WARNING: Summaries are generated by a large language model and may be inaccurate. We suggest that you use the synopsis, short and long summaries only as a loose guide to the topics discussed. The model may attribute to the speaker or other participants views that they do not in fact hold. It may also attribute to the speaker views expressed by other participants, or vice versa. The raw transcript (see the bottom of the page) is likely to be a more accurate representation of the seminar content, except for errors at the level of individual words during transcription.


Technological advances have enabled us to explore the world in new ways, from our smartphones to complex infrastructure projects. We must consider both the big and small implications of disruptive technology, such as the rise of Artificial Intelligence, automation, and rewilding. These changes can lead to improved product quality and more efficient land use, but also ethical questions and a decrease in confidently stated facts. We must also consider our own limitations, such as fatigue and attention to detail, and use technology to our advantage.

Short Summary

The transcript discusses the potential implications of disruptive technology and suggests that although the big changes are more obvious, it is important to consider both the big and small changes that it may bring. It gives examples of how the internet has changed communication, and how the sudden transformation of mobile phones in a less developed country was due to both access to the phones and the infrastructure built out. Artificial intelligence and machine labor are expected to exceed human limitations, and even small differences in precision can make a huge difference in product quality. It is suggested that the cost of precision in terms of attention to detail can have a big impact on the quality of products.
Humans may struggle to attend to detail consistently, leading to costly errors and mistakes in our work. Machines, however, can achieve consistent precision and attention to detail, potentially eliminating these failures. This suggests that we may be externalizing our limitations and projecting them onto the world. This could lead to a world with a higher quality of products, but still the same costs. AI and reinforcement learning can be used to automate industrial production lines, allowing robots to handle fabric and automate the production of clothing. It is still uncertain if attention will be given to complex tasks, as this has been a limitation of humans for many years.
The speaker discusses Artificial General Intelligence's potential to perform complex tasks quickly and without fatigue. Examples of AI being used to automate infrastructure maintenance are discussed, as well as the cognitive resources required to enable AGI. Human attention and fatigue are discussed, as well as Ritalin's ability to shortcut the task switching process. Finally, the development of AI-controlled land-moving equipment is discussed, as it shows how the physical world can become more fluid in response to our limitations.
The transcript discusses how different rates of change create different experiences and perceptions. Smartphones have enabled people to fact check conversations, leading to a decrease in confidently stated facts. This can lead to people making unfalsifiable claims as a way to avoid being called out. Online conversations have become increasingly toxic, with people attacking each other's intentions rather than facts. GPT bots could be used to create a space for people to explore ideas without the influence of established opinion. The speaker suggests that it is important to have a space for naive conversations, and to give people the benefit of the doubt.
The transcript talks about how the world has become more intelligible thanks to advances in science and technology, but that this is still not enough to understand all aspects of the world and how it works. It is suggested that many people have a limited understanding of things such as their smartphones, government, and food production. Furthermore, the emergence of conspiracy theories is a form of psychological disorder, and land is not actually scarce, but our perception of it is due to our tendency to congregate in a small proportion of the land area. Finally, it is suggested that technological advances such as autonomous aircraft will continue to blur the lines between rural and urban living.
The transcript discusses the technology required to travel in and around land, and how assumptions about the land may change in the future. It is suggested that humans could explore the third dimension more and build infrastructure such as underground transportation tunnels, although this is expensive and difficult with current technology. It also discusses the three main categories of environmental land use: conservation, preservation, and rewilding. Rewilding is a complex ethical question as it involves deciding which ecosystem structure is best, and technology is making these questions more than just philosophical thought exercises.

Long Summary

The transcript discusses the potential changes and implications of disruptive technology, and how it is difficult to predict the exact consequences of such technology. It is suggested that although the big changes are more obvious, it is also important to consider the little changes that can occur, such as those caused by the internet. The speaker gives examples of big changes that could occur, such as transportation being completely automated and energy being clean and inexpensive. It is concluded that the implications of disruptive technology tend to surprise us and that it is important to consider both the big and small changes that it may bring.
The internet has changed many aspects of life, including how people communicate and interact. In the past, people had to plan things in advance and be reliable in order to meet up with friends, as there was no easy way to get in touch with anyone. With the introduction of mobile phones, people could easily change plans or communicate at the last minute, making it a social norm. This allowed for more flexibility and spontaneity in social interactions.
The speaker talks about the sudden transformation of mobile phones in a less developed country, and how it was due to both access to the phones and the infrastructure built out. They then move on to discuss how artificial intelligence and machine labor will exceed human limitations, and ponder a method of examining the small implications of this. As an example, the speaker looks around their desk and mentions the various widgets, electronics, scissors and stapler they can see, concluding that artificial intelligence will have small implications on everyday items.
The transcript discusses the correlation between the cost of products and their quality. It uses examples such as staples, coffee cup lids, pens, and flashlights to illustrate how even small differences in precision can make a huge difference in how well a product works. The speaker wonders why it is so hard to make a product of higher quality, when it would seem to be such a trivial task. The speaker suggests that the cost of the product throughout the supply chain may be a factor in why higher quality products are more expensive.
Humans have a limited capacity for attention to detail, which can affect the quality of products. This is particularly evident in Japanese culture, where even small objects are given careful attention. Attention to detail is time consuming and cognitively exhausting, often requiring more effort than physically. This is especially true for tasks that are boring, frustrating, or repetitive. The cost of precision in terms of attention to detail can have a big impact on the quality of products.
Humans experience difficulty attending to detail, resulting in countless failures in our work. Machines, however, can achieve consistent precision and attention to detail, potentially eliminating these failures. This suggests that human limitations may be externalized and taken as properties of the world, rather than consequences of our shortcomings. This could lead to a world where everything is built with meticulousness and attention to detail, unlike the cheap electronics we are used to.
Humans have limitations that can lead to us projecting properties onto the world that are actually functions of our flaws and limitations. An example of this is the way we thought about the world before fast travel, where unknown areas were thought to be inhabited by dragons or UFOs. Once these areas were explored, we were surprised to find that the world hadn't changed. This is a general framing that can be applied to many examples. Technology such as AI can be a source of surprise when it reveals our limitations.
The speaker is discussing the issue of aftermarket laptop chargers not lasting very long, and speculating on the reasons why this could be. He suggests it could be due to lack of attention to detail in the manufacturing process, or the quality of the materials used. He also notes that even if machines are used to build components, they would still require the same amount of energy to do so. He wonders if this will lead to a world with higher quality products, but still the same costs.
Humans may struggle to scale cognitive energy, but machines can be used to focus more attention on tasks. Spell check is an example of this. In production lines, companies can make two separate investments to improve the quality of production: training and a more robust QA process. Apple's quality may be partly due to intelligence required to design the product, but also better training and QA processes. The easiest thing to scale with capital is the QA process, so Apple may just test more and reject more. Robots can also be used on production lines, such as for spell checking documents or inspecting circuit boards on a conveyor belt.
Humans have traditionally been used to perform complex tasks in industrial production lines, due to their difficulty in quickly performing complex tasks. However, with advances in AI and reinforcement learning, robots may soon be able to handle fabric and automate the production of clothing. It is not clear if AGI (Artificial General Intelligence) would be beneficial in this scenario, as it is uncertain if attention would be given. This may be a limitation of humans, as it has dictated our industrial processes. However, it is not the only way to produce complex objects at large scale and for low cost.
The speaker discusses the potential of Artificial General Intelligence (AGI) to perform complex tasks quickly and without getting tired. He suggests that AGI may be able to perform complex decisions and context switching, although it is not clear how this would work in practice due to the cognitive resources required. He also suggests that AI, such as computer vision, is being used in infrastructure maintenance and suggests that the fundamental elements of intelligence may come back when AGI is built.
Software is being provided to the Australian government for maintaining power lines. This includes taking photos of power boxes and running them through a system to check for further attention. This task is currently done by humans, but drones and cognets are being used to automate the process. This is a disruption in process, as it industrialises the maintenance of infrastructure, allowing for more attention to detail and a production line with automated components.
Humans have a low resolution understanding of brain function and why attention feels subjectively effortful in some circumstances and not in others. The explanation likely lies in energetics, as calories were difficult to obtain in our ancestral environment and brains could only become so large before affecting reproductive success. Evolution has been selective in terms of cognitive function, and it is likely that our brains are energy economizing due to this selection.
Ritalin and other pharmaceuticals can make paying attention easier and require less effort. Part of the theory behind executive function is that there is a lot of task switching and metacognition involved. Taking these drugs can shortcut this process, allowing someone to stay focused on one task without feeling pulled away. People have described this as being like being on Ritalin, where they can suddenly find themselves having spent an hour on one task without realizing it.
Humans have physical and cognitive fatigue which affects their performance, while machines may not experience this. This creates a feedback loop of adaptation for humans as they become more fatigued. This has been taken into account when designing systems, roles and occupations. An example of this is infrastructure development in China, which has been designed around the limitations of machines and humans, creating a form of conservation by default.
The physical world changes slowly due to human limitations, but with the development of AI-controlled land-moving equipment, it is becoming more like the digital world and changing more rapidly. This is due to the ability of machines to be given high-level goals and figure out how to achieve them. This will change how we think about the physical world, as it will become more fluid and less stable, and it will be realised that the unchangingness of the physical world was due to our limitations, rather than an axiom of the physical universe.
The transcript discusses the various rates of change in the world. It explains that some things are effectively permanent, while others are dynamic but unchanging. Plants, animals and humans all move at different speeds. Technology has enabled us to create machinery that works at rates of change much faster than humans can perceive. The transcript also suggests that things which seem permanent or slow to change, could quickly shift into different categories and become more rapid.
People are now less likely to confidently state facts without checking them, as smartphones have made it easy to fact check. This has changed the way conversations work, and can lead to awkward moments if someone is wrong. If there was an expert present to refute an opinion, it would be hard to argue against them if one doesn't have the same knowledge. This has led to a decrease of confidently stated facts, as people are more aware of their accuracy.
The transcript discusses the implications of having easily checkable facts in conversation, and how this could lead people to make unfalsifiable claims in order to avoid being called out. An example of this is the speculation around Elon Musk's behaviour on Twitter; as intentions are unfalsifiable, people can make claims about them without fear of being proven wrong. The conversation also touches on how the presence of a real expert in the room can change the dynamic of the conversation, as people may feel sheepish to express their limited understanding.
The transcript discusses how conversations online have become increasingly toxic, with people attacking each other's intentions rather than facts. It is suggested that this is due to the ease of access to information, and the possibility of being called out by someone with a smartphone or GPT bot. The speaker suggests that the best way to avoid this is to give people the benefit of the doubt, and to avoid making claims about intentions as they are unknowable. This could be compared to a situation in which a fact-checker pulls out their phone to call someone out.
The speaker discusses the introduction of GPT into conversations, and how it could affect social interactions. They suggest that it may be better to have a space to explore ideas without the influence of established opinion, as these opinions may be subtly wrong and incomplete. The speaker suggests that there should be a space to talk nonsense, similar to the proverb of a bar being a place for "small fish" to explore ideas. They conclude that it is important to have the ability to have naive conversations, and consider the implications of introducing GPT into conversations.
The transcript discusses the idea that in the past, people had social institutions that allowed them to be disinhibited and have their mistakes forgotten, but that this has been replaced by social media which records everything and haunts people forever. It is suggested that this is an interregnum, where the world is intelligible and there are no mysterious agents. People can now understand the world beyond their social circles, machines and buildings, and don't have to worry about monsters in the dark forest.
Humans have spread their influence across the planet, creating a world that is designed for us and is largely intelligible. However, if other forms of intelligence were to arise, our sphere of influence might shrink and the products of our intelligence may become unintelligible. This would make the world feel more mysterious and unpredictable, like it did in the past when there were dragons on the map and Golems around the corner.
The speaker agrees that the world has become more intelligible in some ways due to advances in scientific understanding, but points out that many people have no idea how their smartphones work or how their government works, and they don't know how things are made or where their food comes from. This lack of understanding leads to the emergence of many black boxes and misunderstandings. The speaker also notes that there are people at the bar who think they know how the world works, but they don't, and they often have the wrong idea.
The transcript discusses how conspiracy theories are a form of psychological disorder and how assumptions about human limitations lead to the belief that only a small proportion of land is habitable. It is argued that land is not actually scarce as there is plenty of Farmland and desert, but with sufficient energy and resources, it could be used productively. It is suggested that our perception of land scarcity is due to the fact that we tend to congregate in a small proportion of the land area.
The transcript discusses how technological limitations have impacted our perception of land and how human mobility has changed as we have become more mobile. For example, places that were once not habitable due to their distance from city centers have become viable places to live thanks to automobiles. It is suggested that in the future, autonomous aircraft could extend the radius of viability, allowing people to live further away and still have access to city centers in a relatively short time. This would continue the trend that began with automobiles and further blur the lines between rural and urban living.
The transcript discusses the technology required to travel in and around areas of land, and how the land may have been used by Aboriginal people. It also mentions how humans have a two-dimensional conception of land, due to the thin veneer of the Earth's surface. It is difficult to build an intuition of how much material is under our feet, and this is compared to the sci-fi short story, 'Flatland'. It is suggested that assumptions about the land may change in the future, with the example of the moon being green due to millions of people living there in the future.
I recently gained a new appreciation for how much material and effort is required to move earth manually. We may never build deep underground structures, but we could explore the third dimension more and build infrastructure such as underground transportation tunnels. It is expensive and difficult to do this with current technology, but it is not impossible. Building on the surface of the Earth is easier for humans, but it is difficult to figure out what to do with the material brought up from underground.
The ultimate destiny of a technologically advanced civilization is often thought to be a world of urban landscapes and buildings, however this is unimaginative. It is possible to have a world of wilderness with greater productive capacity than we have today, with all production happening underground. Nature is infinite and we must be aware of our capabilities and potential for harm when reshaping the Earth. We must ensure that the cognizance of our capabilities is present, particularly when engineering new species, as this could be a level of power that changes the world.
The transcript discusses the three main categories of environmental land use: conservation, preservation, and rewilding. Conservation is the protection of ecological integrity for human use and benefit, while preservation views the value of an area as intrinsic and not for human use. Rewilding involves active human intervention to determine the meaning of ecological integrity in an area. Conservation and preservation are agnostic about what an area ought to look like, while rewilding requires choices about what the structure of the ecology should be.
Rewilding is a complex ethical question, as it involves deciding which ecosystem structure is best. For example, a piece of land near me has changed drastically over the years - from a climax forest to farmland, to a parking lot. When rewilding, the question is which time period should be used as a reference, and whether extinct species should be brought back. There is no clear answer, but technology is making these questions more than just philosophical thought exercises.
The protagonist of the novel, Abel, is punished by the Val Father for breaking an oath. He is told to put on a magical helm which reveals the world as it truly is and sees the truth of the situation. Abel then cuts his arm and offers his blood to the elf queen, Siri, who drinks it in great gulps. This scene demonstrates the illusion of depth from complexity, as Abel is able to see the truth and make a selfless decision in order to save others.
The transcript discusses the idea that many things in the world appear meaningful and deep to us, but this may be an illusion caused by our limited perceptions. Plato's allegory of the cave is referenced to suggest that reality may actually have more depth and complexity than we perceive. It is suggested that artificial intelligence could potentially give us a 'helmet' to see with far-seeing eyes, and this could change our perception of the world. It is questioned whether we are mature enough to handle this and still recognise the depth that was an illusion before. Mathematics is highlighted as an example of something that appears deep and meaningful, but this could change with the introduction of AI.
The transcript discusses the growth and development that occurs from childhood to adulthood, and how this can be both a disillusioning and enriching process. It also mentions how we idealize certain people such as our parents and significant others, only to discover that they have flaws and imperfections. It suggests that this can be a shock, but that we can learn to love them in a more mature way. Finally, it touches on the idea of heterogeneity versus uniformity, and how flaws and imperfections can create interesting differences.
Artificial intelligence has the potential to create greater variation and diversity in comparison to humans, as its flaws and shortcomings are not as immutable. As a result, differences may still originate from flaws and shortcomings, but they will be more voluntary. This could lead to a greater range of variation and diversity, and it is unlikely that there will be a convergence towards a boring uniformity. Instead, it is likely that artificial life will be vastly variable, with over time the potential for even greater variation and diversity.

Raw Transcript

efp told's lipid hmm that's interesting I don't know how an apple could get up here yeah what was the theme for today Adam um well I was thinking uh there's so many things that that Ai and AGI and then Sapient AGI would change it it it's it's tempting to say well absolutely everything and we can and therefore we can't know anything this is this is a this is a sore temptation to do that I mean and I'm sympathetic that to some extent right I mean the the Singularity metaphor is resonates for a reason I mean so many assumptions that we make about the world today will be undermined or obviated or or otherwise um casting Now by this incredible new technology but to an extent you could say that for every new technology that has any disruptive potential at all I mean it's difficult to say where the the changes that a new technology will cause will stop and and what might remain the same what might be different okay well with all of that in mind um I I thought it might be fun to to say okay it's very difficult all of the possible changes that might occur but it's nevertheless interesting and useful to me trying to think for some of them and yeah it's definitely my team does when we look at A disruption as we try to discern identity which are necessarily obvious uh you you do have to think quite carefully um and it's impossible to to to think of everything even the very largest most profound changes that do occur are not necessarily obvious in advance they're in some sense quite emergent um and so the the consequences the implications of disruptive technological change tend to surprise us they they uh they're seldom easily predictable uh but one of the things that I was thinking about is the little things that change so it's one thing to be to always be focused on the big things like what are the big things that are going to change what are the major shifts what are the what are the the profound Transformations that are likely to result in transportation being completely automated or energy being uh clean as in no carbon footprint and also very inexpensive you know etc etc but but these little changes what about the little things like if and so I I uh I was doing I think it was driving in my car or something like that I was just you know I had a few minutes to myself to just kind of think and I was thinking about just little stuff like what small things did the internet change the internet changed plenty of big things and you know the these very high level Transformations
and impacts yes we can we can all sort of easily identify those especially in retrospect but lots of little things changed with the internet the rise of the internet as well um and uh we mentioned one of those examples in the previous session we said you know oh isn't it funny that that movie plots and television plots used to involve the characters struggling to get to a telephone you know um or to have some piece of information that they they you know they didn't have and that's of course a very very old storytelling device it goes back you know to classical Antiquity that the absence of information um or the difference the asymmetry of information between characters is a great plot device right it's easy to drive plot forward with storytelling if the characters don't know don't all know absolutely everything okay um so I was just in a little example of the you know the the internet changed little things we take for granted like like um in the old days we couldn't it wasn't easy to um get in touch with anybody anyway any time you wanted and so fun a little there were little funny things like how children played was very different you know if you wanted to go and meet your friend somewhere you had to make arrangements in advance and then you had to you know you had to be responsible enough from reliable enough to actually be where you said you were going to be and meet them there um and if you didn't or if not everybody showed up well you just had to roll with it you you know you you you couldn't let life grind to a halt because you weren't able to get in touch with somebody for example I so yeah um little changes that could change sorry go ahead yeah just a small anecdote I remember that transition where I guess it was in high school uh for me where people started getting mobile phones um so it was at some point a social Norm that yeah you made plans beforehand you talked to everybody who was going to be there you arranged a time and then you all just showed up and I I clearly remember the kind of Tipping Point where it shifted to a new equilibrium of you know plans shifting at the last minute being recalibrated and if you it was actually one of the kind of reasons to have a phone I have a mobile was that uh if you if you couldn't participate in that re-evaluation you just kind of it wasn't really an option socially to not be amenable to update in in that way yeah that's really interesting that that I suppose that there was a probably in any given location like any given region
or city or country depending on exactly when the mobile phone uh roll out really really you know really took off and exactly how old you are where you would be in in your life you could easily remember you know probably a one year or maybe a two-year period where that transformation happened for me that uh because partly because where I was living I was living in a less developed country for high school um uh and and partly because uh probably my age um the mobile phone transformation didn't happen until my second year at University I remember that quite vividly um when I went back to the the city that I'd gone to high school in um in Oman um uh you know one year nobody had any mobile phones in the next year everybody had them everybody it was an amazing transformation it was very very sudden and Par some of that was was accessing access to the phones themselves and the affordability of them and I think a large portion of it was the infrastructure the infrastructure had to be built out um and so those two things had to happen simultaneously and then a lot of as you said lots of little changes occurred in people's lives um well one example that I mentioned I'll I'll talk about it here uh to kick off the conversation and I mentioned it to you Dan offline which is um uh with artificial intelligence a lot of human limitations a lot of limitations on um on the labor that we do uh things that translate into a cost in terms of time or effort or you know Financial cost are a function of human limitations and you know artificial intelligence and machine labor will transcend will exceed many of those limitations um in in different ways and it might be fun to think through some of this some of the little implications the little things that that might result in that I don't know how to do how to think through this methodically maybe we can maybe we can focus on trying to come up with a method for examining the small stuff methodically and then uh try to think through a few examples I'm not again I'm not exactly sure how to do this but I'll share one example I thought of which is um as I look around myself even just sitting here at my desk there are lots of little items little plastic widgets and metal widgets and pieces bits and pieces of electronics and there's a pair of scissors and there there are there's a stapler with staples inside it and they're um their various Electronics you know plugs and adapters and that sort of thing um uh there's a coffee cup here with a
lid on it and so forth and some of the products around me are very high quality they're made really well they really work well they perform very well and it's clear that they're made to a very high tolerance you know they're just just the Precision with which they're manufactured is very high and the design that's gone into them was very thoughtful and so as a result they work really well uh but but the things that the the and there wasn't a perfect correlation here but there's a strong correlation between how much these these little widgets around me cost and how well they work and I have a lot of cheap junk around within Arm's Reach as well and the differences between the cheap stuff and the expensive stuff are the physical differences are tiny and I know things that have to be made to a high tolerance if they are to function as machines but it seems to me strange like how hard how much harder is it to make a staple that fits in a stapler and doesn't get stuck and then it is to make a staple that's that's just crappy enough to get stuck and not quite work and I encounter many many things in life like how how hard is it to make a coffee cup lid on a you know a travel coffee mug that doesn't leak that actually just it just works right um does it really quadruple the cost to get the the the Machining and the Precision uh correct there and and so as I and I encourage you to do the same look around yourself or to cast your mind around um and think of the just little things that are kind of like why does this break or why does this thing just not work as well as it could why is a rubbish a cheap pen not as good as an expensive one or a you know um uh uh like a flash I've got a flashlight here a torch as you guys might say um and uh you know that it's it's it does not screw the end the cap that screws down to press the battery contact into place uh is not quite aligned well enough and isn't quite long enough and the spring isn't quite strong enough and the threading on the screw uh is not quite long enough to actually screw it all the way down so that it makes a good contact and doesn't Flicker and so as a result this flashlight flickers a bit unless I absolutely mash it mash and crank down the the end cap and it would be such a trivial thing to do better and to have really work um but and yet this cheapness so what what is so why is it so anyway I'm just giving examples the my thought was well I wonder if a lot of it has to do with the accumulation of the cost throughout the supply chain
uh and all of the capital behind the manufacturer of each of these little widgets um the cost of precision in terms of the attention that they take like the attention to detail so in in so many jobs so many forms of human labor you could do a really conscientious thoughtful careful job and really do it right or you can do a half-assed job a sloppy job and not do it very well and what I'm thinking is that if if if you if enough of those if enough of the high cost of attention to detail accumulates then you get parts that don't quite fit together because they're made by machines that aren't quite calibrated as well as they should be because they're made those machines are in turn made from parts that aren't you know quite as high quality as they could be and what is what is it what's going on that determines at least some of that at every step well at least I mean some of it may be just genuinely just the raw materials and so forth but a lot of it's just the effort and the time and the care that human labor puts into every single little step in the chain right every single little step of the way and there are some cultures where like Japan is an example of a culture historically I don't know what contemporary Japanese culture is like to such an extent but they're a little if you go back a little ways in into history uh Japan is a an example of a culture that is renowned for attention to detail tiny little things are given attention to detail that don't receive attention to detail in other cultures and you can see that in many of the sort of uh physical artifacts of Japanese culture like the way that cookies are wrapped just wrapped just how they're wrapped totally different a Japanese product than a product from India for example um or the United States and I'm I've always been struck by the incredible attention to detail in many Japanese products okay this is one example well that strikes me as something that is a distinctly human limitation in paying attention and being careful and thoughtful and conscientious is is not just time consuming it is for human beings but it's absolutely exhausting it takes so much so much more effort uh uh um cognitively not physically necessarily a little bit more effort physically but massively more effort cognitively to pay very close attention and do a very careful and thoughtful job of something especially something that is boring or uh frustrating or repetitive and and you know characterized by toil and drudgery and so uh it it it's it's it what what
might be I mean it seems like a very small thing not obvious at all but it but it might be quite profound what struck me is that man machines are just gonna they're not going to have that problem at all they will they will uh not suffer the difficulty of uh attending to detail they will simply do the best job that they can to whatever level of detail they are capable of achieving but they will achieve that with more or less the same consistency with me I I hesitate to use the word perfect consistently but the perfect consistency but my my impression is that machines will simply achieve that level of detail and precision every time or more or less every time within some margin of error and and that it will be it won't be they won't experience the same sort of difficulties that we do that translate into failures of attending to detail and um as a result one could imagine that all of the failures associated with uh are many of the failures associated with sloppy work of which there are I mean geez countless cumulative examples everywhere around us um as a result of human labor those could just I mean those would just be they would be they would vanish they would be gone they wouldn't they wouldn't they would cease to exist and what would the world look like if everything were uh built with the meticulousness and the attention to detail um that an app that Apple products are made with for example um as opposed to cheap electronics uh by comparison so anyway that was just it was I I apologize for being long-winded about it I I don't have a very succinct way of articulating um what I'm really you know the the this General Concept in the abstract here uh maybe you can help me turn this from some specific example into a more General set of ideas and principles that work here yeah maybe I'll take a stab at the general comment and then I'll push Matt to explain his um descent there uh so I was thinking about this before the seminar um I think there are many examples of uh we take human limitations as somehow more fundamental like being related to or so artifacts of our experience or our world that are consequences of human limitation we we can instead think about them as really properties of the world rather than we kind of externalize our limitations and uh and take them as kind of facts about the world and not just consequences of our shortcomings um uh let's so I I guess I could give some more examples but um so I think to come back to the sort of the reason you were introducing this
topic so they can be you're looking for subtle examples of things that will surprise us and I think maybe there is a whole category of such things which is which are of the following form so because of some human limitation maybe let's take the game of weighty or go for example uh because of human limitations we might think that the uh we might think that our kind of style of playing go or our achievements are somehow not just uh not just properties of us but we can think about them as being properties of the game right we can think about our limits as somehow being limits not somehow things that are specific to us or over interpret our ideas about the game as being somehow properties of the game itself rather than just our perspective on the game um and when Along Comes say a technology like AI or these other examples of disruptions that uh we can be shocked into realizing that this phenomena that we were projecting onto the world and thinking about as a property of the world is really just a function of our flaws and our limitations and this confronting this Gap is is kind of a continual source of surprise for us I'm trying to think of I mean maybe one example is the way we thought about the world before fast travel right um so we had maps with large gray areas on them and we kind of thought about the unknown the unknown status of the world as sort of being a property of the world right it was obviously a function of our ignorance but maybe unless we're paying close attention to that fact we we often thought about it we thought they were literally dragons there right the dragon is a thing out there in the world that lurks in the unknown or the Bermuda Triangle or or maybe now the example is UFOs or something like that we externalize our state of knowledge and think about it as actually being an objective fact about the world and once we Explore More that goes away and the world of course didn't change really once we explored it unless we changed it but you know if we're just flying planes over it and mapping it it didn't change but somehow we were quite shocked to come across a world that was explored that feels very different whereas in fact the world didn't change so maybe that's a kind of General framing that I would put on some of these examples maybe it doesn't work for all of them but having said that Matt do you want to explain what your comment on this first example was the um examples of good I like the staple example I have an example where um I've been through many third party
aftermarket charges for my laptop since I've got this laptop about 10 years ago and they don't last anywhere near as long as the original Apple one and I think it's a mix of um the quality of the electrical components inside the charger and the the soldering and the wiring and the circuit board and stuff um and like the quality of the the connections they may be aren't as secure and robust so these charges last notice will be shorter and actually the recently I've been um I've been uh buying a new charger and then it hasn't lasted even a few weeks and I don't know what to do now because I can't just keep buying an official charger well that's one option yeah it's much more expensive yeah okay that's somewhat you're going to go through more money worth it no I got chargers than a new laptop too yeah well that's true yeah I should buy a new laptop anyway um yeah I don't know if the electrical component is attention to detail but maybe it is maybe it's also materials I don't really understand exactly enough of that good details of manufacturing these like capacitors and stuff um but I wonder how that these companies have their costs whether it's attention to detail or whether it's um material like the quality of the materials they're using so if it was the quality of the materials then I could see that as like a a different kind of uh skimping that makes them kind of you know you can't visibly detect the difference but it's not one of these uh attention to detail uh Corner cutting cases like uh what Adam was talking about um so yeah I'm not sure about that maybe there are other ways cheap things are cheap um but my main stumbling point was was when you said Adam that um that these machines that we build these like if we build um AIS or whatever that are able to do this kind of design work and stuff um they will be uh able they will have some unlimited resource of attention span and patience and diligence that is required to do this kind of work like uh because I kind of think there's a reason why humans don't have that um except for virtuous humans um who are you know embodying these um qualities and and perhaps the Japanese who embody these qualities of patients and and attention to detail um that you know it costs money to to spend electricity um and so I'm just not sure that we end up in a world where it's much different um like we're going to come up against the same costs but then uh but at a higher level of quality maybe I can anticipate a Counterpoint which is that um once energy becomes
abundant um it may be difficult to scale kind of cognitive um energy in humans but not so hard to scale the um the amount of attention we can um we can budget for with our machines so yeah maybe that's a Counterpoint to know maybe spell check is an example in that direction I mean for a human spelling correctly as a matter of attention partly just knowing how to spell but even once you know how to spell at any given margin there's an error rate and you can pay attention or not and correct those errors or not yeah thinking through the point you just made mad in production lines maybe at a given level of expenditure on materials there's uh and also amount of resources you can spend on attention so if you're manufact manufacturing Electronics I guess it's it's assumed that the people say us in certain components into a circuit board say it's assumed they will make mistakes uh and then there's an error correction and QA process that attempts to correct those errors so there's maybe two separate Investments you can make you can invest in the training and the rate of production I mean you can lower the rate of production so less components go past the people on the production line and therefore they have more time per component or you can spend more money on training I guess although maybe that's at a certain point doesn't help and you could also spend more time on correcting errors so a more robust QA process I assume that when we talk about Apple's quality it's partly in intelligence required to design the product in the first place so that it is easy to manufacture at a higher level of quality perhaps that's part of it but it's probably also better training at the Foxconn factories for example maybe more time for the workers and maybe a in particular a more robust QA process but I don't know much about it but I would expect actually that the easiest thing to scale with capital is the QA process so maybe Apple just tests more and rejects more so yeah I don't know that really the margin they spend money on is is attention but I can imagine that that's more just to do with as Adam says human limitations so there's there's only so much you can do to improve the quality of production of a sort of an individual on a production line and okay so this doesn't get me past um like um using robots in the production line like you you uh a robot that is uh be it you know on a production line of words through your document and it's spell checking each of them or it's circuit boards on a conveyor belt and
it's inserting the components and soldering them with you know mechanical precision and then I get you yeah this thing is like infinitely patient in the same way that a computer can just you can just tell it to add numbers and the CPU will just go around in the loop all day and it won't get bought but we're talking about like Ai and disruption in the future so um when you design the hard part about getting to that stage where you have a precise machine that carries out your process to a really good quality is designing the by mechanizing that process that's hard that there's room for humans who are doing that to make errors um but if but if you this is not really the same thing the thing you get at the end this sort of fixed narrow intelligent machine that is very good at a particular task and you can do it repetitively and it's infinitely patient it's not the same thing as like an AGI for example which um it may be with advances in AI that are coming in in the near future we can do things like I I believe uh sewing sewing garments is really tricky to automate with machines because uh the garments are kind of really difficult to work with mechanically because they're kind of unpredictable hard to model um the fabric and stuff and so that's why we have um countries where um clothing is made um by human seamstresses um well actually um but maybe with advances in reinforcement learning we actually are able to build a robot that can uh definitely handle um Fabric and you know we we automate that as well um but it's that that's yeah okay I can see that kind of advanced coming but if we're talking about are we talking about AGI are we talking about putting uh a factory full of um Asimov style you know humanoid robots that have an AGI um and standing them all in a line and then they all work on the production line where we could have had humans standing in the line that seems to be then like it's not clear that you get um attention in that scenario yeah maybe maybe that's another maybe that's a good example of a limitation of humans that has dictated our industrial processes so humans find it so hard to quickly do tasks that are complex that we have arranged in industrial production line that divides a complex task into a hundred simple tasks which can then be performed reliably or relatively reliably and quickly by human labor maybe that's one of the assumptions that the Assumption being that's the only way to produce a complex object at large scale at for low cost if the individual
pieces namely the humans uh I mean there's plenty of room for machines to be to make quick decisions that are much more complex than we are capable of doing quickly and without getting tired right so at the moment say you're producing a circuit board uh you would have the basic circuit board and then you would plug various bits into it and the the people on the production line probably as I meant to they take a component off the production line they should examine it to see if it's like physically defective and pick up another one if they do discern that uh plug it in maybe do some simple checks to see that they plugged it in correctly and then pass it on to the next station um if the individual can make much more complicated decisions at speed then the decisions required for that step you could have more complicated steps and you would divide the production process differently so maybe that's in itself a um assumption yeah it's not clear to me that the the AGI in a humanoid um robot form is that much better at um context switching and making you know doing a complex task um because we pay for those things cognitively for a reason I don't quite understand the full reason but it's the same as with attention like it's kind of costing cognitive resources so but that seems more to do with the power budget of a brain than to do with some fundamental information theoretic limits I would say yeah that was going to be a question I would was going to raise um uh I'll try to remember to come back to that when Matt is finished the the just how little we understand about how these things actually are happening and working in the brain yeah well I would I would distill my point to something like um we don't know what AGI is going to look like there's kind of a chance that when we build an AGI um a lot of these things that we think are gonna go away because it's a machine actually come back because it's an intelligent machine and they're kind of fundamental to intelligence and I yeah behind that with a lot of weight but that's that's my point and attention might be one of those things and then Dan's example of kind of complex context switching and stuff uh would be another one of those things where I'd be like uh well what if that comes back I'm not saying that it will but well just to take a specific example so one of the places AI computer vision in particular is being applied uh quite extensively is in maintaining infrastructure so maintenance so one of the there's actually a startup in Melbourne that's
providing software to the Australian government for maintaining power lines I believe so you take photos of power boxes along you know at the moment there's a lot of human very low level human labor involved in maintaining power lines people drive around they look at the power boxes is that blown up or otherwise somehow needs maintenance no you move on to the next one you look at it now they take photos or the drones take photos and they upload it to a system that just runs it through a continent and answers yes or no for this needing further attention that's an example of a task that humans would quickly get tighter there's only so many pictures I mean it takes time to move from box to box there's only so many you want to look at and you like you you even structure your interest I mean there's a lot of infrastructure that's built in a certain way because that's the only way that it's actually possible to maintain it with human labor at a given cost right you don't uh you don't build exposed objects that are vulnerable to wear and tear uh all over the place if you don't have the labor to go and look at them and fix them if they fail and we just take that for granted in our infrastructure all over the place I would imagine that as these tools like confidence and drones for taking photos and feeding that back to systems that then send out either robots or humans to fix things once we get much better at that we will build infrastructure differently in ways that would look profligate and like just inviting disaster today because you would just look at it and think there's no way we could send out thousands of people to fix all that if it's fails but we'll just accept the failures and send the robots to fix it okay then I'll just I'll just wrap up I was just gonna say um that yeah that sounds like a good example and um cognets and drones are they are we viewing those as like a disruption in process like the the application of these technologies that are worked out in Academia and are being uh like that they're on the deployment roadmap um to you know the rest of the way we do everything well those examples are already deployed yeah I I don't know to what scale but it's already deployed yeah yeah I I can see more things like that and then it's more like that that's like a that's like industrializing the maintenance of infrastructure right you're turning it into a production line that has automated components and um that's that that makes sense and I agree you can get more attention to detail in that
mode um in that regime um and my concern is only in the well do we expect do we somehow expect infinitely more attention to detail from agis because we want to automate not only things that can be put into a production line but also things that kind of count yeah I I guess I wouldn't there will always be a frontier on which attention is relatively expensive uh maybe it's like the Dyson Sphere construction is uh is cognitively demanding uh but but maybe inside that actually sounds kind of simple okay very large scale thing of a kind of very simple you know you just bootstrap it with like some Nanobots or something yeah let's solve solve all those complex uh gravitational dynamical systems in order to stop it collapsing I I think we can put you in charge then if you think it's easy okay all right okay things like there's things like uh complex cognitive things like research I guess um you know seems to require there's not really many repetitive steps or something like everything is kind of new um yeah something like that did you have further comments on that animal should we try for some other examples yeah I had a uh just one quick thought and I think it's something something comes out of it that might be interesting um so just the general comment first and the specific the general comment here with respect to human beings is that we we I think we still have only a fairly low resolution understanding of um brain function and as an extension of why we uh fatigue the way that we do cognitively uh why attention feels subjectively the qualia of effortfulness and why that is uncomfortable under some circumstances but not nearly so uncomfortable under other circumstances so sometimes it feels more effortful to do something than for example that you're quote unquote bored or uninterested in then that you are quote unquote interested in or excited about we don't understand the interactions between motivation and interest and attention well at all to my knowledge um and uh I I think the the the the the explanation at bottom is almost certainly something to do with energetics right even calories were extremely difficult to obtain in our ancestral environment and um brains could only become so large uh before they you know affected reproductive success and so we know our brains are very much energy economizing and it probably is at bottom some reason why uh you know evolution is pretty picky and and I'm sure it it we were subject to selection on the basis of um our cognitive function being more
energetically efficient in some way and that that is some sort of ultimate explanation but having said that um uh there are sort of counter examples so so I have never done it myself personally so I don't have personal experience but my understanding is that there are a whole variety of um uh you know sort of amphetamine family Pharmaceuticals um uh uh Ritalin and um I forget what some of the other ones are and they can make paying attention much less effortful they can make it much easier to hear I'm sorry to to um uh to to um stay focused basically on something then then would be the case normally and my um my understanding also and this might get back to just sort of what general what what elements are required for Sapient general intelligence my understanding is that at least at least part of the theory behind uh executive function is that it is that there's a lot of tasks switching and a lot of sort of metacognition thinking about your thinking that goes on and so you're you're um one thing with paying attention and it's possible that that Ritalin and other Pharmaceuticals kind of shortcut this process a bit um is that you it's some part of your conscious uh uh awareness is more or less continuously returning to whatever it is you're doing and asking whether that's something that's still worth doing and it's it's almost like a check like a continuous check like you're doing something and then you know a few seconds later part of your mind is saying is this still worth doing and that and then you know okay it is and you keep doing the task and then your mind returns is this still worth doing and you keep it and we can feel pulled away from tasks um if you if you if you introspect you can kind of feel that happening if you're trying to pay attention to something that you don't want to pay attention to um and it's possible that that all of that task switching and then the the running that overarching metacognitive sort of executive layer that's sort of looking at all the different things you're doing and double checking in the moment whether they're still viable um that's probably energetically costly um and uh if you just turn that off then it's easy to kind of just focus on one thing and stay stuck there and I think people who people have described it that's a lot that's what being on Ritalin can be like um uh and then you can suddenly find yourself having spent an hour um monkeying around with your pencil or something and and you didn't realize it and so
um because you're you shut off the metacognitive metacognition um part of your brain a little bit anyway so that's one thing sorry uh too long-winded the specific example that occurred to me where that might be interesting is where it interacts with physical fatigue right so your brain obviously is part of your physicality but cognition cognitive fatigue is different than sort of the fatigue of your of musculature you're sort of your skeletal system's fatigue fatigue of your of your muscles um you know your tendons your ligaments your your uh nervous systems you know all of that stuff fatigues as you're awake and eventually you know uh becomes dysfunctional if you don't rest and uh combine that fatigue with cognitive fatigue and then you sort of get a negative feedback loop where as you get more and more physically fatigued uh you know that your your mind is your brain is having to adapt so your body your body doesn't work identically when you're fresh as it does when you're tired and so as you fatigue your brain is having to adapt and do things differently so if you're sitting on a production line the way you actually do the movements and the way you perform the tasks is going to differ at the beginning of your shift then at the end of your shift because your body physically changes and how it performs over that time one would imagine that that that may not be the case to nearly the same extent or even at all for machine labor right machine and labor may not noticeably or significantly physically fatigued in any meaningful sense at all in which case there needs to be no cognitive adaptation to that fatiguing over the course of a work period and again so you can kind of see how this is a feedback loop anyway um uh so there may be that sort of thing to think about as well for you know and and again like you were saying Dan we build consciously or not we build systems all systems today all infrastructure all manufacturing design all roles and occupations pretty much everything I can think of is designed around those limits and around those constraints and around those you know um uh shortcomings and characteristics that you would have hmm I thought of an interesting example it changes the subject a little um so this is based on my experience of spending some time in China and the pace of infrastructure development there so let's see let me call it conservation by default so one human limitation or the limitation of our current combination of machines and humans
perhaps is the rate at which we can change the physical world right so it's difficult and expensive to for example get rid of a hill and we have to get bulldozers perhaps explosives it takes coordinated labor of groups of people it takes tens of thousands of dollars uh and then bye-bye Hill or hello Road or hello building anything substantial takes decades in Australia and months in China but even in China you know it's it's remarkable to see a city change significantly every year when you go back to visit it but it's still a sort of bounded rate of change and that bound is to do with human limitations um but if I mean okay there are already people working on deep reinforcement learning controllers for industrial equipment like bulldozers and cranes and it works so you can there are now uh AI controlled uh sort of um physical land moving equipment that can put together a wall out of large Stones by picking them up and placing them and figure out how to do that sort of you give it a high level goal and it will figure out how to place the stones and more or less achieve it it's very preliminary but you could you know it's not we're not too far away from setting up the machines and then coming back at the end of a weekend and having you know some Earthworks having been done now once we reach that kind of scale uh with moving the physical world around it will change fundamentally how we think about the physical world so right now we think about like social media or the internet is this fast moving ever-changing fluid world where you don't really get your bearings there's no stable landmarks but the physical world well that doesn't change very quickly you know my road looks the same more or less now as it did three years ago they resurfaced it maybe the playground might get new equipment but it's still where it was five years ago but that's a function ultimately of human limitations not a physical fact about the world and our desire of how to allocate capital and machine time and and the limitations of our machines but as they become more capable the physical world will look more like the digital world it will be more fluid it will change more rapidly and be less stable and will come to realize that we had conservation that is unchangingness of the physical world by default as a but it wasn't really like a was it a an axiom of the physical Universe it was a function of our limitation yeah I think that's a very very interesting really quite profound point the the
slowness and I I mean rate of change is I mean the rate at which things change of course is probably how you really want to think about it but the there's a there's a there's a sense in which certain things have a permanence and I mean I mean it's in in effective terms not in literal terms but there are you could think of this you could think of the world as Loosely falling into a handful of qualitative categories and you could you could measure the rates of change quantitatively and it would be continuous you know from zero all the way up um but you can imagine qualitatively that you know in Human Experience there are sort of things that that are effectively permanent um things that are uh these that are Dynamic but unchanging so sort of like you know ocean Rivers you know they're they're in motion they're Dynamic but they're but at the same time there is a there is a a um a consistency over time and then there are things that are slow there there are there's a there's a slow moving category things that things that move but move slowly um things that do change but change slowly plants are maybe an example of that right I mean the plants aren't like a rock uh the Iraq for all practical purposes from the human perspective is static uh plants move a little more slowly but still noticeable um then you have plants and animals oh sorry then you have other kinds of animals um uh and human beings of course even quickly and I suppose in Human Experience you have some processes or some things that are that are really really fast certain insects move almost too fast to see um uh and and then now we've created a new domain of Machinery that that properly does work at you know there are rates of change within Machinery that that are you know going up from there in many orders of magnitude um with how fast they operate and how fast they change and the it's interesting to see what things will move out of the move around in those categories um as a as technology technology advances um and things that seem permanent or were in basically the permanent category or the very slow category um could could very quickly shift into other categories and become things that that can that can you know change much much more of it quickly I love that yeah I just realized that one's going to be knocking around my head for a week or two now I'm sure I'd be looking around the world very differently I might have some other ones but uh does anybody want to add no go ahead yeah so another one I thought of was um
this one's a bit more entertaining is tolerated so it's interesting to think about the the way that conversation has changed uh as a result of everybody having phones in their pocket that they can use to fact check things so there was a lot more of a certain kind before everybody had smartphones I don't know if you remember but uh people would confidently state facts uh about you know maybe they'd kind of heard something like a historical fact or something numerical fact and people would be much more prone to confidently stated uh not say I kind of think it might be sort of like this but just come out and say it and I guess mostly it was guys usually um that has really changed so people are much less prone to doing that and I can remember the transition because there was you know you'd be at quite a few social Gatherings where people would say something and then someone else would get out their phone and check it and it was wrong and there was like this moment of awkward silence and like oh yeah okay uh but people are just much less likely never do that now um so that's it's a of a certain kind has been uh decreased as a result of having quick access to authoritative sources um not other kinds of are less more resistant to just Googling the answer but as you know okay let's let's suppose for example that we had a um an agent sitting next to us here which we fed a transcript of our conversation and then it's kind of commenting in the chat and uh you know I say something about oh in the in China you know there's there's such and such happened in 1980 and this means such and such for technology and then the bot in the chat just sort of quietly notes that actually this is a pretty superficial take on what happened in the 1980s and your interpretation is wrong for this reason and that reason um you know it's it's not some it's not something you could Google right you can't Google that someone's opinion is kind of shallow but uh if you had an expert around who could put into context what you're saying and then refute it with an argument the way that an interlocutor in a conversation would do I mean if you if you have a conversation with someone who knows a lot more history than you you quickly feel a little foolish right because you have some ideas and then they're kind of like no I don't think so because of this and that and if you don't have it your command the same facts they do you might be right actually but it's it's rather hard to Trump someone who just knows more stuff
than you yeah thanks Matt um so so that's that's the case yeah yeah so I could imagine that uh well the flip side of it I mean saying is tolerated is one side of it but it's a kind of undermining of people's ability to feel agency and conversation um it's maybe another way of putting it and that's it's sort of an I mean the Assumption here being that uh like it's it's a subtle thing right so we're having conversation here and there's what four of us and we have certain amounts of knowledge but uh we would talk differently and not have the same conversation if Henry Kissinger were here right for various reasons but um there would just be a kind of like we'd be a bit sheepish maybe to express our somewhat limited understanding of certain topics knowing that a real expert was in the room um but maybe a real expert could be in the room for 20 cents if we wanted fairly soon right uh so there will be changes to conversation which are of the form not I'll check in my on my phone to see if what you're saying is correct but let me bring gpt4 into the conversation pull out your phone press on and then it's listening and it's kind of pointing out your you know uh superficial understanding of quantum mechanics and explaining to you how your dumb ideas about how that fits into New Age philosophy uh really very shallow that's not our quantum mechanics Works etc etc so the same kind of experience you might get from an irritating friend um who's willing to to do that you could actually get it for free well here's a very cynical uh idea to follow on and add into that Dan which is um if if it becomes the case that that facts become easily checkable uh and this has already occurred uh one might ask and and then go out and observe the world in the social media landscape and see if it's already occurring um would this mean that conversation and would would in in the interest of safety uh and not getting called out would people gravitate more towards making unfalsifiable planes in their conversations that's the first thing and then as a second thing as a second thing we can we could ask um what sort of what sort of claims are unfalsifiable and I'll give you a terribly disturbing and cynical example um claims about other people who you don't know claims about their intentions are unfalsifiable right because you can't read you can't read someone's mind right so Elon Musk is doing all this crazy stuff right now with Twitter and it's a lot of it's inscrutable and and everybody's got an opinion about why
he's doing what he's doing but it's safe to voice those opinions because they're unfalsifiable Right facts like claims about facts it claims about you know the way the world is um it could be that could be falsifiable those are much harder to make and people do make those in the world's full of you know uh of that kind and misinformation and disinformation as well but I have definitely noticed uh the tenor of the general tenor of conversations um deteriorating when they when when people when people devolve into attacking one another's intentions um in fact I've come to have it as a sort of personal principle that whenever I see online a conversation devolve into claims about who what a person or an organization or an entity's intentions are I just I just tune out I I disconnect from that conversation because that those are unfal survival they're unknowables if there's too much uncertainty there and it's very easy to attack intentions as a straw man right that's super super easy to set up uh you know to impugn the intentions of an opponent for example as a straw man and then attack that it it just really you know um juvenile uh uh argumentation and and debate when that happens um and and that I I my experience now maybe I'm just a cranky old fart at this point um but I think that it's gotten worse in the era of social media I think that that we used to give people the benefit of the doubt about their intentions a bit more than we do now and um uh maybe that was because we had to tolerate uncertainty of other kinds especially natural ones like whether or not somebody knew their well enough and you had all the facts now oh you have to be so careful because it's so easy to get called out um by somebody with a smartphone or as you were saying damn by a GPT bot um you know refereeing the conversation that the only safe things to say are unpalifiable things about people's intentions um and uh man that sucks but holy moly is that a recipe for toxicity in a conversation no so that's why I say it's a cynical tape what was your take me imagine a world where where this is the case the the the gpt4 example let's bring that into the conversation and and so on um I feel like that's probably so so with the let me go back to the example of the fact-checking person with the smartphone at the social Gathering that's like a that's like a move you can make in a social situation to say well actually I believe that's probably wrong so let me pull out my phone and let me let me call you out
and then see what happens and that seems to have like um calling someone out seems to have kind of stuck as a you know a strategy that beats claiming a a bad fact or whatever and we've reached this new kind of equilibrium in um social interactions where people are less uh willing to um confidently State something like that okay um it's not clear to me how the social game is going to react to the introduction of this new possibility this new possible move that you can make which is bringing GPT into the conversation and getting an expert opinion on um on these opinions that are going around so like would people want that maybe they would avoid that maybe they would be like you know what I actually feel more comfortable having a conversation without that expert um listening to what I have to say um let alone like the Privacy implications of that um because for example uh you know I my my gut reaction is it would be better if we it wouldn't necessarily be better if we had that constant influence of an um establish opinion voice kind of correcting everything that we say because um it could kind of be stifling for for um new like how do you come up with a new unorthodox take on something that's maybe established knowledge or something like that um it's one thing to become kind of more informed but there also needs to be a space for ideas to kind of uh grow and people to have half-baked ideas and to um to talk about them with with friends and get get feedback and stuff um in a kind of gentle way otherwise these new opinions are kind of going to be stifle but they may actually be like the the Orthodox expert opinions may actually be subtly wrong and so well then maybe incomplete or missing something so it seems important that um we have some ability to have naive conversations and um so one question is if this possibility becomes available will it actually be used and another question is should it be used um or how should it be used or you know should it always be used yeah I like that maybe there should be a an equivalent of the shallows or the mangrove swamp for small fish so yeah you do need to have a space where you you can actually talk a bit of nonsense right it's not actually an ideal that in every conversation all the time you only stayed completely factual things and you never explore ideas by messing around a little bit that's this is what this is what this is the this is what the the proverb I don't know if it's quite the same in Australia but this is what the bar was for for the
longest time in the US I mean you're just you go and you get you drink a few drinks and you you it loosens your tongue a little bit and then people you know we have phrases like shooting the you know and I think that there's a reason for having those those idioms you know it's where yeah you have a place where you can go where you can kind of just spout off and drink and be disinhibited and and not be judged and not be scrutinized and then things are forgotten the next day we and and you know of course that's one of the terrible things about the internet of course is that you drunken tweets haunt you forever right um whereas you know drunken nonsense you say at the bar last weekend it's just gone you know it's just gone um it doesn't haunt you forever and and uh so there was a there was a social a social outlet that built in a tolerance and a forgiveness like you were saying Dan the shallows for the for that sort of thing um yeah so anyway yeah we we we had we had some long-standing social institutions perhaps serving those functions that have been a little bit you know displaced or eroded by by um new uh social media I suppose now we have many uni seminars okay um yeah that's right yeah you don't have a beer in your hand I do okay um hey the other thing is you can't record that they can't be recorded sessions right because I don't want you to come back to haunt you that's right yeah uh okay I had another one um the world is intelligible this is interesting because it didn't used to be true right so uh you can tell that from the way we we think about um the world beyond the sort of limitations of our Social Circles and our machines and our buildings it used to be there's a dark forest and there are things going on in there that are mysterious and Beyond human comprehension and maybe scary now we have dark forests in movies but you know I've never been in a forest and felt like maybe there were monsters somewhere that I you know had to pay attention to uh but maybe this is an interregnum that we're in where the world seems like it's kind of I mean it's not intelligible really right they're nations with millions of people can't really be reasoned about or surrounded by institutions and organizations that are deeply mysterious to all the people in them we don't really know what's going on but that's not really what I'm talking about what I mean is that we don't feel that there are agents in the world like localized units that are small like animals or people
that somehow uh well I suppose you could think about the the population of dolphins and whales out there and we don't know what they're doing or what they're saying but we think about that very differently that's like a subject for biologists and uh and um Marine scientists to study it's not So Much Anymore something that's mythological or deeply mysterious and kind of an indication of the world being beyond our power to comprehend given sufficient time and tension um but we could soon repopulate the world with deeply mysterious things right whether the AIS or just animals that we engineer or uh or just things are changing quite quickly so um so an assumption based on human limitations I don't know how this fits into this category exactly but um so uh intelligent people means intelligible to humans so yeah this one's kind of maybe it's it's really the the the version the version of this that is the you know if if we made the the fallacy of believing that the world itself is uh mysterious when actually it's just unintelligible to humans or the world is knowable really it's just the current world is understandable to humans yeah maybe it's it's kind of two things that I'm saying one is that the human sphere has spread out to cover a lot of the surface of the planet so that we can basically live our lives more or less interacting with human products or things that are human clients so you know dogs cats machines uh we can organize our world so that we interact very little with things that are really Beyond humans so that's one fact and then within that sphere of influence most of the world is designed for us and intelligible to us right things have buttons with labels there are signs around machines come with manuals if we choose to read them pets are not very intelligent relatively speaking so they they surprise us but not in ways that we find deeply disturbing um so that's kind of what I'm getting at and both the sphere of human influence might shrink if there's other kinds of intelligence around and within our sphere of control the products of our intelligence may may not be very intelligible to us so we we may move more back into a period that feels more like the past where there really are dragons on the map and uh uh Golems around the corner whose intentions we can't really judge well I I I know if it may be a maybe there's a little bit of selection bias going on just in in this group um but the the world is already massively the human built world is already massively unintelligible
and I mean it already that so on the one hand in some sense for some people the world is more intelligible than it was in the past uh you know 500 years ago for example and I think that's fair perfectly fair to say okay it it certainly to in a general sense or to a first approximation okay okay but having said that um oh man you know the how what does your average person and I include myself in this absolutely has like next to zero idea how in the hell their smartphone works at all it's just totally unintelligible magic to me I have no I know there's a battery in it and I know there's a computer in it and I know there's a camera and a microphone and beyond that I couldn't tell you anything and so it's just just completely unintelligible and there and so and I don't just mean like machines like Electronics but so many things are that way many people don't have any idea how their government works um uh the we don't know how things are made we don't know where our food comes from you know it's it everything like that's exactly right there's so much of our world is a black box now and um so I think the starting point well I agree with you Dan that the world will become more intelligible in some ways thanks to advances in in um our scientific understanding of the world uh for civilization as a whole the world becomes more intelligible for individual human beings existing within the world if you assume we are our our current limitations persist out to some timeline you know before we're all upgrading our brains and everything um then uh yes the world will become the human built world will become less intelligible as more and more black boxes of more and more marvelous capability emerge but we're starting we're not starting from from a from a really high bar already let's I guess that's my main point where there's there are a lot of black boxes and a lot of misunderstood things um uh and and then what's even worse that and this is there so we were left about the bar a few minutes ago right he's shooting the at the bar there's always the guy at the bar who's melting off about stuff he has no idea about right of course because that's what the bar allows um and so there so it's not just that the world is unintelligible but there are people who think they know how the world works or some part of it and they don't they've got completely the wrong idea they're missing for it um uh such as old political enemies yeah exactly this I mean they'll put the entire I mean you've got you know bless the my
heart goes out uh because I genuinely believe it's a psychological disorder but the entire conspiracy thinking conspiracy theorizing smear I mean we're talking not just a small number to millions or tens of millions of people around the world who generally genuinely believe they have an understanding of how things are operating and working that and of course it's deranged it's it's delusional and and not doesn't map to reality at all um uh so that you can be worse than than mystified or unformed you can be you could be laboring under uh under actively harmful delusions uh we can comment more on that one but I have another one that to to take a different topic to AI for at least one of these um uh where is it right uh so assumptions based on human limitations only a small proportion of land is habitable slash land is scarce this is a well we talk a lot about the influence of property prices on behavior and allocation of talent and quality of life and all of that uh family formation there's plenty of literature on some some of the problem to do with population growth or people not being able to have the sizes of families they would ideally like as to do with property prices in many places well why is why is land scarce in some sense it's not scarce there's lots of Farmland around and it's not super expensive there are lots of places you could go and live and property isn't that expensive um but still we tend to congregate in relatively small proportions of the land area you know in Australia this is very noticeable there's lots of Farmland but lots of land that basically is just too arid to productively use for most people now of course there's animals on it I'm not saying it's unoccupied or maybe it has value to people as it is so I'm not giving a normative judgment on it should be green and lash and you know suburbi Suburbia uh but there is a lot more land than we tend to think okay you fly across the U.S and it's amazing how much is occupied and family and it's really stunning how much Farmland there is in the US but if you fly from the west coast to the east coast the first couple of hours are just flying over what looks like complete desolate garbage Land from the Mars right it's just desert or something like desert uh but in principle with sufficient energy there's plenty of water on the planet it's got salt in it but with energy you can get the sold out you could do plenty of things with all that land so in some sense the scarcity of land or our sense
that there is an infinite high quality green acreage out there waiting for us is to do with technological limitations more than a fact about the world I I completely agree um one thing to add into this is the the um and this has already occurred uh to a limited degree and that's that human Mobility our ability to move our ability to travel distance you know given any given distance in any given amount of time as that has uh changed has as we have become more mobile um our sense of space and distance in in human in Practical terms um has changed um places that are actually enormous if you're on foot feel small if you're you can access uh vehicles and places uh in the US is criticized uh heavily for this of course um places that were once not habitable because they were too far away um became habitable thanks to Automobiles and so suburbs and Perry Urban Suburban X urban areas and out into rural areas at the distant periphery those became viable places to have um to have a home because it was it was possible within some reasonable and reasonable debatable but within some fairly manageable amount of time to travel from that place of habitation to some place of work right so you could live a hundred miles which is an absolutely stupefyingly large distance if you're on foot but you could live a hundred miles away from where you work every day um and people do in the United States if you have an automobile and you can drive for hour and a half each way to get to and from your place of work and live a very very long way away and so that's a technological constraint I mean one could imagine you know at some point in the future uh autonomous aircraft that travel very fast could you know extend that radius of viability and and that's assuming you you still want to access certain places um now maybe what it might not even be for work it could be that that you it could be that psychologically it feels like you're living near the big city as long as you're within 300 miles of one because you can jump in a in a you know an automated air taxi and be there in 40 minutes or something like that um you can be downtown in 40 minutes because the air vehicle moves 400 miles an hour or something like that um one could easily imagine a circumstance like that emerging that would continue the trend that we've that we began with automobiles um so uh there's an interaction certainly between our perception of definitive land uh and the availability of energy to make that land habitable
and the technology required to uh to travel within and to and from that area of land yeah I suppose also there were I'm thinking about the area where I grew up a large part of it is basically swampland uh was swampland now is Suburban uh developments that land wasn't really usable for human habitation until you could move large amounts of Earth around and redirect rivers and and so on so it would have been perhaps very likely productively used by Aboriginal people there would have been lots of animals living there that could be uh could be hunted and other resources that could be gathered so it would have been part of a productive sort of sphere of human activity no doubt but it wouldn't have been a place that you could build habitation or regularly live in um so yeah I guess it's land sort of falls into cognitive categories that are based on the current level of technology and we just without much questioning think of well there are large areas of the planet that are freezing cold and therefore uninhabitable or deserts and therefore uninhabitable um but I remember uh as an interesting scene in I don't know which episode but in Star Trek where they're they're talking about well in my time the moon is green because there are millions of people living there um and I thought that was a at the time it kind of struck me as I never thought about that uh I mean obviously you can think about moon bases but the Moon being green somehow is being thought of as a place like in the future in that future the moon was just another place where people were it wasn't some distant forbidding territory um and there are right now many places on Earth that feel that way but that that is just kind of an unexamined assumption perhaps that might go away I think it's also um just while we're on land um the we have a fairly two-dimensional conception of land I mean the the for obvious reasons but the the the surface of the Earth where all human activity occurs is just vanishingly thin veneer on top of the entire volume of the planet right um and it it I think I think it's it is remarkably difficult to build an intuition of just how much material is under our feet um because we really are sort of Flatlanders right you might I'm sure you guys are familiar with that quite famous you know sci-fi short story flatland um it's about two-dimensional beatings um and we're not far off of that right I mean we're pretty close uh but the um speaking of Earth moving um I I over the last six months because
of a summer project I was working on in here in in my home um I was doing some Landscaping work and I gained a completely new appreciation for for just how much volume of material there is so when you try to when you try to move Earth with a shovel manually any significant amount um it's it's it's just it's mind-blowing how much material there is and how how massive it is how much it weighs like a cubic meter of of soil is so much material and so heavy and so much work to shift even the smallest distance and um so I I personally again just a completely new appreciation for so anyway just all I was to round out the thought very quickly um I don't think that we are going to as a species uh you know uh anytime soon be building deep underground hundreds and hundreds of stories or anything like that but the scope is certainly there to to explore a um uh explore the the third dimension explore beneath us and above us to a greater extent than we have already and in many um uh places on the the the the the the the near surface underground is very well suited to a lot of infrastructure that would be um completely transformative so for example um uh you could easily I say easily in the in the in the science fiction uh context so in in science fiction terms it's straightforward to imagine undergrounding all transportation for example you could just all Road Transportation you you could conceivably build a tunnel underground where every street is today where every Surface Street a roadway on the surface is that's not inconceivable it's not even particularly far-fetched it's just really really expensive if you have to do it with with current technology but it's it's it's that that's not uh you know that's not building a Dyson Sphere you know building building a thousand or even ten thousand miles of tunnels underneath a city like Los Angeles is not building a Dyson Sphere it's it's you know it's it's it's relatively plausible um and uh a lot a lot of other infrastructure could go under there I mean we we manufacture um all of our manufacturing facilities are built on the surface of the Earth right now but again that sort of do accommodate human beings because human beings don't like being underground like moles and you know it's expensive to dig out you know from underground and so forth and it's and in the limit it would be hard to figure out where to put all the material you brought up from underground you would have to have some some uh idea of what to do uh with all of that stuff
um but you you could certainly imagine undergrounding a massive fraction of you know the productive the productive production base right of the of the our civilization and freeing up the surface of the of the Earth for you know other use so the idea that that uh the ultimate destiny destiny of a technologically advanced civilization is to Simply pave over the entire surface of the planet with Urban landscape or buildings or something is pretty unimaginative I mean you know you could imagine a society that would made that choice and there'd be lots of famous examples in you know in in sci-fi in the literature coruscanton Star Wars and um I forget what the Homeworld and uh the center the center of the Galaxy is in Foundation but the where the entire surface is just an urban area um but that it strikes me as unimaginative um and uh you know you could easily imagine an entire world just being Wilderness and and still having you know it orders of magnitude greater productive capacity than we have today and all of it being underground just for example so yeah on that note uh yeah Matt contributed one in the chat uh which is the idea that nature is infinite and this we've come up across I mean this was a a theme of the 20th century 19th century uh very much and maybe were more attuned to it now but at the beginning say with fishing or any other cutting down trees it was easy to imagine that the scale of human activity which is not fundamentally affect the reservoirs that were there in nature and of course we came relatively quickly to to appreciate that our activities were at a scale where that was not true um but maybe there's further lessons to learn in that direction that's a bit related to the conservation by default I guess where uh not only I mean of course we know and we understand every child learns over and over and over again in school about the capability of humans to send animals extinct species extinct um but we could I mean we will reach another order of magnitude of capability of reshaping the Earth I don't know and how will that feel different beyond our current uh it feels like a topic that is perhaps over um explained to young people this capacity and it's uh potential for harm but how could this cognizance of our capabilities be any more present maybe engineering new species would be a kind of additional level of power that changes well I can I can offer some thoughts on this it's a it's a set of questions I've given quite a bit of thought to um but uh I don't want to monopolize the
conversation if anybody else has any any thoughts immediately on this anybody else okay um well uh I have a new book coming out shortly and one of the topics that I discuss in that uh is this is the question of um conservation versus preservation versus rewilding those are three broad categories into which uh um environmentally conscientious land use tends to fall they're not the only conceivable ones but they're they're certainly the ones that we cover the most when we take students through environmental ethics um in in in you know undergraduate and graduate courses and the conservation ethic is is the is I'll just very quickly run through it the conservation ethic is the idea of protecting the ecological integrity and and um uh possibly the geographic and geological Integrity of an area it could be landscape or a volume of of um Marine biosphere for example uh but in any at any rate um maintaining and protecting the Integrity of that ecological area for human use in other words in service of human interests um so that that ecological area can continue to provide what are called ecosystem services um so this is a utilitarian and instrumental uh uh perspective on the value of ecological Integrity an area is is valuable because of how useful it is to human beings um the preservation ethic uh argues or or views the value of an ecological area as and the ecological Integrity of an ecological area as innate or intrinsic Not instrumental not the the value is um the value of ecological Integrity is self-evident under the preservation ethic it is assumed it's axiomatic and that Integrity um uh and the right to it the right for it to persist over time is is yes it's intrinsic is the correct word there and so this is protecting nature for its own sake not for our user purposes um the rewilding category is different from preservation in that it involves an active effort and involves active human intervention to determine these the uh meaning of ecological Integrity in a given ecological area what that means is that with rewilding you have to make a set of choices about what the structure of the ecology ought to be and so that's an ought versus an is sort of question right it's it's this descriptive versus a normative kind of uh uh distinction so with conservation and preservation you can be agnostic about what the landscape ought to look like you just leave it alone or you exploit it in a way that that doesn't rapidly um uh reduce the complexity of the biosphere that's there or the the
ecosystem that's there for example but with rewilding that's a different set of ethical questions you have to make a choice about what the structure of the ecosystem ought to look like and so if you were to take an example I'll just pick an example near myself because it's easy so there's lots of Farmland where I am and there's a little bit of urban area so take a piece of land that is now a large parking lot that 50 years ago was a farm or maybe an orchard like an apple orchard a Cherry Orchard we have lots of those here um 500 years ago it was a it was a climax Forest a temporary climax Forest um with possibly a little bit of silviculture from indigenous um Aboriginal Native American populations but but in many cases not or why in the back collect back a bit further 15 000 years ago there were two miles of ice on this land and then twenty thousand years ago we should go back to that so this is well this is the question is is at each of those different points and then you can have sort of just after the last ice age or just before the last ice age here the this ecological um uh structure of the Enviro of the area here they were completely different there were different megafauna different large animals different species different species at every trophic level and animals and plants both so the the the this landscape as a Wilderness looked very different even within geologically extremely short periods of time not over millions and millions of years but just over a few thousand years right so that then it becomes a real question if you were to undertake rewilding to to what time do you wind the clock back this is a a difficult question um and then if you presume you have the technology which maybe we do you were applying it Dan um what if we Could reconstruct you know either de-extinct from tissue samples or perhaps you know construct Jurassic Park style facsimiles of um previous uh previous species should we should we restore you know the woolly mammoths to the landscape and mastodons to the landscape should we um bring back the saber-toothed cats should we you know how far back do we go so this is a I mean it's a there are no clear answers to these questions there's no clear right or wrong questions um but of course what's fascinating is that they won't be just mere you know philosophical thought exercises for undergraduates um to mull over um uh and for people to yak about at the bar and shoot the over at the bar um for much longer you know within our lifetimes and and uh you know with it
probably within a you know several decades these are going to be practical questions we're going to be confronted by them and then we're going to have to answer them and we're going to figure out you know um we are going to have to figure out how to approach that challenge how to answer these sorts of tough questions um um yeah maybe I'll add one last one this is one I think about a lot it's sort of it's sort of come up um in previous seminars but uh and that's the illusion of depth from complexity and I'll explain what I mean but if you'll indulge me I there's a passage from a novel that I often think about when I think about this I don't know if any of you know Gene wolf he he writes in a few genres but mostly sci-fi or fantasy I'm a big fan he has a book called The Wizard Knight which is his take on the kind of typical fantasy uh hero's journey and I'll Briefly summarize it and then read a passage that I think is quite astute on this point so the main character Abel is a young boy in our world and then he somehow gets into this magical world and he's appears to be a grown man a very powerful strong grown man and he becomes a kind of Knight typical kind of story he falls in love with an elf or I guess it's the elf queen who's from a world that seems like it's below hours in some hierarchy uh and he goes on some adventures and at the very end of the book he finally is reunited with this elf queen and uh and there's a scene where the Val father in the novel which is basically God is punishing him so he's broken some oath in order to save some people so he was supposed to not use his powers but he did and then the following interaction happens the Val father gestured to Western you've served your knight Faithfully Western is his Squire I guess you must do him one more service bring his Helm and set it on his head so there's some magic Helm which if you put it on you see the world as it truly is so you see through all the illusions uh so right uh Winston did lovely to Siri who's this elf queen that he's in love with became a puppet of mud and leaves that was horrible but I had expected it blah blah blah you see what you are surrendering the Val father said and no to what it is you go what will you do I drew my dagger pushed up the sleeve of my shirt and cut my arm drink I told the Siri and she bent and drank of my blood not a few drops as elf often do but great sobbing gulps while I clenched and unclenched my fist so that human life flowed freely never stopping
until a small green-eyed woman Stood Beside Me I have to imagine this is from some kind of Mythology I don't know this story I don't know where he got it from but often he's interpreting old myths this strikes me as quite a deep observation so about human psychology and relationships in that case but you could generalize it so there are many things in the world which appear meaningful and deep to us but more or less because our perceptions are limited so uh the when you put on that helmet and you can see with the uh the far-seeing eyes and see through all the depths of complexity and the opacity of of the world maybe things don't seem as meaningful anymore and I've often worried out loud in this seminar what that will do to people when the helmet is put on them or it's available to them through um artificial intelligence so yeah I guess in general I would say that a an assumption based on human limitation uh it's kind of this illusion that things are more rich like we look around us and there's culture and we have lots of stories and there are games that we play and they feel difficult and they feel like they have some depth to them right so personally I was drawn to mathematics for that reason right it feels like it's deep there's kind of meaningful stuff there and you can push into it and it pushes back and you'll never get to the bottom of it but that that feeling feels like it's a property of the world right it feels like mathematics is deep I don't I don't feel that it's an artifact of my limitations that it's deep I feel that it is deep but many of the things we feel that way about May in fact actually not be deep like from some point of view with the helmet on so to speak um and I do wonder the degree to which were mature enough to see with that helmet on and uh move forward anyway and make real the depth that was an illusion before uh so that's um yeah one of the one of the things I I feel uh is most relevant to me personally uh in including my attitude towards mathematics when it comes to the influence of AI maybe there are other examples you can think about well the the perhaps the flip side I don't know if it's a flip side or it's sort of orthogonal but the Plato's allegory of the cave right um where the the the it's sort of the reverse which is that things look simpler and lack depth and they look two-dimensional and flat and simple just Shadows on the wall and that the the the reality is is has more depth and has more complexity um and that seeing that entails on a you
know an Awakening and an appreciation of some sort um so there's there's uh yeah I can kind of see give up some things and get others yeah it's um I don't know um it's tough to say because I mean you can imagine you can also think of your own personal uh growth and development from childhood to adulthood I mean we've all done through that we've all experienced that it's difficult to cast your mind back of course um to the way you perceive the world as a child but I think it's fairly I don't think it's uncontroversial or rather I don't think it's controversial I think it's really uncontroversial to assume that for most of us you know we had a shallower understanding and view of the world as children than and we've come to have a deeper view as adults I'm sure there are exceptions in ways in which that's nuanced but to you know first approximation it's very probably fairly true and of course you could ask the question you know to what degree was that disillusioning and painfully you know and a painful process has to go through and under what circumstances was that an enriching and and um fulfilling process to go through yeah I think that's right and that's what this story is also getting at I mean it's in that case it's the example of a significant other right you I guess when you're young you sort of have this idealized version of the opposite sex and then you you meet and you fall in love and you discover their kind of horrible people just like you're a horrible person but then you you learn to deal with that and it's it's good anyway I suppose you also discover with respect to your parents that your parents are people at some point right before that they're kind of idealized figures if you have good parents they're sources of support and warmth and then at some point I guess usually when you're a teenager you discover that actually their opinions maybe don't make a lot of sense sometimes or or maybe they're biased or just ignorant on some topics and that's a bit of a shock perhaps and then you learn to love them anyway but in a different way a sort of a more mature way um so yeah that's right and we have experience of that with our personal relationships and maybe maybe there's a generalized process that looks like that um I suppose another sort of jet broad category here uh is the you know the idea of heterogeneity um versus uniformity and uh to what extent flaws and Imperfections um create interesting differences and to what you know to what extent are I mean so
foreign this as a question which is for artificial intelligence um uh flaws will not be as unavoidable um shortcomings lack of capacity you know uh undesirable qualities As we judge for ourselves right I mean so so a person can look at themselves say I wish I were different um most of that well that's not signals but but for a for a machine an artificial intelligence um that those characteristics will not be nearly as immutable as they are for human beings today and as a result of that um the uh the the the flaws and and weaknesses will be more voluntary if we have them and difference difference uh May originate from it may still continue to originate from flaws and shortcomings those are things that make a person unique today um it's the reason and we're not all the same and we're very human beings are really quite different from one another and it's a substantial portion of that is um not just in our different sort of um uh capabilities but also in our very different quirks and Oddities and vaults and flaws and shortcomings as we would judge them not necessarily just as other people would judge them whereas or according to any absolute standards just as we as we introspect and Judson for ourselves so I wonder um uh what role the how that might change for artificial uh minds and artificial life that that for which flaws or for which their characteristics are not immutable but are subject to intentional um uh intentional change or intentional assignment I guess and then to what extent are we likely to see difference in diversity will it be valued for the same reasons in the same ways um and I would presume this isn't I might as an optimist I don't think this is likely to be a problem at all because if you remove immutability almost almost certainly you create a greater range of variation that can exist right um and so I'm not concerned that there will be a sort of some massive convergence towards a boring uniformity wherever every every conscious entity is is basically the same that would be very dismal and boring and you know there's some dystopian sci-fi that's like that that suggests that we will all just Converge on some single you know monolithic uh uh uh you know ideal and the world will be very boring because everybody will be just looking like a supermodel but all the same um I think that I I'm more optimistic again I think that shows a lack of imagination um I think we're likely to be you know vastly vastly variable um uh and over time as well as you know at