single stories or experiences and fragmented so I think that connects with what you're saying can you give one example of what you're thinking so one example we did a project around water issues in rural areas of northern in South Africa and so they had different areas of water catchment so some people were getting water from a lake but the lake had hippos and crocodiles so it was a bit challenging for them in different ways and so you've got these different water catchment areas and it's like they had the lived experience of having to go there having to get the water having to boil it because if the water is contaminated you've got to boil it and what it means then if you're then doing that if you've got someone at home who's got HIV or that so then you know so they can't walk or they've got to have so all these questions so what they would do is they would create a fabric map the size of the room about all their different routes to water and then share the stories from eight different or whatever numbers of hotspots and then that became the sort of landscape so in doing so that that wasn't given the answers but it was informing another process maybe a week later which was looking for innovations looking for options for new approaches so that's one type of way that you could reveal something from the local landscape which would never have been available to an engineer from their own knowledge yeah great stuff great stuff so some of what you're talking about is what is our individual narrative and what is our shared narrative like what is our experience of life and what do I as an individual have what knowledge and lived experience can I offer to this group that would help the group to eventually function better or make better decisions right so that's a really interesting question to me for many reasons but one reason is how how suppose we were to move forward this whole project was moving forward how do we communicate with each other richly I would argue that there is no system no functional system right now that a million people or 10,000 people can richly communicate their lived experience their understanding of life their understanding of how what might happen if A, B and C happen I don't think we have we're in need of tools to richly communicate we're in need of tools so that's an immediate opportunity for groups to get involved with just that little question of how would we communicate and how would we make decisions together in some kind of collaborative fashion how do 10,000 people do that on a certain million people that's the question when you can't all fit in a room to really thoroughly discuss these issues and listen to each other how do we do that at scale so that is and I think there could be answers to that that's not an impossible task that's a really interesting and doable task if we just put our minds to it it's related to regimes of attention and active inference perspectives on narratives which were even our first act in live stream number one was on narrative active inference because we think it's crucial and if there's going to be a social narrative like you had brought up well that's something that is not just sharded into a person's brain it's not a holograph where it's the same in the person's brain as it is at the social level things that brains know that neurons don't it's not like each neuron is holding it is amongst neurons or if you think about like Star Wars or a movie it's like each character doesn't know the whole omnescent scope and so we're part of a bigger story and so it's going to be interesting to explore what will it feel like as a person and what will it be like experientially with a personal narrative to be part of a broader societal narrative but those will not be the same thing and the desire for those to be the same thing will lead to system fragility in many cases so Steven and then any other comments yeah I think that that ability to know certain things is a very important point and I think in active inference something that I've increasingly been looking at is we've got external states and we've got hidden states in different ways from the perspective of the generative model of say a water engineer there will be more there's hidden states from everything from everyone to some extent as well as states that we can engage in and so there will be this question about how do you engage the hidden states in a landscape as well as the external states of the regime of attention and where can someone else who's got a different lived experience contribute to opening up and maybe telling us where they think the meaningful aspects of their landscape are so that I think that as well as modeling and happening from our own computational models and being able to maybe use some tools we can model through participation in place based based on that absolutely absolutely so this tool I have in mind it's I want to have my voice heard I will I have a story that is that could be important to others that could be important to any particular situation that or topic that society is trying to decide upon and I would very much like my voice to be heard and not just a yes or no or not just a choose a B or C way I would I have a rich experience that I want to communicate now you know I another person sitting next to me might have a model of water quality or whatever the issue is and they might have something to say about what their model implies but all of those voices right should be heard all of those voices should be part of the cognitive process of the society so certainly data driven and you know model formal models of this and that sure that's great that's fantastic it's an extension of our cognitive capacity and at the same time I'm intelligent and I'm you know I can do things a computer can't at least today you know I can do many things a computer can't so whatever process whatever communication system and process that develops out of this ought to be one where every voice can be heard can be registered can be incorporated into this larger really cognitive process you're talking about a cop you know it's like sharing and we're discussing but that's a cognitive process right for a group that because it's leading towards some better understanding of the world and you know improved decision making about the world John if I could you mention data driven so to again sort of distinguish and contrast with how you're potentially using data driven versus how it might be implicitly used in other settings data driven who's being driven us the data are driving us in our policy oh the policy is going to be data driven okay well how about it's human driven how about the policy and our affordances and the data are like a car and we're driving the car do we need a better car does it have too many wheels where are we driving who whose turn is it to drive how are we choosing which music is being played in our human driven car not how are we going to fit into a data driven human mode and it's just interesting that sometimes what is left unsaid or unstated is where a lot of those values get swept under the rug it's data driven what else do you want to talk about oh values okay well there's time for values it's like there has to be time for values but the data driven conversation makes it sound like it's the beginning and the end right right today yes yes and I think that can be expanded greatly right so that's that's when I use the word richness that's a little bit of what I'm trying to imply is the richness of experience the richness of perspectives the richness of past information that cannot that I as an individual can bring to bear on this issue whatever the issue is and then on top of that what what does the data what data do we have how frequently do we collect it how what is its quality how can it contribute to this discussion right so very much this it's not that the computer has to decide what to do it's that human beings have to decide you know what to do how to given all this information what do we do how do we frame this how do we frame our problem how do we are we even looking at the right problem are we is our focus too narrow are we you know or do we need to broaden our focus to understand better how all of these different problems so very much human in the human in the center computation and to connect that to first and higher order science and I'd be curious what you think the first order project is like build a bridge the second order project is like given what needs to happen is a bridge the best solution or maybe we can go about it a different way and the third order is what really needs to be done here what do we value who are we those types of third order questions which potentially even are on an escape trajectory from science because we're not going to be looking to science in a narrow sense for the answer to that third order truly human question but we do need to expand this discussion from what is the best way to build a bridge to how are we going to go about fulfilling transporting people from side to side or making a landmark for our city there's different reasons why we might want to build a bridge and people feel like a cog in the machine if they're just being instructed at the first order like do this right but when we're including people in higher orders then we're part of a multi-scale cognitive system we're not just a quote cog in a machine so Lee then anyone else yeah I think this is a really important point and for me it touches on you know Metzinger's transparency opacity kind of are we do we understand the beliefs and the frames consciously that we're operating out of so in the bridge metaphor you might say at the third level the third order oh it's really about how are we looking what's the ontology or the set of metaphors that we're even coming at this system through and what effect are they having in shaping well A how we perceive it and B then our actions towards it so going right back to the beginning of this discussion it was quite interesting when Steven asked you John about architecture I'm kind of thinking is that is that a metaphor the architecture of the brain is that a way of looking at the brain or is it something this parallels the discussions about Markov Blankets etc are these things that we're imposing on systems or are they things that are really there I think the architecture one is really interesting because it kind of comes on this whole constructivist tradition and I wonder more whether it's not so much whether it's about the right accurate way of seeing but what particular set of affordances does a model or a way of seeing in the face of this particular problem and this particular niche perhaps and yeah interesting stuff thank you Lee John continue or Steven and then John feel free to continue you know just one last piece I think that this there's some areas that I'm trying to apply these isms now to sort of put some buckets I think the social constructivism is is an important one to think about maybe science in a way kind of fits in that in a way as a human pursuit pragmatism I think is very well aligned with the kind of approach of active influence and this approach here if you take pragmatism from a you know a human not humanistic perspective from a well from the third one which is an ecological perspective so or a ways of knowing as well even perspective so you can drop down into ways of knowing perspectives you know pragmatism social constructivism and ecological approaches and I think that nexus is is quite important to as a sort of way to hold some of these ideas this is you know this is a fantastic conversation and and I would say that the exactly the kind of conversation that would be needed to to move such a project forward it would be conversations like this between scientists within the communities between communities and touching around all these some superficial and some very deep and subtle issues and questions how do we what are we thinking how do we do this how do we measure success what do we want what is our purpose I mean you know we're just maybe starting that conversation here amongst us but I would like to offer to the science world that this could be we know this could be a function in the science world of you know a pursuit of study a pursuit of of attention and if we're I think that the you know the I think that's all that's really necessary actually is to put our attention to this and intelligent people communicating amongst themselves and engaging the all interested stakeholders could definitely make progress in this in this endeavor so great great to watch great to be part of and so yeah John maybe if you want to go to another slide but that's leads the question how do we in our dot one video it's kind of an opening the dot one is where we've had our conversation around with a dot zero's we've hopefully read the papers and digested them written comments written thoughts to our self allowed that to reverberate and then we open the doors and we come together in a group discussion for point one and we also say hey the doors aren't closed more people can join this conversation whether it's next week for point two or on a broader scale so I'm just wondering whether you answer now or later who can be included and how can they be included in participating right well and it's a a few words everyone is everyone is welcome in this conversation in the in the world and communities in different in different different types of communities the science community other communities I mean this is this is a conversation for all of us to be had what what do we want what is our purpose what are we doing how can we improve how can we reduce uncertainty you know about essential variables you know to put it more technically how to do that I don't know I'm open to suggestions of how we make the next step but you know I sort of put out put out what I could put out in this series and I'm hoping that these conversations might lead to to some actions to engage a larger and larger sphere of individuals and organizations so I'm open if we if there's suggestions I'm open and could you flip to see that would be slide to talk to slide eight yet go for it so we touched on all these ideas and I thought maybe this might be a good moment to talk about some of the practical implications when in our original discussions this slide was coming after the after paper one which was all about world view and seeing life as a cognitive process and seeing individuals as overlapping processes of information flow and understanding that when we talk about a community improving its its quality of life and quality of environment we're talking about an extended individual and an extended community that really reaches all the way to the biosphere ultimately right we're all connected with other individuals and all communities are connected with other communities and all of life is connected we're connected with our local ecosystem and with this lead ecosystem so when I when I talk about vitality I'm talking about the vitality of the whole really that's our purpose is we're intelligent we're humans are very intelligent we can think a thousand years into the future of what might happen if we do something today and humans can use that capacity to improve and reduce the uncertainty of the whole world improve quality of life improve the quality of the environment and biodiversity and all that kind of stuff so with that kind of theory theoretical aspects in the past what are some of the practical implications like how would we use any of that information to actually think about design for societal systems right so so this is a list of something some items that we have talked about in previous talks and even touched on today and when we think about what does this mean practically I would like to focus that to a different question really and that is how do we how does any individual any organism solve a problem like what is necessary to solve a problem whether you're a bacteria or you're a human community right what do you do and on the left are I would say some of the universal aspects of cognition for example you have to gather data I mean the intelligent being is processing information and information means that some kind of data or information has to come from the world and then it can be processed it has to remember for example it has to anticipate the future forecast the future with some accuracy in order to survive if I see you know if I see a lion in my path I might want to accurately forecast that if I don't get out of the way I might be the lion's dinner right I mean anticipation is central to cognition so we I've only talked about the first three in the list there but already let's we can talk about what does that mean for a community then so how does a community gather data for example and that might be you know through sensors through through getting the input of others in the community about some particular issue requires like remembering requires that that information be kept in some kind of accessible form some kind of repository of information we can think already of ways that society does these now we for example the Census Bureau collects information annually and every 10 years collects various kinds of information there's financial surveys done every few years some of that information is put in repositories that at least researchers can access for different reasons but how you know how does that differ how do how does what we do today how might that differ from if we were really thinking of communities as cognitive organisms how might we improve on any of these processes right I don't know if there's thoughts on that but I can yeah Stephen guys yeah I mean one thing I think that's interesting is this actually points as well though to an interesting dilemma I don't know if it's a dilemma but certainly a challenge in the sense that these artifacts and these artifacts of knowledge that we have they're kind of we have the option in a community setting to place them in our environment because they aren't they can be equilibrium based artifacts so for instance if I create a book and I put it on a shelf it could be that 100 years right and I come back and it's still a book right so you've got this ability for that kind of storage now when I remember something I with active influence I'm establishing this kind of priors an update of priors which is generated on these lower level sensory flux pieces of sensorium which I have at higher levels potentially started to interpret as what you might call a signal but we get into this interesting question because it's like the error is kind of the signal in active inference although it's possibly at higher levels of thinking it gets a little bit different because we've been talking about that with some of the modeling but there's an interesting question around as you scale out where some of this sense data in active inference is kind of held in the kind of collective knowing of a community which knows how to interact for instance with the artifacts yet you've also got aspects a bit like I've got my clothes and I've got other things that kind of they part of me and they not part of me they part of the environment there's I think there's something quite interesting and I think that's an area I'm really interested in exploring is teasing that out so I wonder what your thoughts are on that well part of I think part of what you're speaking about might be what I've been calling richness of information right so obviously a census can collect information once a year about you know numbers and education levels and things like that that's one stream of information a valid stream of information and yet another is very personal it's what is your gut you know what is your gut telling you about this situation or this you know this constellation of what's happening today what is your gut feeling of what's happening right so how does a community process sense and process all of that information together if you have you know if we're a part of the community and we're trying to decide on some issue it would be good if we heard your gut what is your gut telling you about this is this good or bad or is this direction good or bad or even if there's no like data to back it up what is your sense based on your complexity your wholeness like we can't even put a word on it what you are where that information comes from maybe but what is the sense that you're having and I would like to share my sense too and everyone maybe could share their sense so how do we do that in how do we do that when there's 100,000 of us how do we do that well to give just a few thoughts there's two sides to the conversation they're sharing communicating and then there's really engaging and listening which is part of the topic so 100,000 people speaking over each other is zero conversation that's not conversation right exactly so how do we do that and I don't think that's not it's almost surprising in a way that we have such enormous technical capacity in today's society to do unbelievable things and yet this very basic question of how do we as a community of say 10,000 or 100,000 how do we richly share information in a way that all of us feels heard that we can express our kind of like hard knowledge and gut knowledge how do we participate together in describing our situation describing our fears describing our uncertainty anticipating what is going to happen next if we take choice A, B or C it's kind of interesting that tool doesn't exist in spite of our technological process and there might be reasons why that I mean there might be political reasons and economic reasons why that tool doesn't exist because if that tool did exist people would want to like naturally I want my story heard I want my community to understand what I know because maybe it's going to help somebody and I want to hear in a healthy conversation in a healthy exchange I really want to know what you're thinking and I want to contextualize all this information in light of what our purpose is what our goal is what are we after what do we want right? Can I comment on that because I think this is a really important point and also though as well as see this is what I find I'm recasting speech as an action policy as an extended cognition into a spatial landscape of semantics right so effectively it's actually words could be seen as in a virtual environment that we are projecting out and what that can kind of open up is that when I'm talking to someone and being heard right which basically means words are being heard right so but there's certain things that I can't communicate through words because they're not in that semantic landscape which is the compression of my a useful tool the thing with tools is generally speaking it's not very healthy to relate to tools right it's like tools are things so one non-verbal work can be really powerful we actually had that with our work with the community who use alternative communication you know we really push that so they it can take them three minutes to do to three words because if they're using a head switch they have to scan through each letter and hit the head switch and then so we're really about how about it not being about the words and it being the look of the eye and that building up of non-verbal knowing between people which at some point will probably be well we actually then articulate that through fabric landscapes so then they start painting the room and do world building and then you start making the words so there's something about like you said there may be something that tool well we need the tools but also sometimes tools just can't go there so what do we do where we don't use tools and then when do we bring it into a tool based environment I think this is a good point yeah yeah yeah and I think much of human communication actually happens non-verbally too if we can be in the same room and I'm watching your facial expressions I'm getting a lot more information about you and your intent than if I was only hearing your voice if I was only reading your words so sure there's non-verbal communication there's artistic expression there's poetic expression using words in a different way and obviously we again as an individual I want to be heard and if I'm maybe I'm expressing myself in words maybe I'm expressing myself in how I'm raising my eyebrows or maybe I don't even have a words to express myself but I have an image in my head that I want to communicate somehow somehow we have to be able to communicate richly in in mass and how do we do that and part of the answer to that might be we communicate in smaller groups and that information then is sort of propagated up the hierarchy or something maybe I mean you know the slate is wide open here so this is the question to the world how can we communicate richly in mass such that our purpose might be fulfilled our purpose of maintaining achieving and maintaining vitality into the future right so that's the big picture and then what are the what are the many tools or social norms or educational processes or you know any number of things that lead us to fulfilling our purpose our intrinsic purpose I should say our logical purpose in a sense nice John I hope that people clip that out or hear your question and really consider it and what it makes me think about on this slide and to return it to active inference again is really like the rhetoric of participation because in a reward maximization culture things like communicate it's like but not now we have decisions to make whereas when we reframe about reducing uncertainty on a multi scale so individual and share generative models it's like have we communicated enough to even know each other's generative model because once I know that your generative model includes this dietary restriction I don't need a hundred text messages a day I actually can generate recipes that are going to be compatible with you so it's not like we just need to have more communication or a different kind or these two people need to talk or some sort of secret trick it's actually that we can balance a lot of these sometimes locally contradictory needs like choose to you could find a situation where they align on this list and you could find a situation or imagine a situation where that I've urged and the power comes from being able to have that goal directness and the epistemic gain and then potentially through peer mentorship or through extended cognitive architecture we could be able to mark you know where's their yellow flag where's their red flag where do we need to pull back and share more about one person's generative model or how are we working as part of an integrated unit that's reducing uncertainty given the preference vector rather than well if we each just make as much money as possible then the group will have as much money as possible right so this is a different way that we can go about having that conversation and as hopefully we're drawing out it's something where everyone's perspective isn't just a goop to be formed their perspective is their contribution and their value yep yeah good good we can also look to biology too how does the cells of the body make decisions you know like how does a liver know what to do and when to act how does the heartbeat know when to speed up or slow down there's rich communication that happens multi-scale some local some distant through hormones through messengers there's enormously rich communication that happens between tissues and cells in order for all the pieces to understand what the you know what is necessary now right so there's we can this is an open question it's a clean slate there's great potential for the science world and other communities to move forward of this question of how do we richly communicate and make decisions together how do we do that and and there's lots of biological examples for us to look at to gain inspiration from so a question from the chat just before Steven from Yvonne just about the tools how many people have used the kinds of tools that you're suggesting or sounded like you suggested that it didn't exist but what is the preferred field for a startup from Yvonne where does this tools where is this coming out of this sort of tool for communication and collaboration I'm not sure I understand the question exactly where's it coming from he asked what is the preferred field to make a startup in for this kind of a tool like would it be in the field of industrial engineering or well so I have ideas in my head of how we might you know start to build such tools and the reason I'm thinking about it is for this very this very problem of how to communities communicate how if this club model we were to move forward with studying this club model and eventually developing a field trial with the club model how does communication happen within that club model so that's my interest you know is the application to this problem of societal transformation and demonstrating new ways of of cognition at the local level via clubs but obviously if you had a tool that could do this there would be tremendous opportunities for others to use the tool too but this is just one example of thousands you might think of but I work in the utility industry and I've seen how messages get transferred from a caller calls in and says there's an accident there's a wire down, there's sparks flying there's children around there's some kind of rich situation happening that others need to know about and by the time that message gets to the person in the field or in the truck it might be quite garbled it might be not unclear at all what the problem is or what the dangers are or what the situation is so that would just be one example of how industry might use a tool for better communication and there would be many many more examples you could think of so my guess would be that if the science community moved forward with this problem of how do you richly communicate with a large group those tools and the concepts that would come out of that could be applicable to a wide variety of industries and to this particular question of how does the club model work and how do communities cognate Thank you John, Steven and then anyone else with raised hands Yeah thanks it sort of reminds me of some of the principles actually, systemic coaching can be an interesting area we're looking at coaching groups and the system, it seems to be that and I was just wondering if you've encountered the viable systems model Stafford Beer's viable systems model because I think that some of the active inference work can recast that because you've got not just levels of regulation on the body on the nervous system of a frog or different nervous systems but that doesn't necessarily have that bottom-up teleology that you get from active inference doesn't necessarily have the idea that the lower levels also know something it tends to be the higher levels of regulating the lower levels so that might be something to look at as one area that would be interesting and I just thought I'd mention that No no, great and those are the kinds of ideas that you would want to explore as you build the kind of tools we've been discussing and on that topic the goal is not to build one tool that does all of this we could have many tools different groups could explore a variety of approaches and tools and some depth of simplicity and complexity and there's room for many discussions on this topic one other, just to continue with you, the idea of coaching I've been around enough folks in my life to realize that some groups and some people are really skilled at communicating and talking with others and facilitating group conversation and things like that and many times I've been just so impressed with the ability of certain groups or certain individuals to facilitate communication and show by example what healthy communication looks like and at the same time I've watched TV I've seen TV shows of terrible examples of how to communicate examples of dishonesty examples of manipulation examples of holding back examples of confusion examples of shallow thinking social norms part of this whole process would also be like the club model would be to develop social norms that are useful for the community that are reinforced to healthy behaviors or healthy ways of communicating and healthy skill development that would help the community to better process the information it has so that kind of coaching is available too and this is something in the narrative as active inference we talked about when we see representations of communication they are our priors are updated to think that those types of communication are more likely or that they're prevalent or even that they exist at all we fulfill that prediction by reducing our uncertainty about our own behavior and our own thinking through other minds of other people's behavior thinking well what happens when there's two people and they need to and then you're not surprised when they do it you're not surprised when you do it but there's also a way to show representations of that not happening so that it will not surprise either person when they both slow down and that shows how our intersubjective level of computation is something that is influenced by both the bottom up the sensory input and the top down with our cultural scaffolding and so systems in change need to be able to resonate from the small to the large or there's going to be some pathologies so Stephen any other comments say this actually could be an interesting or could inform anyway the types of questions that could pose to Carl Friston himself around some of these practical implications so we could have these in mind so I was just thinking do you have sort of I wonder if you actually have thoughts about are there holes in this or are there questions around where this whole field of active inference is sort of going in the sort of global sense and whether you've thought oh I wonder what what Carl Friston would think about that just a question to put out there Carl if you're listening feel free to to join in no I would love to one of the many ways to start this whole process off would be to have a conference where we get together some folks who have experience in these fields in a variety of fields who could offer their thoughts and cautions and encouragements whatever they may be well John on June 21st 2021 Carl Friston will be joining us for a little symposium and in another sense we have a nano symposium every week but any type of logistics or organization that are the enabling architecture is what active lab is there for great fantastic fantastic fantastic you know I maybe should say too we've been talking about active inference and obviously all of us are interested in that topic but you know it could be that the active inference 2.0 is called something else it goes you know like we learn something in the next decade or whatever and active inference is no longer you know like we moved beyond that and now we're talking about some other thing you know the important aspect of this for us is how does cognition happen how does you know how do biological organisms structure themselves and orient themselves and organize themselves in order to process information and you know how does that work so whether it's active inference or whether it's you know something else sure everything evolves we learn as we go so it might be you know next year might be a different word but the concept is how do we achieve and maintain vitality in the future you know that's how do we fulfill our purpose and even today for different non-English human languages and for computer languages the word is already not active inference and so that's why it's always so rewarding and valuable to be very clear on what active inference is which parts are drawing from Bayesian statistics from complex systems from predictive processing from all of these other areas instead of thinking that we need to reinvent the field because we certainly don't we're also just a part of the cognitive architecture and it'll be a quite interesting ride for sure so maybe in the last 15 minutes of this point one John help us via another slide or by raising another salient question what can we as listeners and participants think about in this next week to kind of simmer during this week between our two sessions okay maybe next week we'll talk more about we've only covered two slides here and I think like 60 slides we could have looked at we covered two of them I remember maybe three of them but the the goal here was to talk about part three of the paper which is design so I think maybe for our listeners the question of like how would you actually do this what would in part three I spend some time talking about a prototype called the lead when I call the leader framework local economic direct democracy association framework the emphasis there it is the concept is a full integrated systems approach but the emphasis in the simulation trial that I published is on economics so just take that alone what would a fundamentally different economic system look like how would it function and how could it apply to a club model you know like implemented at the local level what might that look like how and when I say economic I mean economic slash financial slash monetary institutional you know like all of those things wrapped up into the word economic what how would that look what are the possibilities great great question one thought is econ eco and ecosystem and economy it's like they start with the first three letters they're so close they're almost on the right track with an ecological perspective on economics yet you add one more letter and it changes from ecology to economics and all of a sudden it's like never the twain shall meet so it's an awesome question to be rethinking how these different intertwined systems relate just like intertwined systems in an ecosystem of like primary producers preanimals predators renewal etc so on that you know just to follow that up just a little more what is an economic system you know before we go about designing new ones what is the purpose of an economic system we ought to ask that question first before we jump into potential designs right so what is what is it supposed to achieve in the first place and the I think for some listeners or for you know the general population at least if you ask that question the answer would be like the economic system is about you know allocating resources back and forth you know the choosing who gets what and and how resources are used but I would argue that that's not at all what an economic system you know like intrinsically potentially what an economic system is I would say an economic system is exactly what we've been talking about so far today it is a cognitive system it is a different variety of cognitive system that allows a community to cognate in a slightly different way like a governance system allows its community to cognate in one way a legal system allows a community to cognate in a different way and an economic system allows it to cognate in a different way So if the goal is functional societal cognition, what does an economic system look like? What one thought on that and using the body's intertwined signaling systems like neurological, chemical, etc. From the point of view of the ribosome, the protein translating machinery, insulin is just another peptide. It's just linking the amino acids together and it's making a protein. To the point of view of a receptor, the insulin molecule is an affordance for binding. And then to the point of view of the downstream signaling cascades inside of the cell, we start to see semantics enter into the picture and then there's a turnover for the insulin and there's lactate being shuttled from the muscles which are using it and they're calling out for help through these different mechanisms. So it's like a cargo ship might be carrying something and it's putting it in a box and it's just moving it like an encrypted message. It's just moving the information. It doesn't need the semantics to be able to relay the information. And similarly, when there's these intertwining systems, it's like that one insulin peptide can be handled by a producer, by a transporter, by a signal detector, by a degrading enzyme. And the idea that all of those pieces are, you know, you could just change one out or you could remove one. It wouldn't be the same system as a whole. So, Stephen, any other thoughts? I like the hormone reference. As I think that the nature of the hormone system where the hormone gets released into the blood, let's say the blood, and it travels around the body, it is actually maybe closer to this idea of a message passing kind of signal going out there, you know, sort of sense in sort of a general way, and maybe this is something that the economic system does, right, is, and you can't do that so easily through necessarily just dynamical processes with people, right? It's like, OK, you can, you can kind of infer things, you can sort of build things up, but OK, you set up a pipeline of value distribution, or at least the information in relation to value distribution and how that can be processed, and you can release your hormones, which is basically money or budgetary predictions into that system. And it flows, right? It flows around it and it gets picked up and it becomes more like a traditional signal that can be computed. But you can't put out the dynamical system information because it it's not a thing. It's dynam, it is dynamical. So there's something interesting there, I think, around that that analogy to that. It's the system it's the system we're working with to be able to explore how message passing algorithms instrumentally used by us as curious and motivated investigators allow us to act in the world to reduce our uncertainty about blood sugar. So following on that, the question to for the public is what is money? You know, like in its best sense, what is money? So we'll try to we'll try to tackle some of that next week. I'm going to meet again. Yep. So just on a closing note, we heard two big questions right there at the end. How can we actually do it? And it would be awesome if people could read, especially the third paper in the series and look at the Leda framework that John has that Leda that John has proposed. And there's a simulation there and it's a scaffold for a lot of rich discussion. So on the how can we actually do this? We want to hear many perspectives as well as people's thoughts on the third section of the paper. And then the second question, what is money? Which is, of course, a fascinating question for history, anthropology, game theory, so many different areas converge on what is money and it's a system in flux as we see quite literally day to day, depending on how closely you follow crypto. Right. Thank you, everybody, for joining. This has been fantastic. Awesome discussion. So everybody who's watching live, thanks so much for participating. And next week in number 21.2, we'll have a follow-up discussion with you, John. So thanks again for your engagement and for really giving us this foundation with the multiple precursor videos because it really helped. And I think this is going to be a great conversation sequence. So peace out, everybody. See you later. Thanks. Bye. Thanks. Bye-bye.