<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Kristian Stålne on Medium]]></title>
        <description><![CDATA[Stories by Kristian Stålne on Medium]]></description>
        <link>https://medium.com/@kristianstalne?source=rss-792e23cb7029------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*NWqv2D40EUx6lHP90C3MzQ.jpeg</url>
            <title>Stories by Kristian Stålne on Medium</title>
            <link>https://medium.com/@kristianstalne?source=rss-792e23cb7029------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 17:27:51 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@kristianstalne/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[An Adult Development Perspective on AI]]></title>
            <link>https://medium.com/@kristianstalne/an-adult-development-perspective-on-ai-2ad4cca5a398?source=rss-792e23cb7029------2</link>
            <guid isPermaLink="false">https://medium.com/p/2ad4cca5a398</guid>
            <category><![CDATA[adult-development]]></category>
            <category><![CDATA[complex-thinking]]></category>
            <category><![CDATA[meaningmaking]]></category>
            <category><![CDATA[ego-development]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Kristian Stålne]]></dc:creator>
            <pubDate>Sat, 26 Oct 2024 19:41:23 GMT</pubDate>
            <atom:updated>2024-10-26T19:41:23.327Z</atom:updated>
            <content:encoded><![CDATA[<p>Like many others, I have experimented with Chat GPT and followed discussions about AI’s future role. My university has held many seminars on generative AI, both from a practical perspective regarding our teaching and a mathematical perspective regarding how machine learning and neural networks work, so I have had the opportunity to reflect a lot myself.</p><p>In this article, I want to try to answer questions like if AI can evolve past us in intelligence (whatever that means), when AI becomes AGI (artificial general intelligence) and what it is, if AI can become self-aware, if AI development is a threat to humanity as, for example, Max Tegmark and recent Nobel laureate Geoffrey Hinton believe and why so many have become nervous or overwhelmed by the latest breakthroughs. I want to answer the questions using an adult development perspective, which is my main field of expertise, and I will try to stay at a fairly basic level to be able to describe more general principles around how we, as well as different forms of AI, work. I conclude with some sources I have found useful in this discussion.</p><p><strong>An introduction to our thinking</strong></p><p>I want to start by describing a common and simple notion about intelligence that I feel is prevalent in AI discussions, as well as about how we, as humans or organisms in general, work, although it is rarely explicitly stated. This model that I illustrate in Figure 1 below consists of two parts, the reality we are surrounded by and how we think about it. Reality could be the physical reality of matter, energy, time, space, etc. The other half illustrates how we think about the world and represent it. It is built up by the information we take in according to the upper arrow. According to the lower arrow, we can influence and change the world by how we act. In this way, we are in a constant feedback process with our surrounding world. We see the world, understand it, act on it, possibly change it, etc. A simple feedback loop.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/222/1*3Iba6930vZhmrbNlL2CLHg.png" /><figcaption>Figure 1. The relationship between our inner, which is understood as only our thinking, and reality, which can be, for example, the material world.</figcaption></figure><p>The arrows at both ends, top and bottom, are here associated with commonly expressed ideals. When we take in information from the external world, we often have the ideal of being objective and representing reality accurately. We should not let our personal opinions influence this view and we should watch out for different kinds of bias. At the bottom, the ideal is often that we should be rational and act in accordance with this correct and accurate understanding of the world and what produces desirable results.</p><p>As you can imagine, there are a number of objections that can be raised to this simplified illustration, but first we can illustrate what happens when it doesn’t work as intended. If we take the climate issue as an example, it is usually argued that we should base our understanding of the climate on the best scientific sources and act in a way that reduces the negative effects. Problems at the upper end are usually understood in terms of climate-denying actors who spread misinformation and doubt. Although we have been discussing what true knowledge is since Plato, in recent years, it has become even more complicated with our abundance of (not always accurate) information and concepts such as fake news, deepfakes and post-truth. At the lower end, the gap between knowledge and action is often discussed — how we continue to act against our understanding of what we actually need to do.</p><p>One might think that this illustration is a strawman and an oversimplification, but I perceive that many arguments about if and when AI will become more intelligent than us are based on this kind of understanding, albeit implicit. For many, there is a preference for starting from as simple explanatory models and building blocks as possible, which we usually call Occam’s razor. Especially among physicists and programmers who are at the centre of the AI revolution. However, I am interested in what happens when we make this model a few steps more complex and I want to start by questioning how we understand the central component — our thinking.</p><p>An earlier description of how AI could take over the world came from Nick Boström’s “paperclip maximiser”, a machine whose task is to produce as many paperclips as possible. The risk in giving a machine such a clearly defined task is that it may reason that humans are somehow getting in the way of this overarching goal and decide to do away with humanity. This reasoning contains a view of thinking and information processing that is based on formal and rule-based logic, like an overzealous bureaucrat in an authority who follows his directives to the letter without any other considerations such as the well-being of the organisation, society or individuals. There are always unwanted side effects we tend to forget when we try to formulate such goals. This was probably also a reasonable description for most digital systems, at least before the breakthrough of neural networks and large language models (LLMs) like Chat GPT. But here are two fundamentally different ways of designing digital systems, such as an AI that we can illustrate by describing the development of chess computers.</p><p><strong>A brief introduction to chess computers</strong></p><p>Traditionally, and still today, chess computers such as Stockfish have been built on formal and rule-based logic. It works according to two steps that human chess players also use. The first step involves calculating possible moves and variations ahead. Here, of course, computational performance is an essential asset, where the computer follows different variations or branches of sequences of possible moves where the number of combinations grows exponentially the more moves forward you try to calculate. The second step involves being able to value a position in terms of how good and desirable it is for oneself. In the classical programs, this evaluation is done according to various criteria such as material (how many pieces you have more than your opponent according to a scoring heuristic: a pawn=1p, a bishop and knight=3p, a rook=5p, a queen=9p), the pawn structure (doubled on the same line or isolated is often bad), terrain or space advantage, piece activity and how protected one’s king is. All of this is added up to a total score that makes up the evaluation of the position. What is important here is that all the steps are explicit, you can follow how the computer (or engine as it is often illustratively referred to) reasons in which sequence of moves and exactly how it evaluates the position. This has been very successful and since Deep Blue defeated the then-world champion Garry Kasparov in 1997, computers are considered superior to humans in chess.</p><p>But in recent years, a new way of programming chess engines has been developed based on deep learning. First, it was done with the game Go, where AlphaGo defeated the undisputed human champion, Lee Sedol, in 2016. Then the chess computer AlphaZero was developed, which was trained according to similar principles by playing against itself many times to train itself. Although we build the education system pretty much according to the more logical functioning above, this way is more reminiscent of when children learn to walk and explore the world. Our young children do not receive an instruction book with rules that they practice according to but learn by trial and error. They get up, stumble, fall, get up again, etc. AlphaGo and AlphaZero use neural networks and a big difference compared to traditional chess engines is that the programmer or bystander cannot see how the computer is reasoning. A neural network consists of a lot of layers with different nodes that send information to each other, which results in a lot of weights (typically numbers between 0 and 1) that show how strong the connection between two nodes is. But it is very difficult, if not impossible, to see how these weights result in how it reasons on the chessboard. It is not transparent in that way but more of a black box. It can make us uncomfortable, and even if AI is increasingly applied in various areas and industries, there is some pushback in, for example, healthcare, to recognizing cancer at an early stage based on images from screening, or in the construction industry, because you typically don’t just want good and accurate answers or estimations but also want to understand <em>how</em> it reasons. In contrast, if you are a stockbroker, you might not care that much as long as it increases your profit.</p><p>Something that was also observed when analysing AlphaGo and AlphaZero playing Go and chess was that the playing style differs from the traditional engines. They have been described as more intuitive and creative, more human-like. Many experienced chess players, such as grandmasters, also have a more intuitive understanding of the game and can evaluate situations a bit differently than traditional engines that use more brute force. Skilled chess players can sacrifice pieces to achieve long-term strategic advantages that are not easy to conceptualise. Based on their vast experience, this intuition has a parallel in that the new chess computers are not rule-based but rather experience-based means of machine learning. This has made many fascinated and some concerned. But why?</p><p><strong>Thinking fast and slow</strong></p><p>We return to our model around our psychological functioning with thinking at the centre. Can we see similar reasoning between how chess computers work and how we think ourselves? Yes, here I think there is a clear parallel with Thinking fast and slow by Nobel laureate Daniel Kahneman. In accordance with our model above, it is often thought in behavioural economics that we are rational and profit-maximizing and make decisions accordingly. Kahneman problematized this and the title of the book shows two different systems we apply when we think and make decisions.</p><p>System 1 is our instinctive, emotional and fast system. We typically use it when we judge other people’s facial expressions, evaluate distances in traffic and make associations in puns. (Answer quickly: Who starts in chess? What colour is the snow? What does the cow drink?)</p><p>System 2 is the logical and slower system that we apply when we calculate 36 × 17 by hand, evaluate logical reasoning and carefully prepare for a race.</p><p>In experiment after experiment, Kahneman demonstrated that in many cases, system 1 was involved when we made e.g. financial decisions and in how we evaluate risks. We often use both, for example, if we are going to buy a house, we might compare price, location, size, condition, etc., with a list of our wishes and budget. But then we go to the viewing and can feel “This is where we should live!” or “ Njaee…”. We are not as logical and rational as we (and many former behavioural economists) often think.</p><p>If we now compare this development in AI, the traditional programs and chess engines have a lot in common with our system 2. That is often what we thought we should have the computers for, to offer a lot of computing power and follow a logic we assign. This is applied in the field of structural mechanics, where finite element calculations are performed in the construction or automotive industry where you make computer models of houses, cars or products and then, according to mathematically derived principles, calculate whether they break and how they deform under a given load. Those of us who teach this have always argued that there is always a need for a person who understands the principles of structural mechanics and the computer programs at a fundamental level so that one can evaluate the calculations give reasonable results and thus take responsibility for the results. If you calculate incorrectly and the house collapses, you, as a structural engineer, may be prosecuted.</p><p>As for the new AI models, like LLMs such as ChatGPT that use neural networks, they are closer to system 1, i.e. the fast, emotional and instinctive system (although it’s hardly emotional). For a long time in the AI development, the models were fed more and more training data without anything significant happening, but then a threshold was reached (they have now used the entire open internet as training data), whereupon the language models suddenly started to give off sensible, nuanced and, seemingly empathic responses to complex questions such as moral dilemmas, adapted his style and wrote poems according to instructions. They can also be wrong and hallucinate, as many do when they answer above that the cow drinks milk.</p><p>I think we have been fascinated and troubled by this for two reasons. The first is that before machine learning and big language models, LLMs, we saw computers as something that complemented us, they were logical according to system 2 and were superior to us in that sense, while we could be more intuitive or instinctive, emotionally evaluating and creative according to system 1. But now, the AI systems seem to surpass our abilities in that sense too. What, then, will be our role when the computers are better at us than anything else? Will it outcompete certain professions, something we are already seeing signs of when AI bots replace customer support workers? Here one can compare the AI revolution to the mechanization of agriculture or the automation of the manufacturing industry. And is there any risk of the computers taking over? This is sometimes referred to as the “AI-alignment problem”, i.e. how we can ensure that the interests of the AI will be in line with our interests. Another variant of that question is the extent to which a lower intelligence (we) can control a higher intelligence (the AI).</p><p>However, we should be aware here that Nick Boström’s paperclip-making machine is based on a computer or AI that functions according to system 2, i.e. the formal rule-based logic, and does not seem reasonable for the neural networks and LLMs that perform well when weighing different perspectives against each other. When I test ChatGPT on structural mechanics problems, it reasons correctly overall but gets a lot of details wrong, even though GPT4 is better than GPT3. It can follow a logic, but above all, it is nuanced and easily weighs different perspectives and points of view against each other. It is not governed by an overarching logic. Here, I sense that the two perspectives are mixed up in the discussions — on the one hand, AI is nuanced and creative (mainly generative AI) according to system 1, on the other hand, it is so blindly governed by rules according to system 2 that it will get us into trouble for that reason.</p><p><strong>Emergence</strong></p><p>The second reason I think we become fascinated and concerned is emergence. Put simply, it means how more complex entities are produced by combining simpler building blocks. These new entities usually have properties that cannot be predicted or derived from the properties of the simpler building blocks. For example, water has properties such as a certain density, surface tension, freezing and boiling point that cannot be predicted based on the properties of the components hydrogen or oxygen. Within complexity theory, we can model (and observe) complex behaviours of large flocks of starlings even though each starling follows simple rules of the game in terms of distance to the starling next to it.</p><p>Another example of emergence is the widespread belief that our evolution can be described in terms of matter emerging into life and then emerging into consciousness. This is a common form of materialism seen today, which means that even if you consider consciousness as an aspect of reality, matter is seen as more fundamental. Consciousness is considered from that understanding as an epiphenomenon to matter. Emergence then becomes a very powerful principle or phenomenon that explains how life and consciousness arise. A limitation of this way of understanding reality is that we have difficulty figuring out <em>how </em>consciousness arose from life and matter, as well as how consciousness now relates to matter, what we call the hard problem of consciousness. More of that in another post.</p><p>What is often found worrying about this view of emergence is that we cannot control or predict what characteristics the AI will acquire, for example when it is used to develop itself. So far, LLMs are just an advanced autocorrect that guesses likely sensible answers to a given prompt, but since it is a black box, we cannot know for sure what is happening on the inside apart from what we can observe on the outside. Something that is sometimes discussed here is when the AI can develop self-awareness or own agency.</p><p>I think we can see examples of emergence in the LLMs, who, with a sufficient (enormous) amount of training data, suddenly began to show evidence of a response that many had difficulty distinguishing from humans. Remember how a Google engineer claimed their chatbot was self-aware two years ago? There are also many examples of people creating closer contact with AI than with their fellow humans. The film Her (2013) is a reality today.</p><p>According to our reasoning above, can an AI that becomes sufficiently complex and intelligent in its information processing become self-aware and develop its own will? After all, we can never know for sure what happens inside an AI, just as little as we know for sure what is inside our fellow humans. They may all be zombies and you reading this may be the only truly self-aware being (solipsism), something we can never disprove. And what does self-awareness really mean? Now we will move on to adult development and reason about how we can understand intelligence and development in general.</p><p><strong>The different branches of adult development</strong></p><p>Within adult development, we study how people, individuals and, to some extent, collectives develop. The area of adult development I have worked with the most is stage theories because I think they give good answers about where we come from, where we could be going, and how that affects how we function today.</p><p>A common division of adult development theories is into two different branches, where the first can be further divided into two. The first branch focuses on complex thinking with, on the one hand, stage theories such as the Model of Hierarchical Complexity (MHC) and Lectical Assessment System (LAS), and on the other side of dialectical thinking. In MHC and LAS, hierarchical stages of complexity have been formulated, which can be understood as increasing levels of abstraction in thinking, from simpler to concrete, abstract, formal, systematic, meta-systematic and so on to even more complex. Dialectical thinking is not stage-based but describes an increased ability for fluidity in thinking and the ability to think more organically in terms of process, structure, relations and systems that are transformed. If we compare the two fields with Kahneman’s two systems, I argue that the stage theories are close to system 2, which can be seen when you test which stage someone operates from through logic tests at different levels that you either pass or fail. Dialectical thinking is closer to system 1 because it is not about building logical reasoning but being more flexible in one’s thinking. It is evaluated in interviews, which are assessed according to how many and deep aspects of the various elements are presented in their reasoning.</p><p>Although it is rarely made explicit, I think that Figure 1, which we used before, also works to describe how we function according to the MHC. It is a behaviourist theory where it is assumed that human behaviour can be understood as problem-solving at different levels of complexity. You don’t speculate much about what happens inside the person; you keep it as a black box. We judge what we see, the performance of solving a given task or producing something like a text. This branch of adult development, as I see it, thus does not provide any further clarity on the question of possible self-awareness and agency in an AI.</p><p>The second branch deals with ego development according to Jane Loevinger and Susanne Cook-Greuter and meaning making according to Robert Kegan’s subject-object theory. Here, development is measured using sentence completion tests or interviews. These are more holistic and include different aspects of our inner world and psyche, how we take perspective and relate to others and our inner self. Criticism is often directed between these two branches or camps. From the first camp, it is often heard that ego development and meaning are vaguely defined concepts and that everything can really be reduced to complex reasoning or behaviour, where the development of behaviour can be measured with greater precision. From the other camp, you can hear that the first miss the point and reduce something complex to something too narrow where the baby (the subject) is thrown out with the bathwater. Personally, I have worked mostly with MHC in my research, but I think that ego development offers a more relevant measure of the development that I think is relevant here. Being a complex problem solver is not necessarily a good measure of personal development and maturity. Figure 2 below illustrates how we interact with our world through our way of making meaning rather than just thinking. Although I think it is more accurate to see us as meaning making beings than to reduce us to only thinking or problem-solving beings, it does not help us much in the discussion of whether self-awareness can emerge from thinking if we cannot further explain what meaning making is.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/222/1*_mgVY7HDtNCsueD8mpLl9g.png" /><figcaption>Figure 2. Relationship between reality and meaning making. The reality around which we create meaning can also be part of our interior.</figcaption></figure><p>There is also a middle way, which was initiated mainly by Tom Hagström, which involves identifying fundamental developmental aspects of meaning making based on Kegan’s subject-object theory and MHC. I believe this path to be most fruitful in the issues we are discussing here, and this is also a route I take in approaching inner development, although here with ego development theory, according to Cook-Greuter, as a point of departure. If one assumes from an MHC perspective that ego development or meaning making can be reduced to complex thinking, then it implies that complex thinking constitutes the most basic aspect of development. This path implies that we need four basic dimensions to capture ego development, namely perspective taking, complex thinking, identity and action logic as illustrated in Figure 3 below. This is part of the framework that I call CDA, collective development archetypes, which I am developing and which is currently in review in a scientific journal.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/266/1*tCINfvZbh5KDGfamtTsN5A.png" /><figcaption>Figure 3. Meaning making divided into the four aspects perspective taking, complex thinking, identity and action logic, which are shown here as isolated parts although they may be interwoven. Reality can be the physical world but can also include our inner self and our relationships.</figcaption></figure><p>Perspective taking means how we take in information and if we look at Figure 1, which illustrates us as thinking beings with the ideal that we take in information from our world objectively, then the vast majority of social scientists would see this as problematic — and so would I. When someone claims to be objective, to me, it rather means that you see an object or phenomenon from a 3rd person perspective. But I and others in ego development believe that there are higher order perspective taking, 4th and 5th person perspectives etc. In my theory development (another article in review), I define perspective taking as a relationship between an object and a subject — and there is always a subject that is constituted in a certain way. This form of perspective awareness is included in the higher orders than the 3rd person perspective. We also need to mention 1st and 2nd person perspective, which means taking in information about the surrounding concrete and physical surroundings by means of one’s senses, respectively seeing the concrete world from someone else’s eyes, something that can be tested with the Sally-Anne test.</p><p>Of these aspects of development, complex thinking is fairly well researched, while the others do not have as elaborate theories attached to them within adult development research. Identity/being describes the emotional or affective aspects of meaning making and this is easiest to see as intertwined with the other aspects. We tend to identify with how we take perspective and with our subject. Many, for instance, Jean Piaget, considered thought to be intertwined with emotions, and an action logic usually has an emotional component that drives us forward.</p><p>Action logic, as Torbert called the ego development stages in his Action Inquiry, describes not only how we think about the world but also how we plan and carry out actions. Perspective taking and thinking can also be seen as actions. The gap between understanding and action in climate research illustrates that they are not necessarily connected. The action logic describes a logic for acting but also an agent and driving force or own motivation to act.</p><p>All of these can be considered different components or facets of meaning making. They can, as I see it, be understood by themselves or as deeply intertwined. This is a very brief introduction, and there remains work to elaborate and refine the theory around these, something I have been involved in in recent years. The adult development perspective also describes how these different dimensions interact in our development when we have gone from more physically oriented beings to more mentally or relationally oriented ones.</p><p>The most important point here is that these four aspects of development are fundamental and basic, which means that none of them are epiphenomena that can be reduced to any other. For example, we cannot reduce perspective taking — how we take in information — to complex thinking — how we then organise the information. Our action logic may then connect with our complex thinking but does not necessarily do so. And our being, our feelings and identity, typically in relation to other beings, cannot be reduced to how we organise information either. Our ability to take perspective is not developed from our complex thinking.</p><p><strong>Back to the questions</strong></p><p>But now, we will return to the questions we started with and tackle them one by one based on the view presented in Figure 3.</p><p>What do AI and AGI mean in relation to adult development theories? If we consider AI in the form of machine learning, neural networks and LLMs, the models are just ways to encode and organise information. It corresponds to our complex thinking, but it is not an action logic or its own form of agency. Although it can be generative, we are the ones who give it a prompt at which the AI answers.</p><p>AGI is usually defined as an ability to solve a broad and general set of problems that is comparable to or superior to that of humans, as opposed to narrow AI that specialises in a certain kind of problem, such as playing chess. However, both cases are based on the view that intelligence is equated with complex thinking and problem-solving. Both of these can be illustrated in Figure 1, which only shows thinking in relation to the outside world.</p><p>AI, like LLMs, is thus exclusively about complex thinking, but doesn’t it have some perspective taking ability? This is an interesting question and one of the first things I did with Chat GPT 3 was to test perspective taking according to the Sally-Anne test in text, which it actually passed. Children around 4–5 years of age normally pass it, while children with autism typically struggle with it. There was a similar test developed by Robert Selman and colleagues that children usually pass around the age of 10–12 that I estimated it to pass as well. But I’m still not convinced that it really takes perspective or that it just organises information in ways consistent with passing the test.</p><p>Can AI and LLMs, if they become sufficiently complex in their thinking, develop self-awareness as in subjective experience and meaning making? Based on the framework and reasoning above — no. Perspective taking, a sense of identity or subjectivity, and action logic cannot emerge from complex thinking because they all constitute different fundamental and basic aspects of development, none of which can be reduced to any other.</p><p>What about the AI alignment problem? How can we ensure that AI’s interests will be aligned with our own? Again, complex thinking and information organisation is one aspect and having a self-interest and agency is another. So far, AI only gets the ability to act based on the information and action possibilities we give it. Agency and self-interest will not emerge from a sufficiently complex organisation of information. Today, AI systems are primarily capable of solving abstract tasks such as playing chess, while seemingly simpler motor tasks, such as physically moving a chess piece from one square to another, have proven significantly more difficult.</p><p>Will AI evolve past us to control us rather than us controlling it? Based on the reasoning above, I think meaning making is a much better measure of development and more relevant in this case. And AI does not engage in meaning making as everything is missing there except complex thinking and organisation of information. So, I don’t see how that would happen. But there are many different ways in which we become dependent on AI as a technology and forget our own ability to solve the corresponding tasks, just as we are dependent on most other technology today.</p><p>Another and, in my opinion, more interesting question is how intertwined we will become with AI in the future. If we consider AI in a broad sense as a technology that can become an extension of ourselves, much like our language, then the boundary can become diffuse. And even if you usually think of this as a tool for more complex thinking, language and information can be designed in this way so that we can access the other aspects of our being. Also interesting is how we create a relationship with a generative AI by projecting on it a subjectivity that is not there.</p><p>There is much more to say about AI in relation to our intelligence and development and the presentation on how AI works here is simplified in many ways. What I want to do here is primarily to connect these questions with adult development perspectives and show how our different metaphysical assumptions about our consciousness and thinking are relevant to what we expect about the future of AI. My view is that we humans are more complex than many in the AI field describe or assume. It is something I think we should discuss to a much greater extent.</p><p><strong>Useful sources</strong></p><p>Black box in The Guardian, episode 1 is about Geoffrey Hinton, who just received the Nobel Prize in Physics: <a href="https://www.theguardian.com/technology/series/blackbox">https://www.theguardian.com/technology/series/blackbox</a></p><p>Matthew Sandler Gamechanger on how AlphaZero plays chess <a href="https://www.youtube.com/@SiliconRoadChess">https://www.youtube.com/@SiliconRoadChess</a></p><p>The Google researcher who thought the chatbot was sentient: <a href="https://www.theguardian.com/technology/2022/jul/23/google-fires-software-engineer-who-claims-ai-chatbot-is-sentient">https://www.theguardian.com/technology/2022/jul/23/google-fires-software-engineer-who-claims-ai-chatbot-is-sentient</a></p><p>Hagström Stålne — The Generality of Adult Development Stages and Transformations: Comparing Meaning-making and Logical Reasoning: <a href="http://www.integral-review.org/documents/Hagstrom%20Stalne,%20Vol%2011%20No%203%20Special%20Issue%20European%20Acad%20Trends.pdf">http://www.integral-review.org/documents/Hagstrom%20Stalne,%20Vol%2011%20No%203%20Special%20Issue %20European%20Acad%20Trends.pdf</a></p><p>In his book The Developing Human Mind, Tom Hagström has a section on AI that is roughly in line with my reasoning: <a href="https://link.springer.com/book/10.1007/978-3-031-28647-6">https://link.springer.com/book/10.1007/978-3-031-28647-6</a></p><p>The Sally-Anne test: <a href="https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test">https://en.wikipedia.org/wiki/Sally%E2%80%93Anne_test</a></p><p>Her (2013) <a href="https://www.youtube.com/watch?v=dJTU48_yghs">https://www.youtube.com/watch?v=dJTU48_yghs</a></p><p>Ex Machina (2014) <a href="https://www.youtube.com/watch?v=EoQuVnKhxaM">https://www.youtube.com/watch?v=EoQuVnKhxaM</a></p><p>AlphaGo — The movie: <a href="https://www.youtube.com/watch?v=WXuK6gekU1Y">https://www.youtube.com/watch?v=WXuK6gekU1Y</a></p><p>Daniel Kahneman — Thinking, Fast and Slow <a href="https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow">https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow</a></p><p>The Routledge International Handbook of Dialectical Thinking <a href="https://www.routledge.com/The-Routledge-International-Handbook-of-Dialectical-Thinking/Shannon-Mascolo-Belolutskaya/p/book/9781032324678?srsltid=AfmBOoojq8ALG9z3lr1ygCfq7MVw9kHXf46lID2h1uBMYcNqBlR2erAN">https://www.routledge.com/The-Routledge-International-Handbook-of-Dialectical-Thinking/Shannon-Mascolo-Belolutskaya/p/book/9781032324678?srsltid=AfmBOoojq8ALG9z3lr1ygCfq7MVw9kHXf46lID2h1uBMYcNqBlR2erAN</a></p><p>Yann LeCun with Lex Fridman: <a href="https://www.youtube.com/watch?v=5t1vTLU7s40">https://www.youtube.com/watch?v=5t1vTLU7s40</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2ad4cca5a398" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[emerge vs Integral and beyond]]></title>
            <link>https://medium.com/@kristianstalne/emerge-vs-integral-and-beyond-651132bcba65?source=rss-792e23cb7029------2</link>
            <guid isPermaLink="false">https://medium.com/p/651132bcba65</guid>
            <category><![CDATA[integral-theory]]></category>
            <category><![CDATA[emerge]]></category>
            <category><![CDATA[integral]]></category>
            <category><![CDATA[shadow]]></category>
            <dc:creator><![CDATA[Kristian Stålne]]></dc:creator>
            <pubDate>Wed, 09 Feb 2022 12:10:07 GMT</pubDate>
            <atom:updated>2022-02-09T12:10:07.177Z</atom:updated>
            <content:encoded><![CDATA[<p>On a sunny Saturday this autumn, I was sitting on my chair packed into a circle with seven other human beings in the cellar of the bUm in Kreutzberg and discussing the view on the individual in the metamodern era. It was a bit emotionally overwhelming after 18 months of physical distancing but at the same time a very familiar setting.</p><p>The session was part of the <a href="https://www.whatisemerging.com/gathering"><em>emerge</em> gathering</a> in Berlin, where we engaged in discussions around the broader question on promises or expectations and shadows around the metamodern project. The workshop was hosted by Johan Ranefors, focusing on the role of the organisation in the metamodern project, Ellie Hain, outlining different imaginaries emerging throughout history, and Alexander Bard, promoting the “dividual” that he borrowed from Deleuze as a way forward from the limiting view on humans as mare individuals. Or does this undermine the individual as one of the central pillars of the modern society resulting from the enlightenment and breaking free from authoritarian dogma? Do the dividual transcend and include the individual?</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nWKDGNk7a6bq56jE.png" /><figcaption><em>Photo: Tomas Björkman presenting at the emerge gathering.</em></figcaption></figure><p>This discussion around shadows triggered some reflections on the difference between the integral community where I have some experiences and everything around the <em>emerge</em> gathering and realm. Although these two worlds differ in many aspects that I will address, there is a significant overlap in terms of theories and outlook, such as</p><ul><li>aiming for building a space beyond postmodern deconstructive critique,</li><li>embracing a transformational attitude,</li><li>involving developmental perspectives informed by adult development theories,</li><li>integrating a multitude of strands of culture and life, and</li><li>aiming for bridging science and spirituality.</li></ul><p>There is also a significant overlap in the participants where I recognised and connected with several individuals from the integral scene.</p><p>Let’s return to the shadows, but first some differences between the two worlds.</p><p>Integral, its worldview, culture and movement, was characterised by and centred around a single person, a single idea, and a single framework. The person was, of course, Ken Wilber, the main idea was development or evolution by transcending and including previous levels, and the framework for describing (and prescribing) this was AQAL. Although there was a trajectory of criticism and bringing forth different alternatives and theorists, the integral approach can thus be described as quite monolithic, which had its advantages and drawbacks. One advantage was that it was easier to have a common language and frame of reference to unite around — or to critique. A clear view, definition, and perspective around what development means facilitates your own development in many ways, but even more so the development of the community. You can build the Tower of Babel higher when everyone speaks the same language.</p><p>In contrast, the <em>emerge</em> context is more characterised as a network or an eco-system of many actors, ideas, frameworks, and initiatives. The very notion of <em>emerge</em> implies that no single theory, framework, or perspective can predict what may arise from combining these different perspectives and approaches. It is more appealing and inviting for many to come to a world that is not pre-defined or a set table, but where you can co-create everything, including the very frames and definitions of what <em>emerge</em> or metamodern is. When Integral was more of an answer that looked for questions and applications in different areas, <em>emerge</em> is more of an open and living question without a fixed answer and closer to the actual challenges around sustainability.</p><p>This more open attitude may appear more complex or mature. Still, in terms of the stagewise perspectives such as those building on hierarchical complexity, it is acknowledged that you reach higher when you have support rather than having to invent everything by yourself. This represents, however, a slightly more restricted view on what development is and how it should unfold. This stagewise view does not contain emergence more than in a pedagogical sense, that of developing uniquely but reaching yet another pre-defined and general level or stage. And even though the <em>emerge</em> movement wants to see itself as the leading edge of cultural expression in the western world, it is probably considered a step back when assimilated into an integral perspective. But it all depends on what you mean by complexity and by “reaching” a certain level in terms of understanding it — or carrying it as your own insight.</p><p>A more open view on development and emergence is represented by e.g. the branch of complex thinking that adheres to dialectics where emergence is both a thought form and a core aspect. This also concurs with my experience from transdisciplinary research. Although a single person may reach higher in complexity than a group having to find a common language, negotiate and compromise the different perspectives into a coherent framework, a collective approach may be easier disseminated and have a broader impact. The more you involve those affected by the perspective, framework or approach, the likelier it is that they will engage in it.</p><p>In my understanding, another difference between <em>emerge</em> and Integral was that spirituality was more at the centre of attention in the Integral world. The Integral view on spirituality was more radical in its approach and many key players were connected to spiritual practices and communities. Evolution was fundamentally seen as a spiritual endeavour. To be fully enlightened, you need to develop vertically (stages) as well as horizontally (states), which is the mission of existence.</p><p>Integral went further in terms of how one should relate to the world. Integral was not only a set of ideas or a community. It was also a worldview, an operating system, a way of making meaning, acting and identifying. It was more than only using the terminology and the integral concepts, but also <em>being</em> Integral? Not only watching the river but immersing oneself completely.</p><p>Integral was more exclusive in terms of the higher pre-requisites of the participants. Although everyone was welcomed, Integral was doubtlessly more elitist as it implicitly placed higher demands on the participants rather than being inviting and inclusive (that’s sooo postmodern). Much attention was given to analysing other frameworks, groups and individuals in terms of to which extent they qualified as being integral or not. This was also the case for the internal discussions, to ensure that they held an integral standard in complexity and embodiment. Ideally, an integral life practice was not something you only did in your spare time, or even integrated with parts of your life, but of something your entire life should be an expression.</p><p>There was, not surprisingly, a flipside to it. Or rather, the dark side — the shadows. So, what does shadow mean here? In the integral context, the shadows are described as disowned parts of ourselves that we tend to project on our neighbours. In Freud’s terms, which Wilber quotes, the goal is to integrate the shadow to make it part of the self — “Wo Es war, soll Ich werden.” In a broader sense, the shadow can be that blind spot you don’t know you are missing that comes back to haunt you when you least expect it. Shadows can also be applied to collectives, such as entire cultures as studied by Thomas Hübl, and movements as if they had a personality.</p><p>Although Integral’s core, ambition, and very meaning were to integrate all available perspectives, and although shadow-work was valued as a core practice, we are always only human. Being more clearly defined and first on the integral scene, the Integral world struggled with some apparent shadows. Here I don’t mainly refer to its exclusiveness in terms of “being Integral or not” and claims of being all-encompassing, both of which were intentional ambitions. But rather, in the abuse of integrally acclaimed and endorsed gurus, I won’t describe in detail here. With this very clear idea of what development is and a strong sense of purpose and urgency, some believed that no sacrifice was too big in the name of evolution. Some just got lost in the space beyond conventions.</p><p>A core shadow in my eyes was a consequence of the notion of the “transcend and include” view of development. This means that a new stage of development includes all previous stages with all their aspects and qualities. You don’t develop into something else, but into something more. Although this principle may be accurate in some areas such as mathematics (if you know how to perform integral calculus, then by definition you can add), it doesn’t hold up that well for most other areas such as understanding what a human being or a culture is, what it means to have a healthy relationship with the earth or recognise different berries. This led to some overestimating one’s own abilities among integral practitioners.</p><p>Back to the promises, expectations and shadows of <em>emerge</em>. Should we expect any shadows? Always, yes. We can and will likely mess things up. Hopefully, we’ll learn from it and take the consequences. But in comparison with Integral, I see three reasons that the shadows will likely be milder:</p><ul><li><em>emerge</em> is more diffuse and ambiguous in its multitude of theories, actors and issues. Unambiguity and light cast darker shadows. But here, there is no apparent telos or higher purpose and no single framework to get attached to.</li><li><em>emerge</em> have emerged a couple of decades after Integral and can hopefully learn from previous mistakes. Raising the question as Johan, Ellie and Alexander did is a good preventive initiative.</li><li>In the <em>emerge</em> context, there is more emphasis on complex challenges around sustainability. The situation is direr now than 20 years ago. We are now more aware that <em>complex</em> means that challenges can’t be solved, only addressed. In the Integral context, there was a common implicit assumption that we can solve everything if we only evolve high and fast enough.</li></ul><p>There will, however, be shadows here as well; rest assured. One is that the notion of emergence might be taken too seriously. That solutions to all problems will magically emerge if we only bring our different perspectives and good intentions. A good assumption is that we are always in a bubble, however inclusive we try to be.</p><p>Also posted on my blog: <a href="http://fication.se/?p=908">http://fication.se/?p=908</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=651132bcba65" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>