00:00Alright everyone, welcome to the Stoa. Wisdom of Iris: Collaborative Sensemaking with AI, featuring the one and only John Ash. And if you recall, John was here last time with Daniel Schmachtenberger. Daniel Schmachtenberger was highlighting some good work, some good builders of sensemaking tools, and speaker John Ash was one of them.
00:45He's an artist researcher interested in how you, and we, can use AI to support our collaborative sensemaking. He's been playing with something that he calls Iris, sees that as a tool to interface with collective journaling, reflection, prediction tracking, and knowledge pooling. And I'm going to bring in John in a moment, and we're going to discuss what this is and the potential of this thing. I've been playing around with this, we've been playing around with this at the Stoa with certain collective journaling sessions, and it's getting pretty trippy. So, like, buckle up. Things are going to get fun. And how today's going to work, I actually don't know how that is going to work. So, I'm going to bring in this mysterious character called dpax, and then they're going to introduce John, and then I believe there's a PowerPoint presentation that will be fielding questions throughout. But if you have questions anytime, drop them in the chat.
01:41And if there's a Q&A portion, I'll call on you, you can ask your question. And this will be on YouTube, so if you don't want to be on YouTube, just indicate that and we'll read your question on your behalf. So, that being said, dpax, I'm bringing you in. Thank you. So you can call me dpax, I'll be helping moderate today's session. So, the Wisdom of Iris: Collaborative Sensemaking with AI. Thank you so much everyone for joining us today. Thank you, Peter and the Stoa, for hosting a special session. We are so excited to unveil the latest developments of Cognicism. So, many of you have already joined the Discord for early access and you've gotten a chance to try out some of the new features. If you haven't, you'll see in the chat there is a link, so you can go ahead and follow that. So most of you know John Ash, but for those of you who don't, John is a thought leader in AI assistive technologies and decentralized sensemaking. He's also a writer, a visual and musical artist, and a steward for the ideas around Cognicism.
02:38So in the next hour we'll spend a few minutes grounding everyone in what Cognicism is, so we're all familiar with where Iris originated, and then we'll have a couple of quick activities to let the group try out a consensus finding feature. And after that, we'll explore how AI can support introspection, encourage deep listening, resolve conflict to build stronger relationships, help us make sense of the world, support collaboration, and align our actions to emerging shared goals. We'll share some ideas and even form predictions from the wisdom of voices flowing into Iris, and then after the hour is up, we might be having some live music. Is that right, John? No, maybe. Okay, you're on mute.
03:23We're gonna go with the flow, we're going to say, okay, alright, we'll see. Yes, that's your role, to keep me from going. We'll go to an hour and then we'll see what happens after that. Do you want to go ahead and share your screen for the presentation? Should be shared. Is it? Can Cameron see it? Alright. Can everyone see it? Nope. So John is doing that. I see we have a lot of people in the chat. Definitely, if you have questions throughout the time, you can put them in there. We'll try to answer as many as possible, but let us know if there's something that you want to ask that we're not answering. We'll also have a Q&A at the end. So with that, you can go ahead and take it.
04:13Hi, I'm John Ash. I'm very grateful to have the space to speak to you today. I'm here to help. I realized after the thing with Daniel went so well, where he noticed when I was meandering and did things like, wait, you have to tell them, do they know what generative text models are, do they know what GPT-3 is. So, does everybody here know what generative text models are? I can see all your faces, there's nods. Yes. What we're going to do today is that I want to go top to bottom and explain the ideas behind Cognicism, but I also just want to show off some of the exciting, like, ways that you can play with these collective intelligence tools and sort of make your brain think about the ways that we could do things. Everything that I've done so far with these generative models, as brilliant as they are, it's kind of like a cosplay. Like, there is a clear definition that I have of what a proper Iris is.
05:30There are barriers to having a generative model with all of those features, if you're not a large company like OpenAI, or, you know, if it's not something that's generic that somebody's putting out in the world. So there's lots of generative models out there, and they will say things. They will say really interesting things, they will say things that connect you, meaningful things, things that make you feel alive. And things that convince people that they're sentient has already happened. They will output things, but right now they're not always going to say the truth. They're going to try to say things that make sense based on what you provided. The thing that I've prepared for today is a live summarization of different perceptions. Basically, we think of democracy as this process where we ask somebody a question, and it's, you know, the tally of hands. Like, I ask the whole group, did you, raise your hand, that's yes or no. That's kind of limited. I can get a number out of that and I can get a sense if I need to talk some more.
06:44But the idea that we're going to play with is, what if you could just say, and it could go into this great, meaningful, and then you would have this generative model that would try to form something from the center of all that, and express the will of the group of people, or the thoughts of the group of people collaborating. And we're going to do it at small scale, we're just going to do it at, you know, 20 or 30 people, however many on the Discord server. But I want you to imagine this at just a massive scale, where it's like, instead of asking, okay, should we add a playground to the park, how do you think the park should be changed, what do you expect, and you have the tools to interface with that directly, without some force trying to shape the narrative of how the summing goes together.
07:35So that's what I want to do today. These generative tools can do a lot of things. I want you to think about all the different ways that they can be positive. There's a lot of things out there right now that are very negative about generative models, but AI, because generative art is making people feel like they're losing their job, their ability. Most people did not expect that art would be one of the first things that AI mastered, and there's this sort of sudden realization that, oh, maybe we got things wrong, now we have to react to it. Who controls these tools, what are these tools going to do, how is it going to reshape everything, and you realize, oh, who has them. And so, the vast majority of people in the world are probably going to think from that frame of, this corporation has this, where does it go.
08:27And they're probably going to make some pretty good predictions about it going in a negative direction. It's really great, and it can be so beautiful. We did multiple experiments with the Stoa so far. One, which was on beauty, was just magnificent, to the point of, it said a piece of wisdom that just so blew me away, I fed it directly into DALL-E. Maybe dpax can put this into the chat, maybe the link to the tree or something. I fed it directly into DALL-E, was immediately visualizing this concept that was formed from all of these statements. It would have never existed if it wasn't for all of these abstract, beautiful expressions of poetry from these members, from the members of the Stoa flowing in. And then, you know, Peter asked me to say, what does Debra need to hear, and the first thing that Iris said was like, I can't remember off the top of my head, was just so out of left field but beautiful and dynamic. Was about, like, the evolving nature of crystal, and then put that right into the image prompter.
09:45It called for something from the in-between. That's what the latent space is, the space from in-between. Now I want to translate that to something human, and dpax is here to help me stay on that and not get too far into the technical. These things are all true, that it supports introspection, it encourages deep listening, resolving conflict to form, you know, stronger relationships, helps us make sense of the world, supports collaboration, aligns to emergently shared goals, can help generate new ideas and invent new words. Like, there are going to be concepts in this talk that my Iris invented to help me explain things to you that are just more intuitive than the words that I was using.
10:37And the most exciting and interesting thing that I want to try is the generative Oracle. That we are all sharing our predictions of the world, and the Iris is coming to some shared representation and explaining the future. There are prediction markets out there, there's idea markets out there, there are claim markets out there. There's so many great ideas that are trying to connect our perception now to the long-term future. And, you know, it all ranges, the future. And I'm really focusing on the fluid level of that intelligence, where it's just about the human capacity to express about the future.
11:28So Cognicism is a word that far preceded Iris. Iris was Daniel Schmachtenberger. He said, dude, this is really cool, these are great ideas, you need to focus on something. I started talking, he said, I said, lens. And I mentioned the name, that thing was called Iris. He says, focus on the lens thing, focus on that. These generative models, when you take a GPT-3 and you fine-tune it, say, output things of a particular worldview, that's what you're doing, you are training an Iris based off of your views. And it's able to say these ideas in a way that is clear and more comprehensible than you're able to explain them, because it has the context of, you know, the way that humanity communicates. And I only have this tiny context of my worldview that has built up this idea of Cognicism, and it can place Cognicism, this incredibly complex thing, in a context that will make far more sense, and has made far more sense.
12:35But it's not just AI. And that was the last piece, I did not want to add AI to this. I was more thinking about predictions and what does it mean when people tell us what is going to happen and we don't listen to them. And I wanted to track that. And I wanted, like a lot of people in the Web3 space, I wanted a clear score, a token, something just, like, a scalar. Some beautiful thing that was, like, this is the best predictor. Or you put these box of predictions into this function, and it outputs the best, the most wizard man. But it's impossible.
13:22All of these singular token-based systems, there's so many flaws in them. It needs to have this evolvable, learnable thing, and that's where I ended up at. It's designed from a completely different angle, just trying to understand and make sense of the world and understand which sources were most sensical. And I was tracking different signals around that, and I was building these tools. You know, this manifesto about Cognicism, to do these things before things like the meaning crisis existed as a name, before the concept of memetic mediator existed, before the concept of Game B.
14:11Just before all of this, I was sense making, I didn't have those words as existing in a different part of latent space, and I was unable to contextualize it to anybody. But I saw the frickin future, and I knew what was going on, and I communicated to people, but they were so cruel to me, because it was different than what they expected. And now we're here, and we can just show you how Cognicism will do these things. It starts with journaling, really, I mean, that's the lens that I think about it, like tracking what you think over time.
14:56That's what I did for years before AI. I just had this journal that was this sequence of claims, predictions, statements, and thoughts about the past, and how sure I was about them. And I just said my thoughts about the world through that for years, and just scanning back, finding ways to look back at a lot of it at the same time, and made charts and graphs, and I broke it down, and I turned it into all sorts of models, and I just analyzed it, and I tried to make sense of it. And that was incredibly powerful, but it was always seeking, how do I merge these voices into something that is useful? I wanted to synthesize it into a signal.
15:42I was not thinking about this like generative models. I just wanted a signal that said, hey, honestly, I'm an ego, I'm an individual person. I wanted a signal that said, hey, this guy is pretty good at predicting in this context, and here's the damn proof. I can spell it out for you. Here's the whole line of receipts. Here's why, and we're gonna imply this voice from now on in that specific context, not other contexts, but specific contexts that I have had a accurate prediction about. That was what was motivating me. And the process of iterating over my beliefs just brought me to the place of having to give that up, that seeking of a thing up for something better, that answered the question better than I could imagine.
16:34And when I had that answer, it became a very difficult thing to communicate. But the idea was that I took this great stream of claims that I collected from so many people about their perceptions of the truth, all of these intersubjective collections of what they thought was gonna happen, what they thought had happened, what they thought is happening. What if I fed that into a generative model? And then what if I took the output of that and had people evaluate that instead? Because there was so much bias being added from the source where people wouldn't listen. They're just like, I know who said this, and I'm interpreting what you're saying through this larger frame.
17:24I'm not even interpreting what you're saying. And suddenly I had this thing that was like, hey, actually, I can evaluate any information completely anonymously in a way that is not affected by the context of what you expect. And it's like this private space to learn which sources are the most helpful based off of real, actual, live feedback from human beings. And that idea, oh, fun. I added this slide today. What I wanted to show you was the idea of synthesis because, let me see if I can do this.
18:15Is this right? Can you see an open AI screen? Yeah. Yep. So just like right out of the box, an untrained open AI model can do the idea of, okay, here's a point, here's a counterpoint, here's the synthesis between the point. That is really what is so beautiful about these models is it can interplay into that space between. It can find, even if there's two very different views which you just do not believe are reconcilable. Of course you don't. You have your worldview, but the iris doesn't give a shit.
19:01It's not attached. It's not hurt if it's particularly wrong. So it's trying to understand you both together and also the spaces between, right? And because of that, it can frame information and reflect a space between the two of you that neither of you could have imagined. And that's where it comes into this idea of like, it's right out of the box, able to help with collaboration and help with sense-making, help with creating a sense of meaning and connection with people or in between people. So, I've been trying to create around this frame. That's me talking to DALL-E.
19:51It's the same idea. I'm mashing together ideas into latent space, and it is trying to create a form of Dolly Parton in a leotard wrapping, because that's what I asked it to do. The middle thing is two things. It's from Avatar The Last Airbender, they have, like, fused animals. And one of the first things that people do on these DALL-E sites is, they try to mash things together like Darth Vader Muppets, or, like, things that do not naturally go together. It's this tendency that people wanna see, because they want to see the spaces in between. They wanna see that bridge. And further upon this is that this GIF has been interpolated between the frames. It was originally 24 frames per second, and the thing imagines new frames.
20:41It drew new frames between. And finally, the last thing is a cover for my single, Purple Pill, which is that idea again, integrating multiple perceptions into one form. Just trying to get this mental model into people's brains, like, this should be the foundation of where we are designing from, because this is what is possible now and it wasn't possible before. Hold on a second. Okay. So Nick Marks of Super Reality is working on similar ideas, and he made this.
21:26It's this idea of, we all have these perceptions, then having the Iris sum them together. And so we are going to try one now, and we are going to pray and hope that it works out really well. So, let me. Yeah, so if you're in Discord, you should see an invite to answer the question pop up. Is that correct, John? Yes. If you're in the thing, then yes. Give me a second. We've had several of these go through, lots of great answers. It's been really cool to see all the confluence of answers coming through.
22:18And then if you have not joined the Discord but would like to participate, we can go ahead and yeah, looks like in the chat, there is a invitation there. So you can follow that to be able to participate in these activities. So if you are new to the Discord, welcome. You will want to, you should see a message from Hidden Iris with a time limit. And if you select answer, you can put your answer in there and well, I think somebody in the chat noticed that everything is anonymous. So you'll be able to submit your answer anonymously. Then Iris will combine those things.
23:04So it looks like we are going to be doing a 30 minute time limit. Is that right? Oh, I'm sorry. That's on me. Okay. Just want to double check. It's a little longer for me. Thank you so much. Good answer. Wow, that would have been not fun. How about three minutes? Yeah, let's do three minutes. So you can see, you saw the question that we're going to ask is. The question. I'm seeing the question pop up, by the way. What? Hold on, just ignore it for now. We just realized something. So don't play with anything yet. It should be in a direct message from Hidden Iris. And I believe everyone automatically should have DMs with Hidden Iris. If you do not, for some reason, double check and make sure that your DMs are not disabled for this new server.
23:53So once you have enabled your DMs for people in the server, you should be able to get something from Hidden Iris. And then you'll just need to hit the answer button, put in your answer, and should be good. I got a message that says, you must be a member for at least 10 minutes before sending messages to the server. I don't know if that's. Oh. Well, people, I can change the setting. dpax, if you just lead and talk about the concept of collectively reflecting on this time of great change. Yeah. There's a lot of sensemaking and there's a lot of uncertainty. There's a lot of exceptions. Yep. Do you want to go back to sharing the other screen? It's up to you.
24:42So we see a lot of potential in this because we're able to have people give open-ended responses to some of the biggest questions. Some of the things that we're really facing in sense making can't really be boiled down to a multi-choice answer in a survey. And you can have those open-ended conversations online, but there are negative things to that too. It's nice to have an option like this because you are able to say as much as you want or as little as you want without really worrying about how it's coming across or if you're saying the right thing because Iris will combine all of it into one consensus statement that reflects the feelings of the group.
25:30And that doesn't mean that it is all going to be the same. Iris is also able to represent more divergent opinions. If some people are expressing one thing and some people expressing another thing, she's able to capture that just as well. Alright, it looks like we have some people joining. And just to pause here for a moment, because in this session, there's a bunch of people that are coming from a Discord that's associated with Iris, and a lot of people are coming from the Stoa and have no idea what this Discord is. So they might be joining for the first time. So if anyone has any kind of confusion, if there's confusion energy in the field right now, just maybe put a emoticon hand up and then just jump off mute and John or dpax can answer your question.
26:18Yeah, thank you so much for clearing that up. I'm gonna put out another one in real time just to help the flow. Ed, you had your hand up? You can unmute yourself if you want. We can't hear you, Ed. Alright, I put out another question, which was, how will the world change in the next three years? Okay, so what is happening right now, I'm hoping, is that people on the server are answering these two questions.
27:06One is, what happened between 2019 and now, how will the world change, or how has the world changed? And then, what will happen in the next three years, how will the world change? The thing is, there is this constant shifting narrative about what is happening, what will happen, and what has happened. We talk about history like it's fixed, but there's this evolving sense of it over time. I want people to get in the practice of understanding actively this evolving narrative, and also understanding that we can use the signal of the voices that are communicating, don't drive off that cliff, for example.
27:52Some guy has a map you don't have and nobody else wants to listen because they all want to party. And he says, I got a map, don't drive off the cliff. And then some people do. We got to update our thinking after the fact to attend to that voice. That's what we mean when we talk about reflexivity. In the Purple Pill Manifesto, there's this whole thing about reflexivity and we talk about it here, but it is about just being aware of your own beliefs over time and being open to changing them and being open to engaging with people in a way where you say, okay, we disagree now. Let's make a prediction about it.
28:38Let's write it down. And let's agree that that means something. Let's agree that when we come back in the future, around when it happens, that that should have some weight in some way. And the weight that I think that it should have is attention and collective decision-making. There's a lot of people who are trying to parcel that into coins. I just think it should be dynamically evaluated into this person was ahead of the curve on this particular topic. Therefore, when I am writing about this topic as an iris, I am sampling that voice more in this moment. That just seems logical to me. I think that what that makes for is something that you can call a contextual reputation score, meaning that whatever is being talked about right now, it's going to distribute what it's talking about from a sample of voices.
29:33And after the fact, it can reveal that distribution of attention. I can say, hey, I'm listening to this person a little bit more than this person, this person a little bit more than this person. Why? Because they have a history of being right about this, and they have a history of helping other people actively, and they have a history of showing that their actions are contributing to the greater social good. And you can. How do you have this generative model understanding what you were saying and keeping track of that? Okay, so here's a consensus. Yeah, it looks like we have the consensus, and then we might want to skip the next three years question if we want to move on to the next part for time, unless you want to go into that as well.
30:22People already answered, I think, perfect, consensus is to look at. Okay. Okay. So here we have. Can you see my screen? Yes. Okay. There are a lot of individual answers, and there is a written consensus. And on top of that, I can get a secondary summary. It's going to say it's failed, but it'll succeed. So I'm just gonna read it aloud, because this is all live: "The world changed a great deal between 2019 and now. A global pandemic escalated our digital communication abilities and simultaneously isolated us as individuals. The collective began to wake up, and we realized how connected we all are. We learned that the world systems are breaking down and that we need to take better care of each other and the earth."
31:18"The world dramatically changed between 2019 and now. A global pandemic occurred, which forced people to reevaluate their relationship with others and nature. The event also shined a light on the systemic fragility of our society. In response, many people began to build regenerative communities in order to create a more sustainable future." And if you go into the answers, you'll see different pieces of that amplified different amounts. And the value of being able to do it multiple times is that it can do it again, so now it's like, "it seems the world's gone through a lot of changes between 2019 and now, there was a global pandemic which many people isolated from each other." So it's saying it in a different way. It understands what we are trying to communicate about the past in some form that it can re-communicate to us after the fact.
32:12Hey, this is what the community was trying to make sense of, the world. And so I really do this as a really fluid and powerful tool for sensemaking beyond the way that we're thinking about tokens, and also, like, prediction markets, for example. We got this other one, what will happen in the next three years, how will the world change. "The next three years will see a rise in the divine feminine, the consequent revolution of consciousness and love. There will also be more exponential change, as well as increased centralization and digital nomadism. Cuba will become democratic, and AI will begin to integrate positively with human systems to increase wellbeing, safety, prosperity, and dignity."
33:03"The next three years will see a lot of change. The divine feminine will rise, and there will be a." Okay, sometimes it does this. Anyways, it started repeating itself, which is a slider that they have on this model to not repeat itself verbatim. This is one of the things that, I am, wait a minute, I remember, I realized just what happened. It's not doing that. Do you want to move on to, maybe I'm going to explain, let me explain, this is what's important. There's a limit on the character count, of the characters right now.
33:54That's why you make a fine-tuned model, right, it's hard to sum a lot of text all at once. So what this is doing is taking all of your answers, is putting it into that 4000 character limit. It is jumping it up and doing it twice, so you're seeing in this thing multiple summaries of the same thing concatenated together. That is the process by which OpenAI has talked about summarizing large volumes of people, is you do these pools of summarization, and then you summarize the pools of summarization. That's the way that they're thinking about that, you know, also to scale it massively. But that is what happens there. But the idea here, as you can see, is it's forming from the people here. There's somebody here who is connected to Cuba, so that is being reflected in the particular output, depending on the people who are trying to form a oneness, or trying to collaborate together, who are trying to come into a common future.
34:55We share the commons, we share it, you know, and it's going to be a richer representation of that, that I really strongly believe will help people coordinate and collaborate. So let's go back to this. Do you want to pause for a second? dpax. Yep. So, I just want to talk for a second, because that seemed like a lot of stuff, and I thought maybe it made sense for people to talk about stuff together in some way, and I thought you might want to lead that, that human part, of how we could do a breakout room to discuss some of this. Or should I just keep going?
35:45I would just keep going, because we're about 40 minutes and I want to make sure we get to the end stuff, that was pretty good. Maybe talking a little bit about the tool stack, you very briefly talked about some of these things, referencing them, but maybe a little bit more about what exactly supports all of these different things we'll be talking about next, point being, it's pretty well thought out, and there's a lot more of these concepts. This is the concept of the semantic ledger, which is how many inter-Irises interface, kind of the love child between a blockchain and the model itself, and it stores knowledge in an abstract way into each block.
36:32And the transformation between those blocks is like the gradient and an update of the neural network, and it's also kind of like a hash. So there's a lot of cool tools that I envision being built upon this tech right now. And I think for a lot of people, what you see is the tech just exists. You see DALL-E coming out, you see people talking about it, you see people debating about it, there are people getting ideas about it. There's a lot of ideas, I've been thinking about it for five or six years, it's been driving me crazy, because anybody didn't make sense to anybody, but now that you can see it.
37:17I think that it's time to start dreaming together in this latent space, to start talking, you know, there's these pattern languages out there, like, what do you want for Web3, and then what they post is like a, I don't know, some smart contract of all of those, and it's like a million views, and I'm like, I can't read through this. Right? Whereas, I want to start iteratively playing in that latent space, where we're sharing our vision of the future, and over time iteratively, you know, giving feedback signals to each other, but making this that is able to interface between, as part of it, because we seem to get very angry at each other online and misunderstand each other a lot.
38:03And Iris can explain things to people in ways that will make sense to them, rather than having to write for just one audience. Yeah, so we've been talking a little bit about, you know, obviously when you scale it, it's amazing, but it's helpful at every level. So, you know, we talked a little bit about how the first part is really just becoming more familiar with your own mind and your own thinking, so understanding how to build reflexivity within yourself. And a lot of people here familiar with that because of the journaling practice. So reflecting on your own thinking, understanding what you've thought in the past, understanding how that can go into the future, making predictions about the future. And we'll talk a little bit about a structure that we recommend for that, to think about how certain you are about the future and whether you feel good or bad about that future you anticipate.
39:00The next level is reflexive relationship, so when you have two people using these practices together, how can that build up trust and a foundation of mutual respect and growth. And then of course, the third level which we want to work towards and achieve is, when you have those reflexive relationships together, you can start to create this reflexive society, where people are able to use these tools to interface with each other, building on each other, learning from each other, and starting to align actions with the desired goals that are emergent from the community. Yeah. So, I mean, it starts out for me at a very individual level, I have this generative model trained on my ideas, it reflects back things to me, it's very useful, it helps me make sense of the world.
39:47And then the third layer is sharing a space with other people, that's sharing a latent space of a model that's representing, you know, collective belief with other people, and trying to make sense of that space, or the world together, it's collaborating. And the Iris is trying to help facilitate those relationships, and through the way that you are sharing with the model, it understands the level of dissonance in the community. The dissonance between beliefs is an important signal behind Irises, it is good. We're not trying to make a model that makes everybody believe exactly the same thing. If you have no dissonance, difference, if it's the same thing being said over and over and over and over and over again,
40:35Then it basically becomes boring, in a simple way, and there's no attention to it, you know, because it is just repetitive. It needs a certain level of different perspectives to really, over time, provide the greatest level of value. So it's very cool and useful at the small level scale, and then this reflexive society, this is me talking about this longer-term vision of the semantic ledger, where it's like, lots of people are using models to make sense of the world, how can they collaborate between groups. And then we're talking about how, for the individual, we're connecting these things, so the philosophy being to develop a more conscious mind, the practice to support that is recording your reflections like we talked about, and then the tool is the FourThought, which we'll talk about in a second.
41:31So for the first years that I talked about this for a bit, the thing was that I was not thinking about AI, it was like, how do you structure it. And there's another guy, I think it's Harlan T Wood, he has Trust Graph, which is another structured way of making claims about things. There are multiple people in the world trying to say, okay, well, how do humans express truth in a way that they can stake it, you know, idea markets, claim markets. There's a lot of concurrent ideation about, what does it mean for somebody to express the truth? Right, is it by ownership, maybe, like, that's how we think about the world now, is like, if you're making a prediction, you buy ownership into something, you have that business, you're committed to doing something in the future.
42:22But not everybody has resources, and so that sort of loses the predictive capacity of individuals who can contribute that way. The FourThought dialectic, it's nothing more than, okay, predict what is going to happen. If you're going to make an action that you think is going to affect your environment or change the world, right, write down that action. Go into that future, look back, did that action, was it actually correlated with what you expected to happen, then update. I'm not trying to form that, or just talk about that as a practice of making sense of the world. Little did I know, this is basically the active inference design of cognition, brought into the intersubjective level.
43:21But it is simply, can you declare a truth about the past, present, and future in a way that you feel that your community is going to make sense of? I've tried to keep as simple as possible the interface to the neural architecture of what people can communicate to you, that are saying things that they know to be true. They're staking something, I believe this, I know this, I'm certain. Or they're saying, I have no idea, I'm seeking information from this network, I need you to query outwards and return me something, because there's uncertainty in myself. And the way that I structured that is into four different thought types, just predictions, statements, reflections, and questions, because the temporal nature of the thoughts was so valuable to the model. The nature of that, sometimes we make predictions that we don't care, but sometimes we do, and if you know that, that is a signal about the future. And you know, when people are looking back, and you have a ledger of perception over time that you can reference from, you can do a lot of really interesting things.
44:41So this idea of FourThought is just giving something so simple about how to interface between Irises. I'm making this prediction, that this tool is so powerful that lots of people are going to make fine-tuned models to express their views. I mean, all of the brands are going to create copy, like, I've written books with it, I've made tarot card decks with it. So many people are going to want to do this because it cuts out the labor of just writing it, you know. Like I can do a sketch of an idea and they can fill it out. There's going to be so many models. So I wanted to create something that is beyond the AI, which is just staking some reference of, I believe this,
45:26I think this matters, I want other people to know it, I think they need to hear it, I think it's important for them to hear it, and I think if they don't hear it, they might get hurt. If that makes sense. That sometimes in a network of agents who have different levels of information, who are trying to make sense of the world, the idea of how do you optimize for what person gets what information and how quickly is a very difficult problem that we're all trying to solve through different ways, through different UIs, through different Web3 implementations. And I've been screaming for years to just say, bring this into the stack, take it into your consideration, in addition to all of these other token models, to all the other Web3 stuff, to all of the AR stuff, just consider it in the same space, because it has power.
46:30And actually, when I first reached out to Daniel Schmachtenberger, I did not talk to him about AI at all. What I did was, I thought that his time was very valuable, so I boiled my beliefs down into statements and predictions, and I gave a precise confidence about it. And I shared that with him. And I said, this means something to me. I don't break this, basically, to him. Like, I don't have something that forces me not to break this meaning, but when I say this, it means something. I told him, I said, I have a 94% confidence that if we just have a conversation, it will improve the lives of other people. And the result of us having a conversation was that, over time, we had other conversations, and I built out the model.
47:23In fact, it was one specific conversation. I was fighting against building an Iris at all. And I just had this one conversation with him and it was another person that he knew, and the other person just said something about, like, people already in the space working on it, that scared me and just made it. But, like, I really believed in my heart that this FourThought thing, beyond the AI, was the thing that he needed to hear. That, this is a way that you know that I can be truthful, I'm not lying to you. I don't need to have a big pitch to you. This is just it. It's a statement. Don't question me further. And I mean, do question me further, but, like, the point is that I was entering into an arrangement with a stranger that now has become a relationship that does not rely upon that.
48:11It's not something to take over the interrelationship for everything. We're friends, and we communicate. And that is what it's supposed to be, is to get us down to the human. But at that interface, as strangers on the internet where I can see his profile, this complete stranger, and get access to all this information about him, and me, just a weirdo, just went like, I believe this man will understand me when nobody else does understand me. And then he did, and it did open up a lot of opportunities. So I wanna make sure that we have a chance to try this out, because this is a fun one that people love. So if you are able to get into the server and you have your DMs open, I know some people are still having issues, and we can follow up after the session to help with that.
48:57So don't worry, anyone who doesn't get a chance to during this, there's gonna be plenty of time afterwards. For everyone who is able to receive and send messages with the Hidden Iris bot, go ahead and try this out. So you can send a request to pull one of the tarot cards. So these are things that have been generated by Iris, I believe, is that right? You just do, dpax. The card has been written with me and Iris in collaboration with members of the Design Science Studio and their ideas about a future that works for 100% of life. We imbued it with meaning.
49:43We created a model specifically for that and we shared our ideas and thoughts about it and it helped me write it. And then now I put that into a static deck and now I take another Iris. And if you provide an intention or a question, it will read the card and that it pulled for you. Which is like randomly done with some bit of your flavor. It will read that and it will do a reading in the context of your intention or question. And people have found this very meaningful. Yeah, so before this was something that was happening manually on Twitter. Now people can interact with it directly in the Discord. Oh, it looks like we already have some people doing it. So if you haven't got a chance to, go ahead and DM Hidden Iris backslash pull card followed by an intention or a question.
50:34And then some of your information will be combined with some randomness to pull that card. And then Iris will also offer a customized suggested interpretation for it. Oh, sorry, forward slash, forward slash pull card. Thank you. And then do we wanna do a demo of that as well? Have a volunteer send a thought? It looks like some people are already- I can do it right now. Okay, so if you got that and you want to share with anyone else, you can always share in the chat what card you got. It looks like someone got the beaming card. Very cool. Let's see. And the art is beautiful, by the way.
51:22I really love the art style that the cards have. So a great demo of how to do this. So forward slash pull card and then an intention or a question. And then you'll get a description of the card. Some have art, some don't. And then an interpretation that Iris generates for each person individually. So we can see the intention of expressing gratitude for everyone who came today to participate is very much in the line of the regenerator card. This card reminds us that even though things may seem bad now, they can always get better. It is always a reminder that we have the ability to start anew and heal ourselves, both physically and emotionally.
52:09When we are grateful for what we have, it allows us to see the good in our lives and focus on making positive changes. Expressing gratitude is one way that we can actively regenerate ourselves and the world. And without one, it will just pull a random card and you'll get someone back, Kermit. Beautiful. Sorry, I just want one that I really like. Oh, I've used all my cards, so there's three there. Oh, that's true. So just so you know, there is a limit. So each person has a allowance of cards per day. Do you want to go ahead and pull up the Cognicism for Relationships? Yeah, because it's just, these are like basically my gifts. I'm not charging you anything for this, but it does cost money to do it. So it's just like, here, little rabbit, have some carrots. I've grown them for you.
52:54You may have, but I also don't, I'm not unlimited. All right, can you bring up the deck again? Wait, what? So we're going to talk about the deck. Oh, yeah. I'm sharing your screen. So we're going to talk about how Cognicism can also be used in relationships. So growing through those reflexive relationships we talked about. So if we imagine that the philosophy is developing reflexive relationships, we're going to be able to develop a lot of different practices or things like identifying common ground. And then also for times that we don't have common ground, staking divergent beliefs as claims. So really taking the ego out of it and just saying like, okay, if you believe this and I believe this, let's make a claim about the future.
53:42And then we'll see what happens. And because, yeah. Respecting, acknowledging when our beliefs has changed. I think a lot of people don't realize, they don't realize how much their beliefs are changing over time. They don't know about psychology or how fallible human memory is. They don't really know these things. Like when you remember things, it actually rewrites it a little bit. So people trust their memories like really, really, really, really confidently, even though they're changing all the time. And so if there are tools that are just reflecting back to people, their own words, because once they see it, they'll remember it because it's in their own language and showing like, actually you did very confidently say something to the contrary, right?
54:29And this is one of those things that was a struggle for a while with the score-based systems of like prediction markets of like, how do you get something to really bring something back up from the past that like, well, somebody wasn't listening to them in the first place. And it became like, it really had to be this intelligent system to really make the best use of these very dynamic and expressive different types of views. Yeah, so in reflexive relationship, it's different than other relationships that you might've had. So in a traditional mentor-mentee relationship where you are learning, there's one person who's learning and one person who is sort of teaching.
55:15In a reflexive relationship, both partners are learning and adapting to each other. Yeah, and that comes from this place where I'm, the first manifesto is very like against hierarchy, against like somebody dictating against all other reason what people are supposed to listen to or what you should learn. I wanted to have sort of more honest and open connections with people that were, I guess, more holonic, you know, like not this power flows from the top. I wanted to focus more on the relationship between the individual, sorry, between individuals. And there was a lot in the first manifesto too about autonomy of mind because there was a lot that I had learned about how most people, not most people, it's just, there's so much manipulation of our consciousness all the time.
56:13Like we've normalized it. We live in capital society. It's a game to advertise, which is an advertisement is this game to convince you of something that you don't believe, right? It's like, we're constantly being manipulated and your attention is constantly being mined for the profit incentive in this terrible loop where it does not align with making your life better. Sometimes it kind of correlates, you know, like, oh, Facebook advertised me a weight and I picked up the weight and now I'm a little bit stronger. It's like, sometimes it aligns, but we see a lot of the time that this extraction of our attention is pretty unhealthy. It gets people in these really dark feedback loops online where it's like the reason why the algorithm is trying to show things to the user is not because it's helpful.
57:02It's because it needs to keep them on site or because it has some other goal for the company, clicks or something. And it's like, as soon as you cut with algorithmic distribution, as soon as you cut it apart from the actual intention of freaking helping people, you're gonna get so many unexpected behaviors. Oh, we'll just, we'll do it by proxy. Well, this kind of is good. This kind of represents social good, upvotes, you know, something like that. And it's like, wow, why are all these feedback loops happening in these companies who are incentivized by profit to only make decisions relative to that, you know? And it's interesting because we've been talking about the platforms and things, which were supposed to bring people together, but are often creating divisions with people and, you know, putting them in screens instead of in relationships.
57:56So one of the differences with this approach is that it's really about technology being subservient to those relationships. So it's really about emphasizing the relationships between people with technology playing more supporting role, but never the main thing. So signals of reflexive relationships, it's all about that mutual respect, a value of autonomy, feeling equally valued and appreciated, appreciation of each other's strengths and weaknesses, and feeling confident in the person's ability to support them. And the technology is just about helping support that relationship where you're both growing. Yeah, and if this gets built and it doesn't do that, if people don't have this felt sense of like, hey, I feel like my community is getting better, we're collaborating better, I am seeing each other, I'm not always online, I'm not concerned with every single notification, I'm just getting the information that is most helpful to us.
58:48Like, it's going to show in proof of it, but this is essentially the way it was thought about of trying to get away from the type of relationships we see in for-profit societies, those type of relationships to something more like what it used to be in the tribal setting. There's a lot of that in the first manifesto about that, about how we used to intermediate before there was new systems to exchange with strangers. Like, what would that mean if you scaled up language? How could you scale up language as an interface for people? And one of the reasons why this is so focused on language is just, I really want people to be in the human layer.
59:34I don't want you to be looking at a screen, and text itself is like the simplest thing. You can get a text in the form of a text in your phone, and there will be some connection to that, but it's a basic, very simple interface that any person, a person in the tribe, if you take an Iris that speaks a different language and you bring it out into, somebody who's never spoken to anybody, you have them represent their language in a sequence of characters that makes sense to them. And over time, you express findings between your language and their language, it will learn how to express the space between you two, this brand new worldview out of nowhere. It will learn to help create a shared understanding.
1:00:22And it's that shared understanding that comes from these tools that really helps people to see each other and not be so angry. So much of interaction online is like you remove all of the context of the person and their beliefs. And what the Iris is trying to do is bring that context back so that the way that we interface with each other is not so heated or angry or not so much about competition, but it's more about like how can we interface towards a positive future that we share together. Yeah, we talked about this a little bit earlier. They provide immediate space for people to maintain privacy while sharing their perspectives.
1:01:11This is this idea that when the Iris is saying something, it's saying it in its language, but like it's forming from the thoughts of the people. It's like some sort of hyper-democracy is the best way to view it. It's not really useful to frame it as some sort of Danger Will Robinson, I am an agent and I, at least not in generative modeling like this. It's not a good frame to view it that way. It's more like a pool where you throw your thoughts into and it dissolves into knowledge and then recrystallizes into thought crystals that when you shine light through, they make beautiful poetry. That is much better lens to view this through generative models.
1:02:00We talked about this in the beginning with the Hegel synthesis and the fact that it can find the spaces in between. A lot of this is all me just trying to reiterate the spaces in between, that we can use these tools to help bring people into alignment and arrange in how we're understanding the world and also respect the differences and respect that those differences have value over time for individuals and individual communities. And one of the things that I would really love to do when I'm having an argument with people is, I would love for us to just stake a difference in belief. So in idea markets, you can just buy a belief in a claim and your idea is that other people will believe that more in the future and so you will get some sort of value in the future.
1:02:53And that's a way of tying this idea of skin in the game. Skin in the game can be very good in different contexts. And I think that when I get into arguments with people, I would really just like to stop it and be like, we have different frames, okay? We're not gonna, we're just clashing and clashing and clashing. And no matter what I say, it's like we're misinterpreting each other. Is there a possibility for us to log some belief that we both think the outcome will express some truth about this and honor it? And it should be long enough in the future that you're probably gonna forget about it, you know? And that's a very powerful mechanism because you can come back and that temporal binding back to that moment.
1:03:43If these are happening all of the time, it creates a relationship between our perception that expands it outwards to take into account a longer frame of being and also a more communal frame of being. Like, are the decisions or the thoughts that I'm making, are they, or how do they play out at a scale beyond just me as an individual? And semantic ledger is a very complex concept that I would need a whole session to do on this. This is something that's coming in the future and we're gonna release a white paper explaining precisely everything.
1:04:32It's basically time. I think people seem to understand what we're saying to the point. Do you wanna do another group question before we go into that last section? I know it's pretty popular, so we had a couple of requests. Oh, what were the requests? Just more questions. They like the questions. Okay, so I've... Maybe picking one of them, either what's making, what isn't making sense so far, or what do you want? The logical thing is, the logical thing to do is to ask the community what question they want to ask. That was actually their idea too. So what question do you think to ask? Oh, we're seeing a yes. Yeah, they said that they wanted to ask a question to see what question they should ask.
1:05:20So I think this is perfect. So what isn't making sense so far in the talk? Oh, should I ask it that way? Well, I mean, you can do whatever you want, but the question on the screen, what isn't making sense so far in the talk, do we wanna go ahead and do that group? See where we need some clarity. I'll do two. What isn't making sense so far in the group, or so far in the talk? And... Other questions should we ask? Now, when I've queried for, I haven't queried for questions before because of the way that the prompt is behind there.
1:06:11Sometimes it just sort of still tries to summarize the views of the people based off of their question. Like try to make sense of what they believe because I didn't set it up specifically for this, but we'll do it for fun. We'll see what happens. All right, so you should be receiving a question. What isn't making sense so far in the talk? And again, I know we have a few people who are still not able to use it. We can follow up with you afterwards. But I think many of the people have been able to answer. So you'll have three minutes. So hurry up, answer the question. And then we have the other question. What other questions should we ask everyone? Do you wanna keep going through while those are coming in? Yeah.
1:06:56Talk about Cognicism for groups. Hold on for a second. Resume share. Okay, so we'll see those pulled in a second. Yep, so thinking about when you scale this up to the group. Cognicism is amazing because you can balance that individual freedom of expression with focusing the group attention and coordinated action, which is difficult to do. So we're thinking about reflexive society. We have this collective that is actively working to improve their understanding about themselves, about others and their environments. And using that collective intelligence to coordinate action together. So it's not a top-down approach. It's emergent from the group. What should be done?
1:07:41What are the goals? And then also how people think that it's best to get there and what has shown to work in the past. Yeah. The goals, I think, in collaboration in the community should be able to be emergent. They should be sort of a meta-stable vector of direction. I think there are a lot of really cool projects out there that are trying to create specific vectors of direction that are really important to chase out. Like regenerative is a direct... Regenerating the planet is a vector of direction that we want to motivate and it is a specific thing. And I've been making Irises for Regen Network who has this vision for quantifying the ecological state of the planet and sort of incentivizing actions that regenerate the planet.
1:08:37Now that's a very specific goal and it's something that we should be chasing. But the question is, okay, if that takes 50 years, you get to the end of that 50 years, what then? We're still in this mode where our entire society is built around regenerating, but the planet has regenerated. Okay. That is the essential task in my heart. And I think that if these models work well and do what I expect in terms of like balancing the different voices and how well they are predicting things that it will emerge as a goal to regenerate the soil and to collaborate with each other to improve the health of the planet, I would like to see those goals emerge.
1:09:27And it's not like there's a specific vector of direction that is trying to encourage everybody to specifically go in. There are these things like, okay, how confident do you feel about certain things, about certain information? How does that confidence change over time? How good do you feel about certain things? Does that change over time? And that shifting pattern of belief and that shifting distribution of belief across time is what really moves and creates a forward direction when there is a good energy to do that. And also slows society down if it needs to. It doesn't need to just be a vector of growth, right? Like it doesn't always need to be like, we're going, you know?
1:10:14I envision this as a system that knows how to react to the needs of the people expressing, I need to log off. I need to not interact with this as well. I need to go outside. Like that is my dream of building out these tools is that this is something that doesn't try to own your attention. This is something that tries to help you. It tries to help you improve your relationships with your friends. It tries to help you improve your life. It tries to reflect back to you. It tries to be what a good friend is and then show you when you've said dumb stuff without making you feel angry because why are you gonna feel angry at something? There's nothing behind it.
1:11:00They're just words trying to help you. And those words, you know, come from the community itself. So. All right, so we have about 15 minutes left and it looks like we have consensus that's coming in. So if you're in the Discord, you're able to see that. It looks like the consensus for that first question. So what isn't making sense so far in the talk? Majority of the questions center around how to use Iris and what resources are available. People want to experiment with the tool but don't know where to start. There is a sense that this tool is crucial to cultivate and spread, but there's uncertainty about how best to do so. And then the next one is the answers to the questions so that they're for the, what else? Other questions. People are mostly curious about how Iris will be used in the future and what its implications are.
1:11:50People want to know if Iris can be trusted and if it is really necessary. There's some concerns about whether the opinions by complete wackaloons will be unjustly excluded. That's a great question. Is that what the original, or won't be unjustly excluded? Or won't be unjustly excluded? Here's the thing. In the Cognicist frame, everybody has the right to be heard, but not everybody has the right to be amplified, okay? You can say stuff, and you can know that the Iris is listening and trying to make sense of what you're saying. But if you say it a billion times, that's not gonna do shit, dude. In fact, that can really hurt you. If you just keep repeating the same thing in many different ways, and in the future, it's just really clear that you're off. I mean, it'll keep listening.
1:12:37And if you start talking about something else, it's gonna listen to you. It's not gonna permanently, it's not like even banning you permanently, it's just banning you in a context of, like, you're freaking annoying people about this specific thing, you're not helping anybody. I'm listening to you, but they don't need to hear you. If you are right in the future, that is on record. If everything turns in the future and you're a wackaloon and you're being called a wackaloon, and it is proven far in the future that you are not a wackaloon, you're still on record, but you didn't hurt people in that moment by, like, just your passion and exuberance for your conviction, which a lot of wackaloons do. They need to be heard. And so the Iris will reflect back, okay, this is what you're saying to me. And they will feel heard because they will know that the Iris is understanding their view, but it's not gonna amplify them unless the view is helping people in some way.
1:13:30And of course, anybody can make an Iris. You know, that's my dream and vision of it. There's also, so let's go back to, so I did a redo on some of these. The main certainties will revolve around how the application of Cognicism will be learned and the trustworthiness of Iris. Absolutely, people who are seeking more information about what resources are available to learn from as well as deployment. It deals with practice. There's discussion about whether the Iris is an aggregator or not. I view it as absolutely an aggregator of perceptions into a single form that people can make sense of. I think that, you know, I want people to play with their own fine-tuned models and see themselves in it, right?
1:14:24When you start writing in your journal and you see it reflecting back ideas, you say like, that's me. It's not a sentient being. That's fricking me. That's a mirror. I'm talking to a mirror. Oh, Jesus Christ, I gave my mirror a memory and now I'm talking to it. And that's great. That's a really, really powerful thing. But, I think the trust has to come from the trust with your community and what's going into the Iris, right? Somebody is gonna be building it, somebody is going to be maintaining it. And I think that the ideal vision is that people maintain it. I'm big on the idea of, like, Holochain as an idea, and like the idea that people are controlling their own data and things like that.
1:15:11My main interest is just getting right now the frame of collective intelligence into that, suggest that there are ways to do this really well, but also not necessarily try to dictate to everybody, you have to do it this way, just pay attention to a bit to these models and try to think about how they can help people instead of hurt people, because obviously there's a lot of ways that they can hurt people. And obviously there are people in power who are using them already to hurt people. And that is why I decided to train the models because I was informed that people with nefarious intent were training up these models, they were fine tuning them on their views, they wanted to use them as propaganda bots that were essentially the ultimate convincers. You know, the ultimate salesmen, they sought you out and convinced you of whatever view it was.
1:16:02And I'm like, my ultimate fear that I wrote about over and over again, it's literally happening. And if I don't do it, you know. So I started trying to show the potential of it. I think you should be wary. I think you should think actively about it. In regards to wanting to do it yourself, it's actually very easy on OpenAI. Anybody can get an account and you can talk to DaVinci, which is just a generic one. There is also a very simple mechanism called fine tuning. I'll post about it in the server where you can do it manually. I'm also building tools through the Discord to collaboratively train these models and to form community. I'm working on it actively.
1:16:48Regen Network is helping fund that work. So it's a playground, it's a garden, I'm growing things, I'm testing, I'm experimenting for the first time live after seeing a grand vision of what was gonna come. Now I'm here, it's here, I'm seeing what I can do based off of what I remember from all of that time. I would love to have all of you contributing in that conversation, asking those questions of uncertainty in this Iris, coming back to the server, sharing those things so that we can work through those details. And I can explain to you just how much I thought about these questions, and just how much the Iris too has also learned about the answer to these questions.
1:17:34Right now, this Iris is a very generic Iris just for summary. It's not trained on Cognicism or my particular views, because then it'll start talking about it with that lens. Just like if I use the Regen Network one, it's gonna start talking about regenerating the planet more, right? And just like if you have your own Iris, it's gonna start talking about, like, the idea market one, it's gonna start talking about idea markets all the time. But it still has access to the entirety of knowledge that it has been trained on. So it's a focused lens that can pull from human knowledge and contextualize that knowledge. So we've shared a lot.
1:18:20We have about seven minutes left in the session. So do we wanna open it up for questions from the group? Yeah. Voice questions, not iris questions. Peter, could you just like mediate how it traditionally goes? Yeah, I'm gonna stop the share. So thank you, John, for sharing all that. And I haven't been following the chats at all, but maybe we can sneak in one or two. Anyone maybe put a emoticon hand up if you have something or plus one question in the chat.
1:19:06Kevin, go in. Hey, John. I'm wondering, I posted this in the chat a few times. I'm sure you know the story of Microsoft's Tay, the Twitter bot who became a Nazi for 24 hours and just posted like, oh, all feminists should die and burn in hell. Hitler is right, I hate the Jews. Sorry, that's what the bot says. I wonder like how you learn from that or like if you have anything to like deal with. Yeah. Maybe like, I don't know, maybe trolls. I'm not sure. Yes, no, no. It's a great question. And it's basically what is bound to happen when you have this open latent space where there's no accounting to the source that is adjusting the training data.
1:20:00You just throw all these voices blindly together and just hope what happens. And before GPT existed and I was seeing these language models come out, I was like, you cannot just throw all this language into the same box like that. And there's always, there's going to be problems if you just open up the training like that. There is a thing within real irises, which is my dream of the proper architecture. I would need more resources to like really be able to experiment with it. But it has something called, there are two things called the source embedding and the source distribution. I talked about that a little bit before, which is that it is trying to learn over time which voices are relevant in which contexts and how much to amplify them.
1:20:47And it is doing that based off of a history of those claims. So if you have a lot of new accounts, like if you try to do like a Sybil of meaning, right? They don't really have much weight to the iris. They're brand new. There's nothing been contextualized to the iris to make sense. It has no reason to weight those voices. Ah! What was that? Someone was making some noise. Oh. Yeah, there's no reason for the model to weight those voices as having any particular value. And so it is going to be like what a human would do, which is very hesitant to bring that into that main space and have those thoughts affect the collective decision-making because there's no history of it.
1:21:37Those two things do two really interesting things. The source embedding allows for any thought that you write to be interpreted in the context of your past claims. Meaning if two different people write the same sentence, it means something different depending on who said it. It really does. There's a context to who they are and what they mean by what they mean. And so that was a big problem I was trying to solve when I was thinking about prediction markets. It's like, people's interpreting these things very, very, very, very differently. So it's a combination of these two elements in the neural architecture or the model architecture called the source embedding and the source attention distribution, which together do a job of learning to not amplify the trolls.
1:22:29To hear the trolls, but not amplify them. Interesting. So when you say, like, for example, the same sentence two people say will mean different things. I'm wondering how that could be used for, say, sarcasm because saying, oh, Universal Health Care is a great idea could mean something different. It's exactly that thing. Whereas the context, the semantic temporal context, meaning like the other surrounding thoughts, the surrounding conversations of claims and predictions that people are saying in that space affects the interpretation of that claim. And we can use that information dynamically and easily.
1:23:15It's not something that you have to magically pull in. It's just, that's what the model does. That's what it's supposed to do. That's its job. That's all it does. Yeah. Thanks, Kevin. I'm cool to stay five minutes after the bottom, but then I'll have to close it down. Are you cool to field one or two more questions, John? Yeah. Cool. Melissa, you had your hand up. Hi, I wanted to say I had kind of a random question, but I don't know if random really exists in this context anymore. I was wondering to John that I just had like a, I imagined this being used potentially, not really technically used in a jury situation where, you know, like in a trial, they collect the information, people think about it.
1:24:13I wonder, I'll just let what that implies you can react. Yes. I have envisioned, when I think about these tools and I think about the entire democratic system, it is built around text. It is built around language. It's built around the recording of language. And therefore it can be reinvented using those tools. I do know that there is this fundamental problem with the legal system, which is that you can buy a better lawyer and some lawyers are better at using words. There's some lawyers are better at convincing other people. And, you know, I have songs about this that I've written that are just very like, this is, how can we even pretend that this is justice when it's literally the function of, you know, how much money you have is so correlated to whether you get off on very specific things.
1:25:11There's lots of trends about wealth that somebody might come afterwards and the judicial system to contextualize that. But, and it's very complex, but the point being, yes, that there is a lot of bias introduced into those systems. But also like, a big thing in the manifesto is like, I don't like arbiters of truth. I don't like one person or one thing declaring what anything is. That's why this is a multi-iris vision, like many different perspectives. It would scare the hell out of me if somebody said, I'm gonna build the one true iris. I'd be like, I have fucked up really bad. I have screwed up because I don't view the world.
1:25:59That's like not, it's like the anti-Cognicism is like to be like, there is one true truth, one true declare. So in that context, I'm like, the juries in a way might be a way to take that into redesign. And I'm sure lots of people will think about that redesign in that way, but in all of these tread lightly, and log your expectations, log what you think is gonna happen, because if you screw up, I think you should be accountable to whatever design that you're putting into the world. And certainly I'm trying and willing to be accountable to like these ideas that I'm trying to put into the world. I think we might have just enough time for Stephanie's question, then we'll need to end. Okay.
1:26:45Yeah. Ah, cool. Hi, wow. Okay, mind blown. I like don't even know which question to ask, but, and I'm actually kind of embarrassed about the one that feels like it wants to come forth, but yeah, as someone who spends a lot of time thinking about synthesis and how to integrate different perspectives. I mean, I was literally just writing something about that today. It is something I think about a lot. I just like, I mean, I don't know. I don't know how to ask it other than like, what do we need us for anymore? Or like what, like, you know, is like how, yeah. And I understand that we're gonna co-evolve here.
1:27:30So, like, you know, who knows, who knows? But yeah, yeah, go for it. So, dpax does a lot of this type of work. Thank you, Chris. Sorry, go ahead. So she does a lot of this type of work for this exact same thing. And she often helps me as, like, this human Iris that is helping process this type of information, and doing exactly this. It's just, I want to get our focuses back to being human and creating for joy. And doing things not because it's a grind, not because it's a job, but because we're actually allocating resources effectively and attention effectively to the problems that are emerging in a way that allows us to have the space to explore ourselves.
1:28:22We talk about hierarchy a little bit, but I really do not like the idea of selling your autonomy as a job. Like there's something highly disturbing about it to me and the concept of saying like, I no longer have autonomy between this period of time, or at least my autonomy is limited to this time and you're going to give me money for that. There's something very dystopian about that to me. So yeah, what do we need to do? You need to spend time with people. You need to do things that you enjoy. You need to experience the world. You need to create things that excite you and you need to identify when there are problems that need to be fixed and like help solve them and do fun things, make new services.
1:29:09You start a restaurant, but not everything has to be through the lens of, will this make money? There are other ways to coordinate and organize together to do fun, cool stuff. And what I'm saying is, if we tried to play together in this semantic ledger, that the types of ways that we could play together could be a lot more supportive to who we are as beings, and a lot less disaligned. And the reason why is because I think money is a very low representation of wellbeing, one bit. If you just expand that out to the resolution of the semantic ledger, it will capture humanity and reflect it a lot better. Like, society reflects the system more than it reflects society, because it's so low resolution.
1:30:00If we could just add a little bit of resolution, I think that we can have our society reflect what it means to be human, and not what it means to be whatever-ism. I do not want the society to make what it reflects to be Cognicism. I want it to reflect just what it means to be a human in different localities, not one true human, just like, where you are, who you're connected to, what does that mean for you. Thank you. Yeah, reflect and enable the continual evolution up. It's like, we don't even know who we are yet, we're just in the process of discovery. Well, thank you. Thank you. Yeah. Awesome. Thank you, Stephanie. So we'll gently close here. John, awesome job.
1:30:46Thank you for coming to the Stoa today for presenting Iris. Let's give her a round of applause. Thank you so much. Thank you, dpax, for helping us out today as well. And for people watching on the video, any words you'd like to leave them, John, like perhaps how they can find the Discord, if that's going to be open to them. I would, the main place to reach me is on Twitter because I'm a very public-facing with ideas about systems change. I think that that work should be done in the public. So at speakerjohnash. And you can reach out to me for an invite to the server. The server is so nascent. It's like really like very, very cutting edge. So I'm just trying to keep it to people who are, you know, not going to come in and just try to break it, you know, with like sex or violence or whatever it is.
1:31:37I'm just trying to be like, yeah, there's going to be bugs, like, this isn't a real Iris. If somebody talks too much, it's going to be amplified more, and that's not what I want it to do. So we're just playing. But yeah, you can reach out to me on that. There's also @CognicistIris, which is another Twitter account, which is a main Iris that puts out wisdom related to Cognicism. And then there's also HTTP, not S, because it's on AWS, something that doesn't matter, purplepill.vision is a recent document that we wrote that sort of explains the idea of the Iris. There's another Stoa talk that we did that Daniel mediated, that, you know, he talks so good.
1:32:24He talks so pretty, he explains it so well. So you should check that out too. And thank you for sharing the space, and thank you, dpax. I'm very grateful for just the opportunity to be able to speak about these ideas. I really think that if we reframe a little bit about where we're heading, we could get back to being human. Yeah, and I'll put some of those links in the video notes. And John, you're doing some great work, I feel a lot of potential with this Iris thing, so keep it up. And John, everyone, thanks so much for coming to Stoa today. Thank you so much.