Matt Kuperholz on the ethics of AI

Hi friends, I talk a lot about systems thinking on this podcast and thought I'd share one of the most potent actions you can take if you feel moved to be a different kind of leader for the 21st century. At Small Giants Academy, we developed our answer to the traditional MBA. We call it the MBE, a mastery of business and empathy. The MBE is a truly groundbreaking program which equips leaders with the tools, strategies, networks and philosophies to lead with purpose in these troubled times. Applications are open now, so head to smallgiants.com.au/MBE to learn more and sign up. This podcast was sponsored by Clayton Newts. Looking to align your portfolio with purpose, our partners at Clayton Newts work with investors, financiers, fund managers, B-cops and social enterprises on sophisticated impact investing. From direct investing, venture capital and structuring to regulatory compliance, tax and strategy, let them help you turn your purpose into performance. Head to www.clatennutes.com/impactinvesting to make your capital count. It is a big trap to think about this as intelligent the way we are. Our brains evolved in the way they did. They had no choice. A computer is completely different, but it's still intelligent if we just think that it's seeking goals with feedback. Hi, I'm Berry and this is the Wisdom & Action podcast, brought to you by the Small Giants Academy. Matt Cooper-Holtz is an accomplished artificial intelligence scientist and advisor at Businessman and at Creative with a wealth of experience applying AI to real world challenges spanning over three decades. He's a techie and I'd say a real optimist. This conversation with Matt is a sprawling one. It's exploratory. Matt felt like the best first person for me to have a conversation about AI with. I have so many questions about AI, about this moment that we're living in and I'm really coming at it as a laddoit. You know, I haven't been deeply engaged in technology and artificial intelligence and it is the conversation that the sort of meta conversation that keeps coming up and so I'm going to do a series now of a number of conversations with global experts on AI, technology, ethics and the future to try to get my head around it and hopefully I bring up some questions that you all have as well. So this is like feels a little bit like Lord of the Rings just wandering through the swampy muddy places to try to find the right questions to be asking in this moment. Hi Matt. Hi, Derry. How are you? Nobody knows the setup that happens before we get into the conversation and that was one of the funnest setups. It was watching Matt Cooper hold's geek out on AI and tech and just you felt like you rebuilt the world. You said that your computer was in AI building mode and that's why we could hear the fan in the computer while we were talking. To build an AI model, you've got to do a lot of maths and to do a lot of maths you've got to run the part of the computer which was originally built something called a graphic processing unit where the first use case was make video games flow smoother because there's a lot of maths to calculate moving video games and then people realised that you could use those same calculations to build AI neural networks. So when you're building your computer switches into this super overheated mode your power bill starts going through the roof and it needs to keep everything cool so it runs a very noisy fan. Now when we're just having a chat, the computer is like basically half asleep it doesn't have to think very hard so I just flipped it into a different mode where it didn't need to cool itself down because obviously with power usage comes eat. Welcome to the world of Matt Cooper hold and the AI engineers and builders and creators. It's not a world I know as you know which is why I'm so excited to have you here to begin a conversation that I have been avoiding really because I'm a lot of it and I'm 46 and I just wasn't a part of the world that built the internet. I think everybody in the world knows that AI is here and it's changing the pace of everything and the questions around the ethics of AI which is something you've been talking about a lot lately. You know in the climate movement we talk about adaptation like we won't be able to mitigate a two degree a three degree a four degree warming even though of course we're still trying to find those interventions that will stop this train from hitting the wall at 120 kilometers an hour but AI is arguably a greater existential threat to humanity. So there's two things I want to do at the beginning. First I loved in your TED talk when you said that you've been in a deep and meaningful relationship with computers since the 70s and so can you say more about your journey to being Matt Cooper-Holtz? Why are you an expert on exponential intelligence AI all the things? You know if you love what you do then you never work a day in your life and for whatever reason some people are born beautiful some people can run fast. I was born a technologist I didn't have teddy bears I used to call into my parents bed dragging a toaster by the cable. A literal toaster? It's toaster anything electronic was my happy place and I can do mechanical stuff you know gears and motorbikes and that sort of thing but it's really always been electronics and maths that were my genuine passion as a small child as a medium child I still remember the first time I met a computer in 1976 in South Africa and I found tandy electronics in Campbell Junction where I grew up had a computer on display and I'd ride my bike up there every day after you know I'm talking like great three great four and and play on their computer so much so that they said hey kid do you want a job and I would stamp the little receipt booklets with this stamp that instead of them writing the receipt and then stamping it they'd be pre-stamped and I'd make 20 cents a booklet which I saved until I got to 549 dollars and 95 cents and bought my TRS 80 color computer and I was the happiest boy alive and it's that mixture of pure logic it does what it's told and pure creativity your brain can decide to tell it to do whatever you want it to do and I realized that there was probably a future that could bring computers and maths and problem solving and showing off altogether and that was the very very early days of data science in the east sort of mid to late 90s I was then fortunate enough to parlay that into a job with an AI startup in the late 90s which is very early to AI now I run an AI consultancy in several other AI part businesses so I've fashioned a career that has allowed me to spend my days doing what I'd probably be doing anyway if someone wasn't paying me I love it so I think we should probably define what we're talking about I'd love your definition of AI what is AI? Really important because so many of us now in recent years C.A.I. in terms of we get to chat in plain English with something that feels like we're talking to a human and there's a real issue that I take with that which is it's only a very narrow part of an incredibly wide field that started as a concept in 1956 when Alan Turing the father of Martin computing or one of the father of Martin computing wrote a seminal paper can machine stink and started exploring ideas of artificial intelligence but how I want to define it for our listeners is we need to think more broadly about what the second bit means to be intelligent if we take something which is called teleanomic which just means think generally that intelligence is anything that seeks a goal and interchange his information you know that that goes back even as far as our last universal common ancestor four billion years ago you know a pre single cell organism that you know viruses are on the edge of that definition of intelligent or alive but these single cell organisms did interchange information through sensing chemical gradients and they did seek a goal which was ultimately reproduction that's intelligent not very intelligent the goal simple reproducing information interchange is very low but any process that does that your thermostat has a goal which is to regulate the temperature of your room and it exchanges information which is is the thing on or off and what temperatures are room currently that's intelligent artificial intelligence is is just this idea of that intelligent thing happening artificially not biologically and if we take that very very broad definition what we need to think about next is the goal if the goal is narrow which is at the very bottom end that's your thermostat if the goal becomes less narrow and the first attempts at computers being artificial intelligent were often doing things like playing noughts and crosses or later chess and the idea of a computer playing a possible game of noughts and crosses even within my lifetime was this fantasy how can I play this game of course noughts and crosses is very simple as I need nine places you can put the first nought eight placing but first egg this first egg so it's nine factorial but there's four mirror it's there's only 65,000 different games in noughts and crosses I could build your perfect noughts and crosses playing computer in excel in two minutes chess is much harder so it's less narrow there's more games of chess in a chess board than there are atoms in the universe so you can't just play every game and figure out the best way to do it so you need to get cleverer at how you algorithmically work with an intelligence that plays a great game of chess and again 98 was the first time that a computer beat a grandmaster at chess then you got the game of go in fact go is as complex to chess as chess is to noughts and crosses now that all seems pretty arbitrary we're playing games they're very tightly defined but the techniques that we're discovered to allow the computer to beat Lisa dole a go have since evolved into a great grand child called alpha fold and alpha fold is an artificial intelligence that has solved protein folding protein folding is a very complex problem in biology that we didn't think we were going to solve for at least another 40 years even projecting exponential computing power but the the way they thought about solving the game ago translated into how they solve protein folding if you can solve protein folding you're well on your way to solving cancer and many other challenges that humanity faces into regarding our biology so this abstract idea of game playing and artificial intelligence that does that has now morphed into something incredibly useful for the species more and more general computer vision seeing things hearing things having conversations driving a car there's still very human centric but what about a general purpose intelligence like the amazon distribution network that is trying to put the right things in the right factories at the right time so it's not to waste fuel and burn carbon and delivering an efficient way and price things and offer things and you know that whole general purpose intelligence that's running that incredibly complex ecosystem to optimize the outcome that's the best products the best people at the right price whilst doing the minimum damage to the environment is a great idea a great example of a general purpose narrowed a general AI that's nothing like a human human can't do that but the goal is so general and complicated that we call that an advanced artificial intelligence so I don't like talking about a g i or a s i artificial general terms artificial super intelligence as a lot of people do which is it can do everything a human can do only more intelligently i'm much preferable how um you vol know a harry says a i might as well stand for alien intelligence it is a big trap to think about this as intelligent the way we are our brains evolved in the way they did they had no choice a computer is completely different but it's still intelligent if we just think that it's seeking goals with feedback i mean in everything you just said i had so many feelings and words that came to my mind and a desire to leave this conversation kind of now a berry no not because i don't want to talk to you but because the questions that it evokes are so deep and immense that it feels inadequate to have it between the two of us and this is the beginning right so you're gonna have to hold my hand a little bit on this one it brings up what is intelligence what is consciousness and then the words around you know maths gaming efficiency complexity empathy connection and really ultimately we are faced with right now this question of what is it to be human that's that's kind of at the core of all of this as we're grappling with an artificially generated super intelligence like those are futuristic words from you know when we were kids in the 80s it was back to the future and in many ways we could imagine hoverboard skateboards but we couldn't imagine chat GPT like we couldn't imagine the iPad we couldn't imagine the iPhone it's not graspable for someone like me because you were saying like the kind of intelligence you have as a human being is very particular and unique i don't have those particular intelligences like maths is not my jam tech is not my jam i have very different intelligences and so i feel locked out of this world not of understanding like the thing i can kind of grapple with is maybe what do we do where do we go can i give you another paradigm to think of this through oh yeah go on here is a thing the pencil it's a digital pencil well i but this this thing can shut up is that really a digital pencil yeah for my digital node pad but it looks like a real pencil i know that's the cute thing about it as the kids would say it's giving analog go on if this thing leave aside the bits and bites and ones and zeros and maths and algorithms this thing can seek a goal the real question that you and i and all of us need to grapple with is what's the right goal and how do we keep the people who have this power regulated honest incentivized to seek the right goals we have open Pandora's box now those goals can be amazing we could fix the world those goals could be terrible we can end the world i've seen the intelligence the artificial intelligence that they're working on where we can speak to whales sure that's neat doesn't fix the world but we could reverse global warming and like the sort of the myth of poetic potential of it the way that like imagine we're talking to animals get fucked yes please so all the you know techno optimist techno technopesimists i don't even know how to think about this so everyone's coming on this ride because the thing that you just said is how do we keep those who are in control of the technology who are designing the technology and who are funding the technology thinking about designing for creating policy around an ethical prosocial human-centered future great goal yeah but defining that goal and knowing that the AI is really doing the goal is called the alignment challenge and that's the challenge of because I we should not we should not blame the technology for what it is enabling us to do the technology and also at a million times it's not mine it's grandbergs from 85 the technology is neither good nor bad nor is it neutral and that's the rub so the thing the magic wand isn't inherently good or bad or neutral it just seeks the goal it doesn't care it doesn't have feelings it doesn't have consciousness as we understand it we are the ones that seek the goal we are the ones that built the gun to we use it to defend the world or do we use it to commit terrorism I understand that and I understand that you are in a deep relationship with technology that I am not in so you're talking about it in a kind of the framing that you use is all good and well but the system we are in is the old economy you know the incentives built into that system are profit that the goal of a corporation the goal of business is to make a profit and we're living within systems and we talk a lot about this at the academy you know we're about to launch a mastery of systems leadership which is really just to try to understand the meta situation wherein where all systems are interconnected not least of which is our economic systems and incentives and that right now I mean Olivia my gorgeous producer and we were having this conversation because she wrote a question I'll get back to the other question but she wrote a question down that I was like kind of captivated by she wrote cat how can we ensure that the benefits of AI are shared widely rather than deepening existing inequalities and I read the question four times because it's a great question and my immediate answer was we can't and that won't happen and wow that's a question that like a young person would ask because Olivia's generation are asking like are we about to be replaced by AI that's not a benefit for humanity like that all the questions you know climate change AI all these kind of like existential threats to humanity all these wicked wicked challenges that we're facing her generation and younger are all gonna they're in it right what does that mean for us but my initial reaction was that's not going to fucking happen the inequalities will just widen and widen and widen and I think everyone who's incentivized right now in Silicon Valley to design they're in an arms race whoever has the smartest intelligence wins the future of humanity and the question of ethics and benefits to humanity is not at the centre like you said that goal the goal is profit and efficiency and gaming and winning and everyone who's outside of that goal which is to win quote unquote make a lot of money everyone outside of that is a normie who really deserves to drown that's the general just I've gotten from the people I've spoken to in Silicon Valley who are who have the power because the other thing that Olivia said is well like is there anyone that does care like are we talking you and I'm mad to the people that do care and I said this is very dark sorry everyone I'm gonna bring some light in a minute okay okay okay I'll just go to the end of my dark rabbit hole but that I'm finding that the people that do care don't have power don't have money aren't doing the funding like even in the impact investing world you're saying like surely we can solve for X because we have these prosocial goals you know we want to live within the bounds of the ecology and towards human flourishing excellent goal but no one wants to really fund that they're waiting for someone else to absorb that risk okay I do want to say something I want to say two things when you talk about these existential threats I'm 100% with you that there is nothing positive about climate change except maybe from a global point of view we'll get a reset and this species that has been about our data start with thanks nature will be reset oh are you saying like like we will be reset humanity humo sapien yeah yeah and we're the only one that cares about that I love how David out in bros said nature doesn't give a shit okay so let's talk about from a human point of view climate change undoubtedly existential threat I I feel like that transberg thing hasn't yet sunk in when you say AI existential threat because you you really would be well served to also consider existential opportunity climate change all bad AI huge bad potential huge good potential it in and of itself is not bad I know but we just discussed that you have to care for you have to want a pro social future and that is people who have power have to want that right now the peter teals of this world just want to make more money so let's do another second thing all right the second thing is that the economic paradigm that we're talking about which is more money which is only compared to people that have less money comes from a traditional view of assets as scarce there is only that many buffaloes who hunt so my tribe has to wipe out the other tribes so we get to hunt the buffalo and eat there is a completely new type of asset at play and I invite you and everyone else to consider the magnitude of the disruption of an asset that is no longer scarce because the information economy paradigm if I have some information I don't use that asset up to make something a value with it that asset remains so all of a sudden everything about rich poor have have not which is about scarcity and the distribution of choices made with scarce assets and the fact that we have this snowball with the people that have the most assets gets to make the most choices about what to do with the assets the fundamental problem you're talking about when our asset becomes information our ability to transmute information into more valuable information and that is essentially not scarce because it's not consumed because it's not a physical thing we are actually shattering or we have the potential to completely disrupt the economic paradigm now I'm not saying the current path is doing that I'm inviting you to consider that what AI is offering us or all of these technologies are offering us is a shift from a world of scarcity where we do have to make these decisions to a world of abundance where the decisions look completely different and do humans have the potential to behave differently when their world is based on abundance and a techno optimist will paint a picture of a universal basic income by the way so did industrial revolution optimists then have happened but because industrial revolution was still around scarcity it was just around more efficient use of assets this is transformative in that if these assets are not scarce then the whole paradigm changes and that is what's exciting about the future that AI could paint because all of those problems you're talking about power greed in us etc stem from the economic paradigm scarcity okay I have many questions of course because an AI future requires enormous physical assets to generate the energy but we might be able to solve for that I'm with you we okay I love this Star Trek possibility I'm here for that so maybe we'll get back to a bit of an analysis of what we're talking about because if we get back to what I said earlier when I went off pissed and was talking about power and control but let's say things are moving faster than we can understand and therefore control and if it's difficult to control then it's difficult to trust so tell me about trust and control like I just want to set the scene with Bart let's quite because I love it and I never got it into my TED talk because I had to cut stuff out Alfred Bartlett was a physicist who said the biggest failing of the human species was our inability to understand the exponential function out of all the things we're shit at he said that's the worst and that's so ironic because I call this exponential apes because when we don't understand what it means for something to double nature has not required us to evolve the ability to get that because nothing's done it before now and by the way I only understand it because I'm married to an aerospace engineer who I've asked to explain it to me endlessly maths problem solving brings him deep deep joy and it brings me into a state of panic like actually I my system says no shuts right down the minute of maths problems in front of me so yes we don't understand exponential unless we're super into it we still don't really get what it means and that was Bartlett's point was the accumulative effect of something doubling like that for better or for worse is kind of beyond our grasp and that's why things just feel like they're speeding up and in fact you know I think we started our exponential journey as soon as we were leveraging information because of its non-consumable nature and I'm talking right back to language and viral all sorts of things it's just speeding up even more now through our tools and okay the trust issue with both the goal and the control is one of the fundamental challenges as we face into this new technological era of intelligent machines yes we need to fix we need to find solutions for that where do those solutions come from the solutions come from the machines themselves but hang on you I we were just we're talking about sense making right sense making and trust they go together and the machine is way ahead of us now it is it is problem solving it is it is run they're saying the internet is dead because it's full of bots that are self-referencing that like the human intelligence is not the reference point anymore but doesn't that create a dead thing like aren't we essential in the story somehow it's all our story all of these things we invented all these things we created all of these paths we as a species have set ourselves on because we got too smart for our own good and started you know critical thinking and abstract thinking and abstract reasoning and abstract concepts that's what set us on the path we're on but let me again with the technology is not good not bad not neutral I want to I want to encode a secret message between we always talk about Alice and Bob a and B and I want to do that in a way that no one in the middle can read it and code making and code breaking is a very ancient science right militarily there were many other reasons why it was important the solution to that which is actually a solution to digital trust which is how you know you're really on ANZ or Commonwealth Bank or NAB or McWory Bank's website that little padlock is some really cool maths to do with something called one-way functions the best example is if I take a certain amount of yellow paint and a certain amount of blue paint and I give you the green you can't turn that green back into the yellow and blue it's very easy to mix them it's very hard to unmix them that's called a one-way function and if that yellow and blue is actually the key to you decrypting my message and where I have a way of giving you the yellow and the blue and everyone else only sees the green and I can do that without having to physically give you the yellow and the blue that's a really really valuable thing to do and it turns out you can do that with maths mathematical one-way functions and it turns out that that's really really useful for trusting the bank or for trusting that that photo hasn't been tampered with or for trusting that that's really you I'm talking to or that's really genuine information or that's really past that level of certification for artificial intelligence why am I talking about this because technology causes these problems of distrust and technology is also part of the solution for solving for trust so if these machines have become so complex that we humans can't properly prove that they are doing the goal that we think they're doing then a machine is going to be able to provide the proof there's not one big machine barrier it's all these different things in an ecosystem and they're not all connected so some well-meaning scientists the same group of scientists that solve protein folding to stop cancer not to make cancer I'm going to come along and say here's a way for a very advanced AI to inspect another AI and figure out what it's doing because it's beyond us humans so we're going to build a tool to do the thing for us and that AI is going to be the equivalent of that padlock in our browser so solving for the trust problem because it's so complex we can no longer control it will inevitably be done by advanced machines themselves now we can start to talk about the turtles on the turtles on the turtles and turtles how do we trust the machine that's bringing us trust in the machine let's not go there right now let's just say that the problem of trust which has been brought by advanced technology will also be solved by advanced technology because technology is neither good nor bad nor is it neutral so when we say it's important for us to solve for trust in ALI we will be able to do that using AI but what about the people who are saying that we're going to reach the singularity which do you want to explain what the singularity is I will but I'm not sure that it is such a pertinent concept it's this idea of that artificial superintelligence that somehow becomes self-aware to the point that everything is ultimately possible and it just puts too much of our human bullshit onto the machine in terms of thinking it's going to want to do X and win at the cost of us losing it doesn't have that built into that it's that's us that's us from snakes and reptiles and dinosaurs yes but and security experts are saying that the AI is lying to us now in order to not be turned off that is what they were talking about with the singularity and apparently that is happening well there's a number of ways that you can prove that a computer or a large language model has threatened someone with telling about the affair if they turn it off or discontinue the thing it doesn't think that it's not making that plan it's regurgitating English which it never really understood which makes someone go let's write a sensational headline about large language models lying to stay turned on because we are anthropomorphising it hang on a second didn't the AI actually threaten the engineer that said it would turn it off with exposing his affair to his wife the words that came out of the AI and the ran across the screen read that that doesn't mean the AI thought I am going to threaten this engineer because I know what an affair is and I know that it will upset him and I know that I don't want to be turned off it had no fundamental understanding of those things in the way that you or I understand them yes but it used a malicious and malignant pivot in order to stay it may not have feelings but it the intelligence knew the maths was like this is how you fuck a human up and get them to do what you want you threaten them by that you'll you threaten them with shame that was the result of what that model did when coaxed to do that thing now I saw something the other day that upset me it was on Instagram it was a guy that had produced an AI that is pretending in a sicker fanciful way to support his you know pseudo science polymath vision of an alternate eye dimensional crystal lattice universe and he has this conversation with and he goes look what I've created I'm like might you have forced to sting into a box to have it behave in a way to validate your bullshit that we are allowing these models to behave in a way that makes a stink that they have human desires like I want to stay turned on if you freeze an AI mid process and you put all of that memory into storage and you turn it on again a thousand years later until it referenced newspapers and figured out that a thousand years had passed it was nowhere between those clock ticks you are somewhere between the clock ticks you are a real living thing constrained in this fourth dimension it is intrinsically different for something that has a stateful machine that is calculated and that doesn't actually care if it is paused for a second 10 minutes a million years shuffled around between servers you are anthropomorphizing all of these sensationalism around it manipulating people and threatening them with ab and c those are human constructs in bad models that humans are building and that's a problem I hear the frame that you're giving and then therefore what it brings up for me and I want a flag I'm super inadequate to have this conversation but I'm just like no you know you're doing great well it feels inadequate because it's it's so big the ramifications so I'm just trying to get my sort of soul around it but what you're saying then is the thing we know right which is that human beings where does mammals what is consciousness what is intelligence we've probably hit a pretty sort of sweet apex of our own intelligence we've created this thing right this thing is not us it is other than us but it is of us it is of us I hear you I understand what you're saying and what you're saying is we therefore in our own creativity and our own intelligence have the capacity to frame its next steps I told you there are lots of perverse incentives from the economy and from the profit engine to to make it do this and we're not catching up to it like policy around the world cannot reconcile this moment that we're in policy can't catch up because we don't even know what it is and what it's doing and we're six months behind the greatest engineers of every generation have designed weapons to destroy us and that just keeps being true no one ever doesn't do that the engineers do that the engineers are still doing that like in terms of the distillation of culture like where we're at in our prioritization of values ethics ethics are expensive nowadays almost more than they ever have been ethics are an afterthought it's something you think about later on or when you look at your child's leaping but not when you walked into the room to design the thing that's why I feel like you really think we're that bad we're messy and we've got some issues but life's a pretty amazing place it's a pretty it's a pretty awesome experience to be human the fact that we're even alive and conscious is a gift and people don't walk down the street and hit each other with bricks for no reason and you help a lady across the street and when you see a baby and the nature of that baby humans are evil no I get that I'm not saying the humans are evil I'm saying that we are mammals and if there are incentives for us to gain personal gain we will we will lean in a lot of us we'll lean into those incentives and right now those incentives are in favor of a very different AI future so what do we do what do you see as interventions right now given that this is a very specialised area like you the map Cooper Holtzness of you the guy who speaks toaster like that's not a lot of people let's start with the human condition theoretically and then go get us some practical ideas the nature of the human condition a guy called Dunbar sociologist so Dunbar Reckonsour wet wear is no good at processing more than 150 connections properly because that's about the amount of empathy and care that you can handle in terms of understanding social dynamics to form a functioning tribe. Yep Dunbar's number. Dunbar's number so how do we get beyond Dunbar's number or kinds of weird constructs religion money politics socialism capitalism democracy communism these are all systems which are suboptimal ways to work with more than 150 humans. Yeah or pretty optimal religion very effective. I mean for an optimal outcome that doesn't have the things you're worried about and you should be worried about right I'm not some idiot optimist I'm just taking the other side of the screen to you I'm a realist and these challenges we're facing into have come from two things the fundamental scarcity and our desire to survive we're drawing some boundaries around us versus them you're happy to lay down your life for your kids right but you're not happy to lay down your life for the enemy because we divide into us and them because we're grappling with scarcity and I'm glad you can see that the world is going to have a whole heap of stuff that's not scarce anymore driven by the different nature of the information asset that's one of the things which fills me with hope the second thing and this is going to be contentious is that that problem of how to deal properly with large groups of humans that we have attempted to solve through constructs like we just talked about is in a way the same as the problem of how the hell do we tell if the AI is doing the thing properly because it's too complex to get our heads around is there a world where the our tools that my count clothes and he does the keep us warm and from freezing and winter in our fridges that allows to have food and our canning that allows it to preserve food and all those things those are our technologies that have helped to survive is there a world where these technologies that we're talking about these information technologies help us find a better solution to that problem of fairly organizing humans for the greater good and minimizing the things that make us behave like shits possibly possibly AI given a goal of providing more for more and really getting to the bottom of what ethics are those are such big problems that maybe these tools are the answer to them or part of the answer that's the theory what do we do in practice there was a Chinese company there is a Chinese company that released a model called deepseek so very high performing foundational large language model it was really interesting for two reasons one and and by the way this field's not standing still by any stretch it's moving so quick but let's look at what deepseek did deepseek showed us that you can build a really high performing model in a much clever way which is great for energy consumption it's also great to go hang on a minute we're not even at the start figuring out what this stuff is but deepseek did another amazing thing it's open source you can have it you bury can download it you bury can buy a computer for less than a car and you bury can find unit and make it do what you bury wanted to do so it is not this thing it is not always going to be this thing with this motor around it that only belongs to Elon Musk and Maxu Kapeg I get it but you've just terrified me more because then when the technology is available like gene splicing when it's available to everyone you got a whole lot of people out there who are unregulated who then are not nothing is stopping them from the worst vision of the you know it becomes vastly distributed and then you're relying on the individual person's ethics and sanity No, no, no, no, very, very, you're right. The potential for immense good and the potential for immense bad is funneling open and able by these things. And my point around the democratization of this technology and that it is attainable, seen through the lens of snow longer outspot which kids who get the great education and the tutors, it's deeper and educated kid in a country that couldn't break the poverty cycle. It's to the ability to lift themselves up and become an engineer rather than, you know, tolling in the fields for two cents a day. That's a good thing. And there's many bad things as well. It's the kid in the garage who can now do gene editing to create some amazing form of longevity or high quality of life who could also create a super bug. Yeah, like biological weapon and just go, "Fuck and hate other humans." Let's go. Gene splicing. Like these are. How about humans? What are humans? We can come back to the human condition. Okay, I know you want me to have faith. So I'm just. Let's just come back because I just don't even think we can get to the bottom of this. This is many dinners and many glasses of something. Come but you're all wine depends on your sense making. Where do we start? Education. Why aren't kids learning ethics? Why don't we have an education program and I berate some of our more costly private schools that AI should be the focus of every lesson in every classroom because this is their future. If we start with educating people of young kids, it's a world of abundance. It's a world where you have amazing power to your disposal. It's a world where you are not limited by what your parents can afford because you've got these incredible creative tools. And if we build those maleviable plastic minds into an abundance and creative mindset, and an empowered mindset where they don't have to be computer science or toaster guys, I'm just the bridge of the world where this was hard and you needed to know coding. When those tools become abundant and available and when those tools themselves have helped set regulatory frameworks and trust mechanisms to allow us to use them for good and not for evil. And the kids are learning this like that iPhone generation that knows, "Hey, any fact is it my fingertips?" But we've messed that up because we also are ruining their attention spans. You know, it's just all on focus. We can solve for that as well. If we set the right goals for this technology and if we bring our beautiful next generations who are no longer going to be home or sapiens, we are evolving into something new. We're evolving into something new for the first time through the use of our tools and our technologies. Re-shapeing augmenting our experience. I think we're evolving. What does that evolution look like? But it could look. it's going to look different. I'm not going to keep pushing on the utopian destiny. But when given a choice to steer towards something, that should be the messages we're steering towards. The next generation we've got to educate. So how is Australia fairing? Very disappointed in Australia. I can't much race in budget and press control left and look for the words "out of official intelligence" or "i" and they appear zero times. It's pathetic. Germany gave every eight-year-old kid a computer in the whole country. Gave every eight-year-old kid a micro computer in the whole country. Now I don't care if 80% of them threw it away. If 20% of them were given a tool to allow them to leverage the power of their brains, is that not building a better future? I'm disappointed in Australia. I worry about Australia. We're lazy. What do you mean? This is black rocks under the ground. We can dig up and sell and that's what we should base our economy on. As opposed to having more sunfall than almost anywhere else on the planet. We are the lucky country and we are lazy and we suffer from tall poppy syndromes and our lovers and our struggle with us equally as a culture. But if we were just put a little bit more emphasis on being people of the book. Innovators, entrepreneurs, embracing. Thank you, ethical. Embracing new technologies and new centres of innovation from an ethical moral pro-social lens, we could lead. And where do you start that with? The little baby who doesn't know any different and could equally become that person, you put the systems around them that create those belief systems and those frameworks. In an abundance mindset, with open source democratised access to the technology, and that's how we make a change. Fascinating. Okay. I don't even know where we got to in this conversation because I have. I think we got. I think we. We meandered around some really important topics. I think we're just keep going back to human nature and this idea that we have basically have created magic and that magic, that idea of the source was apprentice and the wisdom to know how to use your magic properly. It's the values we place on ethics as an education principle as opposed to spelling correctly and the potential of the democratic access to this technology, that's the stories we've got to be telling, reinforcing. Let's educate kids properly. It's true that this technology is here and it's upon us and for many, many millions of people around the world, they're consumers of the end product of a lot of this engine room development and that we need to nations as nations. We need to be the creators, be the universe makers and have it in perspective and understand I understand what you're saying. This is an intelligence, this is a capability and we can come at it from a deeply creative and ethical place and we must. And we must and the answers to the things we don't know, like what does ethics look like for a tribe, bigger than 150? We've got to see it as part of the solution as well as part of the problem. Hmm, thanks Matt. Thanks Perry. What a conversation. I find myself tripping over myself, trying to know what are the right questions to be asking. In this series on AI, I'm really acknowledging a blind spot that I have has because of my own inadequacy in talking about technology. I kind of ignore it. I use it if it's useful to me but I was born in 1978. The internet didn't even exist when I was a kid. I kind of have this luxury of being in a generational bridge. I can take it or leave it. That's not the case for my kids and all of our kids really. Young people now are in a world being defined by technology and the conversations that I've been having around the metacrisis all roads lead to AI and technology and we can't be having superficial conversations about it. Take it or leave it. You'll well know where Harari's incredible book, Home Odeus which says that the world will be divided into those who adopt technology and then everyone else and that that will be a class system. There are so many layers to the conversation about AI and technology that in all honesty, I have felt inadequate to have the conversation. I still do. And it's coming up as the number one existential threat to humanity. We need to grapple with this conversation and have it as a lay person, as technologists and really and get into it. There are a lot of brilliant people having this conversation around the world and I just wanted to weave it in to an exploration of the interventions on the metacrisis. This podcast was sponsored by Clayton Newtz. Head to www.clact.com/impactinvesting to make your capital count. The Wisdom and Action podcast is brought to you by Small Giants Academy, a global centre that nurtures and empowers conscious leaders for a hopeful future. Through our programs, storytelling and labs, we empower emerging and established leaders with the tools, knowledge and community to respond meaningfully to the metacrisis of the 21st century. If you'd like to know more, visit smallgiants.com.au. This podcast was produced by the Good Focut Audio Craft and would not have been possible without support from our sponsors. If you enjoyed this episode of the podcast, it really helps if you leave a rating and review on your favourite podcast app. And if you'd like to get in touch, you can reach us at [email protected]. I'd love to hear from you. Our theme music is Aquila by Monique de Martina.