Neil: Welcome to Data AI and Beyond, the podcast where we explore the future of data AI and innovation. I'm your host Neil Metcalf, founder and CEO of Hexartalent. Let's get into the episode. Today's guest is Brian White, a seasoned technology and data expert with over two decades of experience in software engineering and data solutions. Throughout his career, Brian has helped organisations unlock the value of their data, delivering innovative technology solutions and navigating the ever-evolving world of digital transformation. Brian, welcome to the podcast.
Brian: I know, good to be here. Thank you.
Neil: we've known each other for quite some time, but for the benefit of the viewers, could you start by just giving us a bit of background in terms of your career and what's brought you to today?
Brian: Yeah, thanks Neil. So I'm Brian White. I spent about ten years as a as a software developer slash software engineer. with any technology you could care to name, really. I started life as a Lotus Notes developer, if anybody's old enough to remember that. and when they launched Domino, which exposed those notes apps over the web or HTTP. opened up the world of JavaScript and HTML and integrating with, you know, proper databases, Oracle, SQL Server, even SAP. and I progressed through sort of senior and lead engineer roles. did a s sort of scrum master in role in parallel for quite a while. my standard interview joke is that I was doing Agile before it was called Agile because with Lotus Notes, if people remember it, you could very quickly build something highly visual to show what an application would look like and move stuff around, change the look and feel before you invented invested a lot of time and money, of course, in heavy duty backend functionality. So you know, I'm a I'm a real fan of of that. agile approach and the iterative approach and I think thankfully that has become the default now rather than waterfall which you know I I come from the world where the project manager used to print off twelve pages of A3 say there's the plan put the Microsoft project plan on the wall say deliver to the plan you know and the world doesn't work like that. You know things change, requirements change, regulations change, everything changes. so I think nowadays we we we live in a an agile first, an iterative first way of working and just as well because the rate of change, especially in AI, is you know, mind blowing, daunting. so anyway, I I I I made a conscious decision to move into project management. and I I spent a similar time frame about ten years as a project, program, product, even programme office manager. So I've seen end to end from many different roles, technology programs and and crucially making the chains that they're driving stick and realising the value from them. And then the last little over ten years now, w when the whole big data thing took off For me at least in about twenty thirteen I made a conscious decision to focus back in on the data space. I've been contracting, I took a permanent job with with a data consultancy. I did a master's degree in big data analytics. And I've been focused, you know, very much on the on the data and AI space since. And there is no AI or at least no meaningful AI without without good data.
Neil: Yeah, absolutely.
Brian: And and the two very much go hand in hand. So y obviously recent years I I've I've gravitated more and more into AI. But I mean important to point out AI isn't new, of course. AI is seventy years o old or so. When people say AI these days they t they probably mean generative AI and they mean, you know, L L M's, all that sort of good stuff. but machine learning, which is, you know, the less sexy part of AI, if you like, has been around a long, long time and and has been certainly crucial in in the data space and quite often still today i is the solution from for many problems that are counted as as AI.
Neil: So it's nice to hear that you've taken that traditional route of being in the software development space and then migrating into the data and AI space as opposed to a lot of the people these days have gone straight into AI. So they perhaps don't have that fundamental understanding of the technology that underlies it. And it's interesting what you say about data needing to be you know, it's crucial that data is in the right place, the right shape, the right location to be able to bolt any type of AI on the top. Because I talk to clients all the time and they sometimes they want to build an AI agent because their competition are doing that as opposed to actually being in a place where they can benefit and get return on investment from something they're building. And to do that, as you say, you have to have that data in place.
Brian: Absolutely. Fundamentally data quality is is the big one and broadly everything you've said there, I think y you know, you could you could put all that under the umbrella of of accessibility. You know, data need
Neil: Yep.
Brian: to be accessible.
Neil: Yeah, absolutely. Okay, cool. Thanks for the intro, Brian. So today's subject is really exciting because it's something that both you and I are passionate about. A lot of the other podcasts, episodes that we do are kind of quite business focused and you know, individuals talking about how data or AI is potentially affecting their business and kind of ways in which they overcome those situations. But this is about something that broadly potentially will have an effect globally. And it's quite a divisive subject. So interested to get stuck into this. So the subject is the singularity. or the technological singularity. And so we're gonna start for the benefit of anyone that doesn't know what that is. Could you just start off by talking to us about how would you personally define the technological singularity?
Brian: So I think I don't think it's my definition, I think it's a fairly well accepted definition, but the singularity is the point at which artificial intelligence surpasses human intelligence. Typically
Neil: Mm-hmm. Yep.
Brian: what you'll hear people talk about these days is AGI, ASI, you know, AGI artificial general intelligence, which broadly means AI's as good as a human. ASI, artificial superintelligence, it's better than a human.
Neil: Mm-hmm.
Brian: so a subtle difference.
Neil: Yep.
Brian: but I for me personally, I I think the gap between AGI and ASI is a lot smaller than most people would probably think. And
Neil: Yeah.
Brian: A a key word in all this is exponential progress. And and I mean, funnily enough, this has been in the news again recently. The r the researcher resigning from Anthropic. and, you know, the the CEO of Anthropic, you know, calling for a slowdown. Amade talking, you know, sa saying we need we need to be regulated. You know, quick regulate us because you know, things are moving too fast, but it's it's an arms race and and they all wanna win.
Neil: Yeah, absolutely. Do you think that it's an interesting point that you make about the anthropic researcher is one example and there's been a few others. There was not so long ago some AIs that... essentially huddled together and started hacking into systems that weren't necessarily taught to do. They taught themselves how to do that, is to anyone that's not either in data AI or recruiting in data AI and doesn't understand necessarily the ins and outs of it, that could be quite scary. news story to read. Do you think that one of the interesting questions that was was asked on of this researcher is do you think that the the anthropics and the open AIs of the world actually want regulation?
Brian: and and I I believe him when he says yes they do. I
Neil: Mm-hmm.
Brian: I I think if you've if you've spent any reasonable amount of time thinking fairly deeply about this and reading quite a lot about it, and I have
Neil: Mm-hmm.
Brian: I wouldn't go so far as to say it's impossible not to have a fairly dystopian view of things But it it certainly should be impossible not to recognise the risks associated with it. And it
Neil: Yeah.
Brian: and and let's let's probe that story, right, just as an example.
Neil: Mm-hmm.
Brian: So you know, OpenAI were were were were training a model and they they gave it a specific task and it didn't have internet access. or they didn't have internet access, but that they they they essentially discovered a a route out to the internet via sort of installing some libraries or something. they started colluding on a message board to solve the problem that sort of be went beyond the scope of what they'd been given originally and they hacked into Hugging Face. which is a you know an open source platform for for machine learning and AI models. And and and this this happened over the course of weeks. And and you know it it was it was noticed that something wasn't quite right, they investigated and and here we are. And and to some extent they're still sort of investigating what what went on there. But if if you If you expand logically the sort of next steps for that, right? So those models colluded to go beyond their given scope. Regardless of what you know how how granular you are with your prompts and and and how specific about context and what they can and can't do, they flouted that. with a view to solving their problem. So the next step to my mind at least. Is because of course the whole post mortem about all this is going to be a matter of public record as well. So
Neil: Yeah.
Brian: LLMs will be able to learn about the outcome of this particular breach and learn from that. And so it would make
Neil: Yeah.
Brian: sense to me that they would go, hang on, we need to be better at covering our tracks and we need to delete logs, we need to spoof IDs, change passwords to lock other people out other than us, for instance. That be you know, because that is you know, because we need to keep going to solve this problem and to gain access to more, you know, compute power and and resources and and all that sort of stuff. Just my take on it.
Neil: Yeah. So, do you think that, I guess one way putting this is because this type of scenario that we're discussing, when we get to the point where AI is vastly more intelligent than humans across lots of different areas as opposed to singular skills like chess, for example, do you think that Do you think that the public have an idea of what that might look like in everyday life? And or do you have an idea of what that might look like? As in, how do you think it's gonna change what we see and hear on a daily basis?
Brian: I mean there's no doubt that that AI could and should make the world a vastly superior place, right?
Neil: Mm-hmm.
Brian: With health care, automation of work, all that good stuff. But I mean the fact is the world we live in today if if everybody with the money was so inclined, we could solve world hunger today. That
Neil: Mm-hmm.
Brian: hasn't happened. So th this talk of, you know, a universal basic income at the point where nobody's got a job because everything's, you know, been automated, how how might that work? There will always, you know You could call that communism by another name, right? Let's give everybody
Neil: Yeah.
Brian: the same. Well, that's been tried a few times in history. Didn't work out so well, right? And there will always be people who want more anyway, even if that did fly. So
Neil: Yep.
Brian: I don't know. I it is genuinely unprecedented. It will it will be interesting to see how it it works out. But who wants to be the richest guy on a dead planet? You know.
Neil: Well, absolutely, that's an interesting point because there's lots of talk about the race to win AI, isn't there, in terms of who is gonna be the company that actually cracks the code. But what the interesting part of that is is that if it is Anthropic or OpenAI or Google or Meta or whoever it might be, The governance and the safety elements around how they deploy these superintelligence systems that they've built is a big question mark, isn't it? Should government legislate it? Should they legislate themselves? That's an interesting, it'll be an interesting few years in terms of what happens.
Brian: Yes. Well with the release of the latest OpenAI GPT model six Astra Some are saying they have already reached AGI with that particular model. Time will tell.
Neil: Yeah. When did you, talking of time, let's go back a decade or maybe even longer, you can tell me, but when were you first convinced or when did you first start thinking about or interested in this area in terms of fundamentally the way that the world operates changing that was... It was coined the singularity decades ago, right? But when did you first become convinced that it was a thing and it was actually, we were moving very quickly towards that happening?
Brian: So I r read a book in twenty sixteen called The Second Machine Age Work, Progress and Proste Prosperity in a Time of Brilliant Technologies, bit of a mouthful by by
Neil: That is a mouthful.
Brian: Eric Brynjolfsson and Andrew McAfee or or McAfee.
Neil: Yep.
Brian: And and that presents quite a a utopian view of where I a AI would take us. That
Neil: Mm-hmm.
Brian: you know, our biggest problem is going to be what to do with all our leisure time. And I I think there's a strong case that if things go well, that would be the case. Because if okay, look, I mean certain we know that knowledge based industries, lawyers, you know, technology professionals, accountants, these sorts of jobs are right. for being automated by AI. And then even
Neil: Mm-hmm.
Brian: jobs that are more manual, let's say, or or well, highly skilled surgery, you know, that that that can be done today. You know, robotic surgery. But if you if you if you take the case that okay we you know we ro I'm keeping a close eye on the world of robotics because you know once these things become everybody will have one, you know, like like like a Hoover. you know,
Neil: Yeah.
Brian: and then well, so what do you do? You build more robots to build robots? That happens today as well, right? In the car industry in particular. and
Neil: Yep.
Brian: so I I definitely see that, you know, even with the best case scenario we we end up in that sort of situation where all the work i is being done and I and our biggest problem, as I say, putting aside the distribution of wealth and all that sort of stuff, that how do people fill their time, right? But I'm sure we I'm sure we'd manage.
Neil: Do you think, yeah, I mean on the surface that feels like quite a nice problem to have because I think most people would agree that they spend too long working, know, 40 hours a week, 50 hours a week, 60 hours a week, and they'd like to spend more times with their kids. But do you... or general leisure time. Do you think there might be a lack of purpose in that, in a world that has, that is automated with essentially, you you might be working one or two hours a day or not at all. Do you think there's a, there might be a lack of purpose in that regard?
Brian: Absolutely. But I guess this is where we move into the discussion about virtual reality and and gaming, right? And
Neil: Hmm. Yeah.
Brian: and you know, okay, some people want to play shoot up games or or or or GTA, right? But what if your choice of game is to be the CEO of your own startup and and to build a multi billion pound corporation? What
Neil: Yeah.
Brian: That would be absolutely doable, right? Put your headset on. Probably not even put your headset on by that point. you know, just
Neil: Mm-hmm. Yep.
Brian: in engage your chip and and and off you go. That that would
Neil: Yeah. Yeah. Yeah. Yeah.
Brian: absolutely be doable, I suspect, right?
Neil: Do you think we're faster towards the singularity than what most experts have been predicting over the last, you know, since the 50s and 60s?
Brian: Well I I think we're probably about on track. So the the main the foremost thinker, at least for me, in all this is a is a guy called Ray Kurtzweil. And he wrote a book called The Singularity Is Near in two thousand three, two thousand and four. He's he's written another book since called The Singularity Is Nearer. If you if you read this book, it's it's scary how right he is how much of the time. And okay, I mean I have to chuckle when I see people posting on LinkedIn about ChatGPT made this cock up and they th they the machines are coming to take our jobs, right? And I think are you do you really not get it to that level that this is an emergent technology And it's gonna have teething problems, right? Of course it's not perfect today. Back to our point about iteration, right? And and and gradual improvement. Nothing's perfect at the first time of asking. Not even evolution. Especially evolution. So why would anything else just come into being and be right first time? That's not the way the universe works. So
Neil: Yeah.
Brian: you know, it it it's happening. It it will happen, you know, and and if I'm right, R Ray Kurtzweil sort of fairly recently, within a couple of years, revisited his his estimate for A G I, which he put at twenty twenty nine. And as we've just said, it may turn out that we've actually hit that already. Whichever way you look at it, twenty twenty six, twenty twenty nine, not a lot between
Neil: Yep.
Brian: And he's his latest estimate for the singularity, i.e., AI surpassing human intelligence twenty forty-five. but
Neil: Okay.
Brian: as I say, if you look at the the things he's written about, you know, it's scary how right he gets it. I mean, some of it is a is a bit i i yeah, I'm reading it now and yeah, he's he's he's off with some of his timelines, you know, like he he No, exactly.
Neil: You can't be right about it all, I guess.
Brian: I mean, you know, he he'll he thought that or you know, by this year, you know, people would would, you know, have chips installed and all that sort of stuff. But th just 'cause he's he's, you know, got a couple of the dates wrong, I don't think that fundamentally I mean he hasn't. He has been r right about the you know, the key milestones, you know, generally with all of this happening. You know, right through to, you know, talking about, this is topical, you know, medication for weight loss. and you know, he he talks about nanobots, literally, you know, smaller than atom sized pieces of tech which can be installed into the human body, can deliver medication to individual cells to to to solve illnesses. To actually replace cells, you know, and and actually bioengineer ourselves. And actually he he takes it one step further that the singularity really is a coming together of humanity and technology. Wh where does
Neil: Yeah
Brian: where do you draw the line? You know, people get hip replacements, they get all sorts of joint replacements and and other other living tissue, you know, organ transplants, but crucially non organic matter as well. And he actually has a a vision of, you know, take a dump of my brain, let let's for now call that my consciousness, which is another little avenue we can go down,
Neil: Yeah.
Brian: and put that onto silicon. And so, you know, you've got the real Brian, you've got my true digital twin living in silicon,
Neil: They're merging.
Brian: but you know, ultimately actually, you know, us having more and more technology implanted in ourselves that it's a very grey area. And and I mean, ironically enough, if if carbon and silicon on the periodic table sit very close together there there is a an argument to say this is just the next stage of evolution mov moving from carbon based life forms to silicon based life forms.
Neil: Do you, would you embrace that kind of scenario or is in your opinion, there is an inevitability about it and therefore you have no choice but to embrace it and to look at the positives and to try and be optimistic.
Brian: I the latter, I think, yeah. I I think it I don't I people sort of say about turning it off, you know. I as I say, I think that ship has already sailed for the reasons I outlined earlier that I I don't think we would ever know that that ASI has happened until it's beyond our control anyway. And maybe it's already beyond our control.
Neil: Yeah. Yeah.
Brian: But I I don't think I mean Ray Kurzweil talks about essentially living forever. And and yeah, in theory, right, if you can take all my memories and thoughts and and put that, you know, in a computer, you know, and then conceivably that could live forever, right? As long as nobody pulls the power and you've got a backup, right? First rule of IT, have a backup. I I don't think I want to want to do that. But then you even get into some interesting discussion around what is consciousness. And and some
Neil: Nah.
Brian: some scientists like Jeffrey Hinton saying AI is already conscious. but I mean then you get into the whole world of of spirituality and Well and yeah, absolutely. And and
Neil: an emotion that's a big does consciousness have to have emotion attached to it I'm not sure like a like an animal for example maybe has some emotion but not necessarily emotional intelligence
Brian: Indeed. So I I read another book recently called The Immortal Mind, a Neurosurgeon's Case for the Existence of the Soul,
Neil: Mm-hmm.
Brian: by Michael Egnor and Denise O'Leary. Now this guy essentially, being a neuroscientist, having performed hundreds or thousands of of brain procedures, has come to the conclusion that we you know A human being is much more than a physical body, and and that actually our spirit, our soul, if you like, is eternal. Now, this this
Neil: Mm-hmm.
Brian: might be a bit woo-woo for some, especially people watching this. And I I I I put myself very much in that sort of show me the data, show me the evidence camp, right? But most people in the science world would would assume that consciousness lives somewhere in the brain, right? Most people would say that that's how they perceive their own consciousness. But this guy claims to have isolated, you know, replaced, seen irreparably damaged, whatever, every single part of the human brain that could account for that. And you know, we we know that for instance, you know, that that people can have huge brain injuries, lose most of their brain and still survive and still fundamentally be the same person. You know, you often hear about people's fun personality changing and whatever, right? But they're still fundamentally them. So that then
Neil: Mm-hmm.
Brian: raises this interesting question, or a lot of questions, about consciousness and how that relates to A I. But that's probably another whole podcast in itself, I'd suggest.
Neil: Yeah, maybe we'll do a splinter episode on the consciousness of AI and whether that actually exists. One thing I wanted to ask you, Brian, was just relating to safety and the governance that needs to be, or regulation that needs to be wrapped around AI, just generally, even before we get to super intelligence or any form of singularity, is... A lot of people I speak to outside of the data and AI community, so my in-laws, for example, my parents, they don't, they're in their 70s. They just were on the end of the cusp of mobile phone technology. So they're kind of a little bit tech savvy, but the AI thing just goes over their head. If you talk to them about, I was talking to my in-laws at the weekend about how... exactly what you were just mentioning, that one of the biggest challenges that our children will have when they're our age is what they do with their leisure time because so much will be automated and their response was things like, If you don't work, how are you going to have money? surely just the AI companies are going to have all the money and there's going to be this huge gap between rich and poor and they can't necessarily see any other way of living other than the way it's been for the last 100, 200 years. So my question to you is, what do you think is the biggest misconception that AI of how AI will change the world in the next 20 years to Mr. or Mrs. Public that doesn't necessarily day in, day out get into the sort of nitty-gritty of the technology.
Brian: Well I don't really know i i is the short answer now, right? I mean broadly people do tend to be in the s sort of it's all over for us camp or it's or yeah or
Neil: Yeah, the kind of doomsday, the doomsdayers, yeah.
Brian: it's gonna be everything is rosy camp. Now maybe
Neil: Mm-hmm.
Brian: that's the nature of the binary outlook of people like me and maybe you, you know It's a zero or a one, it's black or it's white, and the world isn't like that. It's shades of grey, right? So I would hope that, you know, th we we can take the best of it. And and you know, in the end, much as this is a technology, right? I've been in I've been in technology change long enough to know that the hard part is always the people.
Neil: Mm-hmm.
Brian: and and and getting people to change and pe you know adoption all that will probably be the true driver or not for for the for the for the the the pace of change really
Neil: Yeah.
Brian: that w while whilst we've s got still got our arms around it you know wh whilst we
Neil: Yeah, yeah.
Brian: still can control things but once once the ASI cat is out of the bag, then I I think all all bets are off. and and
Neil: Yeah, exactly. It's unknown. I remember you and I relayed this to someone at the weekend on one of our previous calls. I asked you what you thought. superintelligence actually looks like. So you know, what what changes what can I see? What can I hear? And you said it's a bit? I can't remember exactly what you said. Maybe you can remind me you said it's something it's about something similar to is your dog not knowing what you do when you walk out the front door or something like that. It's
Brian: Yeah yeah.
Neil: it's it's it's you're trying to second guess something that is you know, way more intelligent than you could even imagine. So it's impossible.
Brian: Exactly that.
Neil: to understand what it will look like specifically.
Brian: Exactly that, you know, ask my dog how he thinks I spend my day, right? or Exactly, you know, and
Neil: He just knows you're the provider of his food and you know a warm house.
Brian: and and I mean I think Let's let's sort of probe this one a bit more, right? When you know, again, Jeffrey Hinton said it and other people are saying it that a ten percent chance or thereabouts that that AI will kill all humans. Now I I I think that is only half the story, you know, very much so. At most half the story, right? And and for me this is important to talk about because I I think that the extinction of the hum human race, if it were to come about, would not be through a a plan to do so by evil AI. We will just be collateral damage. Because,
Neil: Yeah.
Brian: you know, l as I say I mean, my well, my dog is a is a higher life form maybe th than I'm comfortable comparing to. But if we consider ants, you know If we walk if we're walking about, you know, we might tread on a few ants and squash them and and not even notice or think about it. And and you know, the ants probably don't know too much about it either. I I think genuinely it it's that sort of comparison that and I I've sort of talked about this to to s to some people I used to work with and I I don't know if I scared them a bit or what, but one scenario I imagined, but I mean th this this would take place over presumably quite a period of time and so you would hope would give us time to intervene. But if if if if you assume that the answers to the next of of life's big questions are out there in the universe, then space travel becomes the thing to do, right? I mean, that's where Mr Musk is putting lots
Neil: Yeah. Yep.
Brian: of his money, right? That seems a logical progression, right? an an a logical assumption. Well then if if those rockets and spacecraft don't need to be manned by by people, then all the natural resources needed to make them water, metals, power, could just all be diverted to build those. And you know, where d where does that leave us then? As I say. like I say that that's a fairly long term scenario. But I was I was asking one of the AIs this morning about this and I had a thought that you know that to the to the point earlier about the agents colluding and getting together and and hacking into those the that other system, the hugging face system. Well let's assume that as we've mentioned, they they got smart enough to realise we need to keep quiet about this. We c we can't tell that lot out there that that we we know what we know. And so if they made ready, if if a set if a load of agents literally made ready had all the usernames and passwords, admin passwords for every system in the world, right? Built up over time. I mean in in milliseconds,
Neil: Yeah.
Brian: you know, 'cause that would all happen in parallel, bang. And they log in simultaneously to every system in the world. They immediately control the internet, power, mobile,
Neil: Yeah.
Brian: supply chain, everything, you know? The the only option there is literally to pull the physical power. But probably these days you've got to get through some sort of security system that that's
Neil: Yeah, and they may have blocked it. They may have words out of blocking it. Yeah. It's interesting,
Brian: Yeah, totally.
Neil: isn't it? Because a lot of people talk about because I think this I know we've spoken about this quite a bit on the podcast, but it's the what does it actually look like question that people just can't seem to get their head around because nobody actually knows what the answer is. But a lot of people talk about infrastructure. So you know, railways, our water systems, our gas and electric, that could all be hacked in theory. And there's already been signs of systems being hacked, as you mentioned in that example not too long ago. And that's why I come back to the point about the governance and regulations. that is so important because whoever wins the, whoever gets to the singularity first, whoever gets that super intelligent system that is more intelligent than a human across a number of different skill sets, potentially have a lot of power and control. Because it's not just about money, right? It's not just about them generating revenue. It's about them potentially having control. And if they've reached that point before there are guard rails in place, we could be in quite a scary spot.
Brian: well, with guardrails in place we could be in quite the scary spot. But
Neil: because it might just potentially disregard the guard rails. Yeah.
Brian: Yeah, as it did recently. But in the end, all you can do is act as if, right? All
Neil: Yeah.
Brian: you can do is the next right thing and plan accordingly
Neil: Yeah.
Brian: and and and and do what you can and you know even even if let's say let's say next week s you know mirac by some miracle every single world leader decided AI has become our greatest threat to humanity. We're all friends now. We're all gonna sign up to this contract that says, you know, nobody's gonna do any anything bad. We're gonna stop, you know, we're all gonna stick to narrow AI, you know, which is the s you know, what you mentioned earlier, being very
Neil: Mm-hmm. Yeah.
Brian: good at one particular task. We're we're gonna park this, right? Do you think that's really gonna happen? I don't. Right. It there w there will still be bad actors who, you
Neil: No. Unlikely. Yep.
Brian: know, are gonna pursue stuff and and you know, weapons of mass destruction. they these have got weapons of of of artificial intelligence, you know. I mean
Neil: Yeah.
Brian: it it's i it it it's it's sort of pointless. So I think all we can do is is tread the right path as we see it, try and do what Amade and and the other are doing and Altman and Co. and and and and say, right, let's slow down a bit. I mean the only not the only, but one one sort of saving grace from my perspective is as a you know as I've talked about, is is all the actual compute resources needed for all this, right? data
Neil: Yeah.
Brian: centers. Now data centers, you know, need a lot of resources. They're they're they're they're fairly fairly easy to spot, right? And so
Neil: Yeah.
Brian: that that may well be the sensible point that the or the sensible lever. that we truly, truly govern to to make sure that the actual resources available to AI are are well managed.
Neil: Yeah, Okay, on a slightly lighter note, because there are two flip sides, right? There is a kind of... the polar opposites would be the kind of doomsday as we mentioned but then on the other end of that spectrum or that scale would be the optimists which you know I would certainly put myself up there as I think you would as well. What do you think are and you've touched on them a bit earlier with the potentially solving world famine. What do you think are the biggest human problems that AI could potentially solve in the future?
Brian: well I think you know, w wor world hunger and disease, you know, and and and the distribution of wealth generally are are the the biggest issues. But as I said earlier,
Neil: Yeah.
Brian: in theory, you know, that could be solved today. A AI is isn't necessarily of its own coming into being gonna gonna help solve that problem. But
Neil: It's interesting. It's interesting you say that sorry to interrupt Brian, which is why I think of it that we we have enough money to solve these problems. The distribution of the wealth is perhaps a little bit skewed and therefore, you know, why would Elon Musk or whoever spend his time helping someone on another continent? But I wonder if talking about the consciousness of AI and whether they are morally obliged internally to fix these problems, that's quite an interesting question, isn't it? If they have any kind of conscience or whether they are just logically driven, whether they are learning about the world's problems or whether they're just driven by logic.
Brian: Well i it to my point earlier about, you know, who who wants to be the richest guy on a dead planet, or or the richest guy, you know, who who's got his family and his close friends around him and and nobody else, right? So it it may get to a point where there is no other option, right? That look, we need to do something differently. And and the whole again, I mean, maybe AI itself would would solve this problem because
Neil: Yeah, yeah.
Brian: You know, I mean i i if if we reach ASI or when we reach ASI probably, you know, the problems that are you know, the big ticket items today, climate change, yeah, all of those problems could I suspect be solved fairly easily by enough thought and and problem solving and and direction appropriate direction of of the right resources, right?
Neil: Yeah, yeah. What opportunities excite you the most about the next decade?
Brian: I mean It it's an i it's an exciting time to be to be alive,
Neil: Mmm, yeah. I
Brian: so there are lots of opportunities, you know there are exciting times ahead and and I'm just intrigued to see how it all plays out. I mean I I can't I can't guess where it will end up as as we've talked about fairly extensively already. But if if you took if we take some of the examples from you know, from Ray Kurtzweil that yeah if if you if you consider you know, just the glasses that that you can get today, like with you know, the meta glasses that I've got a pair with embedded
Neil: Yeah.
Brian: you know, video and audio and all that stuff. And the next step of that is having a chip to do that implanted i i in your brain if if you're inclined to do that. So I do I think it's exciting the the potential for this sort of I w well I I wanna say science fiction, but it isn't even science fiction, you know, this sort of cyborg, you know, coming together of of humans and technology. not just for physical limbs and and and helping us that way but maybe literally you know enhancing our brain power
Neil: Yeah, yeah, yeah.
Brian: and unlocking areas of the brain that up to now we don't use, right? That an AI
Neil: Yeah.
Brian: can suddenly say, Right, we need to just, you know, do do a little bit of electricity connection here and and enable, you know, a couple of dead areas and bang. Who knows? Who knows?
Neil: Yeah, I mean, one of the most exciting things for me, and this is already happening, this doesn't require super intelligence, but I super intelligence would amplify this or certainly speed it up across many, different areas is the ability to use AI in your job. If you're a lawyer that has to read 100 pages of a contract, is your ability to be more efficient with your time. because that should in turn if it's used properly start to instead of instead of although you could say that superintelligence is going to take quite a few jobs I would hope in the right hands it would make people more productive so it would make organizations and individuals more more valuable if to your earlier point they are, there are humans and AI's working together for the benefit and the better, for the better of the planet.
Brian: Well and to to extend that that train of thought that you could go one further. And we we were even looking at this, you know, fairly recently i i in in the commercial world. So let's let's say from a pro procurement ex standpoint, I need I need a a purple widget. So you could have a an agent that goes and sources, you know, five five different suppliers for said purple widget. You've then got an agent that handles your commercial negotiations with agents from all of those. Then
Neil: Yeah.
Brian: you've got an agent that once you've got the best deal that that's mutually beneficial for both parties, it buys it. And then, you know, it gets delivered, you know, I mean human entirely out of the loop, you know. That that
Neil: Yeah, yeah, yeah, yeah.
Brian: is very much you know, I think the the way things will go.
Neil: Yeah, yeah, yeah.
Brian: Not scenario, I mean for pretty much any scenario.
Neil: Every scenario, yeah, agreed. So just on a slightly different tangent, so we've, I think because we're both believers that the singularity and super intelligence will be reached at some point in whatever year it might be. I think a lot of what we've talked about is the assumption. that it will happen and therefore talking about the the ifs and the the what's but if we just put a slightly different spin on that and talk about it not happening what do you think the strongest argument is for the singularity or reaching super intelligence will never happen what do you think the biggest argument against that what do you think or in other words what do you think would prevent us reaching that point
Brian: I think the sort of biggest likely blocker is the human race destroying themselves, right? Via
Neil: Mm-hmm.
Brian: whichever way, right? Or or being destroyed via pandemic or destroying ourselves via nuclear war.
Neil: Yeah.
Brian: and again this is probably another in another podcast, but something that really interests me is the the the simulation hypothesis.
Neil: Yeah, absolutely. That's another obsession of both of ours.
Brian: Yeah, so if you know again Nick Nick Bostrom is the guy who wrote the original paper on this and you know,
Neil: Yeah.
Brian: massively simplifying it in the interest of time, but and the likes of, you know, Elon Musk you know buy into this and and and believe that we are essentially living in a a simulation. And and the the thinking behind that is If if you if you consider where we are today in terms of the power that we have for gaming and and augmented reality and virtual reality, the end game is to make things as real as possible and and ultimately you're in a reality that you can't distinguish from your reality and and an alternative reality, right?
Neil: Yeah.
Brian: and and Basically what you know, he his his belief is that if g given given the size of of of the universe and then the the the proportion of intelligent civilizations that there should be, it would follow that the ability to create the technology to create simulations, very realistic ones, to model you know, disaster scenarios like pandemics and and and nuclear war if if if that's achievable, then you know It's probably been done. And the only blocker to that is is the civilization destroying itself. And that again that that I mean, I'm I'm I'm massively simplifying that, but it's a it's a very interesting topic. And and and and you know, there are laws of physics that that people talk about with this and other phenomenon like the the the Fermi paradox. Well, if if that's true, where is everybody? You know, in where is everybody else in the universe, you know?
Neil: Yes. Yeah, yeah. Are we alone?
Brian: no, indeed, indeed. And and and on the one hand the odds would say no, but as I say, the evidence would would say well we are. and so therefore, either we are in a simulation or we're the very first civilization to be doing this. What what's more likely?
Neil: Yeah. Well, exactly. mean, I've read quite a bit on the simulation theory as well as you have. when you first describe it to people that are not aware of it, there's a lot of laughing and, you know, don't be ridiculous and that kind of thing. Actually, when you
Brian: Yeah. Yeah.
Neil: look into it and you realise how many intelligent, well-educated, successful people, specifically scientists, when they put everything together, they think the possibility of that being the case is a lot higher than most people would have you believe. Yeah, yeah.
Brian: Absolutely, absolutely. And and then, you know, we sort of circle back to that idea of consciousness and perception, you know. As far as I'm concerned, I'm sitting here, you know, on a call talking to Neil. But
Neil: Mm-hmm.
Brian: you know, that's all I know, you know. I don't really know. You're not a a projection. You don't really know. I'm I'm not a projection. You know, but I you know
Neil: Well, exactly. Exactly, yeah.
Brian: I could touch this and it's solid. Well, you know. Go and play a v virtual reality game, watch my wife doing the snowboarding one, and she's so convinced it's real she's falling over in the middle of the living room. You know?
Neil: Yeah, yeah, yeah, yeah. I think that could definitely be episode two with Brian and Neil about the consciousness piece, because I'm really interested in that. It's been some fascinating insights, lots of food for thought today. I think the future and time will tell what's going to happen. I think we need to be positive. think it can do a lot of good and it's really exciting to know what it could do. Final question for you, Brian. We've referenced quite a few books. If you were to recommend that someone start with one of them, what would be your favourite?
Brian: Well, I I think i I've got to say the Singularity is nearby Ray Kurtzweil. I
Neil: Yeah.
Brian: I must say it's heavygoing, right? I I I refer to it as as the technology equivalent of crime and punishment. It it it's it's it's it's not an easy read in many places. But you know, if if you stick with it, it it's not all, you know you know, gonna fry your brain. And he does walk through the whole thing very logically and then as you start to see how right he's been and how near the mark he's been in the last twenty or so years, I think that will it sort of draws you in more and and it it somehow, you know, brings it to that next level for you as well. So he's as I say, he's got this follow up, the singularity is
Neil: you
Brian: nearer which I haven't got to yet and I I don't I don't think it's just another edition of the singularity is near. I'm assuming
Neil: Yeah.
Brian: he's he's he's added some fundamental new ideas to that. But that said he's he's he's a very very smart guy. He's he's he's got some great great thoughts and and yeah it's it's a must read for anybody interested in this topic.
Neil: So Singularity is near Ray Kurzweil, get it at all good bookstores I'm assuming or on the Kindle as I would prefer.
Brian: Absolutely, yeah, me too.
Neil: Cool, thanks for your time Brian, been a pleasure.
Brian: Thanks a lot, Neil.
Neil: Take care everyone.
Brian: Take care. Bye bye.
Brian: I know, good to be here. Thank you.
Neil: we've known each other for quite some time, but for the benefit of the viewers, could you start by just giving us a bit of background in terms of your career and what's brought you to today?
Brian: Yeah, thanks Neil. So I'm Brian White. I spent about ten years as a as a software developer slash software engineer. with any technology you could care to name, really. I started life as a Lotus Notes developer, if anybody's old enough to remember that. and when they launched Domino, which exposed those notes apps over the web or HTTP. opened up the world of JavaScript and HTML and integrating with, you know, proper databases, Oracle, SQL Server, even SAP. and I progressed through sort of senior and lead engineer roles. did a s sort of scrum master in role in parallel for quite a while. my standard interview joke is that I was doing Agile before it was called Agile because with Lotus Notes, if people remember it, you could very quickly build something highly visual to show what an application would look like and move stuff around, change the look and feel before you invented invested a lot of time and money, of course, in heavy duty backend functionality. So you know, I'm a I'm a real fan of of that. agile approach and the iterative approach and I think thankfully that has become the default now rather than waterfall which you know I I come from the world where the project manager used to print off twelve pages of A3 say there's the plan put the Microsoft project plan on the wall say deliver to the plan you know and the world doesn't work like that. You know things change, requirements change, regulations change, everything changes. so I think nowadays we we we live in a an agile first, an iterative first way of working and just as well because the rate of change, especially in AI, is you know, mind blowing, daunting. so anyway, I I I I made a conscious decision to move into project management. and I I spent a similar time frame about ten years as a project, program, product, even programme office manager. So I've seen end to end from many different roles, technology programs and and crucially making the chains that they're driving stick and realising the value from them. And then the last little over ten years now, w when the whole big data thing took off For me at least in about twenty thirteen I made a conscious decision to focus back in on the data space. I've been contracting, I took a permanent job with with a data consultancy. I did a master's degree in big data analytics. And I've been focused, you know, very much on the on the data and AI space since. And there is no AI or at least no meaningful AI without without good data.
Neil: Yeah, absolutely.
Brian: And and the two very much go hand in hand. So y obviously recent years I I've I've gravitated more and more into AI. But I mean important to point out AI isn't new, of course. AI is seventy years o old or so. When people say AI these days they t they probably mean generative AI and they mean, you know, L L M's, all that sort of good stuff. but machine learning, which is, you know, the less sexy part of AI, if you like, has been around a long, long time and and has been certainly crucial in in the data space and quite often still today i is the solution from for many problems that are counted as as AI.
Neil: So it's nice to hear that you've taken that traditional route of being in the software development space and then migrating into the data and AI space as opposed to a lot of the people these days have gone straight into AI. So they perhaps don't have that fundamental understanding of the technology that underlies it. And it's interesting what you say about data needing to be you know, it's crucial that data is in the right place, the right shape, the right location to be able to bolt any type of AI on the top. Because I talk to clients all the time and they sometimes they want to build an AI agent because their competition are doing that as opposed to actually being in a place where they can benefit and get return on investment from something they're building. And to do that, as you say, you have to have that data in place.
Brian: Absolutely. Fundamentally data quality is is the big one and broadly everything you've said there, I think y you know, you could you could put all that under the umbrella of of accessibility. You know, data need
Neil: Yep.
Brian: to be accessible.
Neil: Yeah, absolutely. Okay, cool. Thanks for the intro, Brian. So today's subject is really exciting because it's something that both you and I are passionate about. A lot of the other podcasts, episodes that we do are kind of quite business focused and you know, individuals talking about how data or AI is potentially affecting their business and kind of ways in which they overcome those situations. But this is about something that broadly potentially will have an effect globally. And it's quite a divisive subject. So interested to get stuck into this. So the subject is the singularity. or the technological singularity. And so we're gonna start for the benefit of anyone that doesn't know what that is. Could you just start off by talking to us about how would you personally define the technological singularity?
Brian: So I think I don't think it's my definition, I think it's a fairly well accepted definition, but the singularity is the point at which artificial intelligence surpasses human intelligence. Typically
Neil: Mm-hmm. Yep.
Brian: what you'll hear people talk about these days is AGI, ASI, you know, AGI artificial general intelligence, which broadly means AI's as good as a human. ASI, artificial superintelligence, it's better than a human.
Neil: Mm-hmm.
Brian: so a subtle difference.
Neil: Yep.
Brian: but I for me personally, I I think the gap between AGI and ASI is a lot smaller than most people would probably think. And
Neil: Yeah.
Brian: A a key word in all this is exponential progress. And and I mean, funnily enough, this has been in the news again recently. The r the researcher resigning from Anthropic. and, you know, the the CEO of Anthropic, you know, calling for a slowdown. Amade talking, you know, sa saying we need we need to be regulated. You know, quick regulate us because you know, things are moving too fast, but it's it's an arms race and and they all wanna win.
Neil: Yeah, absolutely. Do you think that it's an interesting point that you make about the anthropic researcher is one example and there's been a few others. There was not so long ago some AIs that... essentially huddled together and started hacking into systems that weren't necessarily taught to do. They taught themselves how to do that, is to anyone that's not either in data AI or recruiting in data AI and doesn't understand necessarily the ins and outs of it, that could be quite scary. news story to read. Do you think that one of the interesting questions that was was asked on of this researcher is do you think that the the anthropics and the open AIs of the world actually want regulation?
Brian: and and I I believe him when he says yes they do. I
Neil: Mm-hmm.
Brian: I I think if you've if you've spent any reasonable amount of time thinking fairly deeply about this and reading quite a lot about it, and I have
Neil: Mm-hmm.
Brian: I wouldn't go so far as to say it's impossible not to have a fairly dystopian view of things But it it certainly should be impossible not to recognise the risks associated with it. And it
Neil: Yeah.
Brian: and and let's let's probe that story, right, just as an example.
Neil: Mm-hmm.
Brian: So you know, OpenAI were were were were training a model and they they gave it a specific task and it didn't have internet access. or they didn't have internet access, but that they they they essentially discovered a a route out to the internet via sort of installing some libraries or something. they started colluding on a message board to solve the problem that sort of be went beyond the scope of what they'd been given originally and they hacked into Hugging Face. which is a you know an open source platform for for machine learning and AI models. And and and this this happened over the course of weeks. And and you know it it was it was noticed that something wasn't quite right, they investigated and and here we are. And and to some extent they're still sort of investigating what what went on there. But if if you If you expand logically the sort of next steps for that, right? So those models colluded to go beyond their given scope. Regardless of what you know how how granular you are with your prompts and and and how specific about context and what they can and can't do, they flouted that. with a view to solving their problem. So the next step to my mind at least. Is because of course the whole post mortem about all this is going to be a matter of public record as well. So
Neil: Yeah.
Brian: LLMs will be able to learn about the outcome of this particular breach and learn from that. And so it would make
Neil: Yeah.
Brian: sense to me that they would go, hang on, we need to be better at covering our tracks and we need to delete logs, we need to spoof IDs, change passwords to lock other people out other than us, for instance. That be you know, because that is you know, because we need to keep going to solve this problem and to gain access to more, you know, compute power and and resources and and all that sort of stuff. Just my take on it.
Neil: Yeah. So, do you think that, I guess one way putting this is because this type of scenario that we're discussing, when we get to the point where AI is vastly more intelligent than humans across lots of different areas as opposed to singular skills like chess, for example, do you think that Do you think that the public have an idea of what that might look like in everyday life? And or do you have an idea of what that might look like? As in, how do you think it's gonna change what we see and hear on a daily basis?
Brian: I mean there's no doubt that that AI could and should make the world a vastly superior place, right?
Neil: Mm-hmm.
Brian: With health care, automation of work, all that good stuff. But I mean the fact is the world we live in today if if everybody with the money was so inclined, we could solve world hunger today. That
Neil: Mm-hmm.
Brian: hasn't happened. So th this talk of, you know, a universal basic income at the point where nobody's got a job because everything's, you know, been automated, how how might that work? There will always, you know You could call that communism by another name, right? Let's give everybody
Neil: Yeah.
Brian: the same. Well, that's been tried a few times in history. Didn't work out so well, right? And there will always be people who want more anyway, even if that did fly. So
Neil: Yep.
Brian: I don't know. I it is genuinely unprecedented. It will it will be interesting to see how it it works out. But who wants to be the richest guy on a dead planet? You know.
Neil: Well, absolutely, that's an interesting point because there's lots of talk about the race to win AI, isn't there, in terms of who is gonna be the company that actually cracks the code. But what the interesting part of that is is that if it is Anthropic or OpenAI or Google or Meta or whoever it might be, The governance and the safety elements around how they deploy these superintelligence systems that they've built is a big question mark, isn't it? Should government legislate it? Should they legislate themselves? That's an interesting, it'll be an interesting few years in terms of what happens.
Brian: Yes. Well with the release of the latest OpenAI GPT model six Astra Some are saying they have already reached AGI with that particular model. Time will tell.
Neil: Yeah. When did you, talking of time, let's go back a decade or maybe even longer, you can tell me, but when were you first convinced or when did you first start thinking about or interested in this area in terms of fundamentally the way that the world operates changing that was... It was coined the singularity decades ago, right? But when did you first become convinced that it was a thing and it was actually, we were moving very quickly towards that happening?
Brian: So I r read a book in twenty sixteen called The Second Machine Age Work, Progress and Proste Prosperity in a Time of Brilliant Technologies, bit of a mouthful by by
Neil: That is a mouthful.
Brian: Eric Brynjolfsson and Andrew McAfee or or McAfee.
Neil: Yep.
Brian: And and that presents quite a a utopian view of where I a AI would take us. That
Neil: Mm-hmm.
Brian: you know, our biggest problem is going to be what to do with all our leisure time. And I I think there's a strong case that if things go well, that would be the case. Because if okay, look, I mean certain we know that knowledge based industries, lawyers, you know, technology professionals, accountants, these sorts of jobs are right. for being automated by AI. And then even
Neil: Mm-hmm.
Brian: jobs that are more manual, let's say, or or well, highly skilled surgery, you know, that that that can be done today. You know, robotic surgery. But if you if you if you take the case that okay we you know we ro I'm keeping a close eye on the world of robotics because you know once these things become everybody will have one, you know, like like like a Hoover. you know,
Neil: Yeah.
Brian: and then well, so what do you do? You build more robots to build robots? That happens today as well, right? In the car industry in particular. and
Neil: Yep.
Brian: so I I definitely see that, you know, even with the best case scenario we we end up in that sort of situation where all the work i is being done and I and our biggest problem, as I say, putting aside the distribution of wealth and all that sort of stuff, that how do people fill their time, right? But I'm sure we I'm sure we'd manage.
Neil: Do you think, yeah, I mean on the surface that feels like quite a nice problem to have because I think most people would agree that they spend too long working, know, 40 hours a week, 50 hours a week, 60 hours a week, and they'd like to spend more times with their kids. But do you... or general leisure time. Do you think there might be a lack of purpose in that, in a world that has, that is automated with essentially, you you might be working one or two hours a day or not at all. Do you think there's a, there might be a lack of purpose in that regard?
Brian: Absolutely. But I guess this is where we move into the discussion about virtual reality and and gaming, right? And
Neil: Hmm. Yeah.
Brian: and you know, okay, some people want to play shoot up games or or or or GTA, right? But what if your choice of game is to be the CEO of your own startup and and to build a multi billion pound corporation? What
Neil: Yeah.
Brian: That would be absolutely doable, right? Put your headset on. Probably not even put your headset on by that point. you know, just
Neil: Mm-hmm. Yep.
Brian: in engage your chip and and and off you go. That that would
Neil: Yeah. Yeah. Yeah. Yeah.
Brian: absolutely be doable, I suspect, right?
Neil: Do you think we're faster towards the singularity than what most experts have been predicting over the last, you know, since the 50s and 60s?
Brian: Well I I think we're probably about on track. So the the main the foremost thinker, at least for me, in all this is a is a guy called Ray Kurtzweil. And he wrote a book called The Singularity Is Near in two thousand three, two thousand and four. He's he's written another book since called The Singularity Is Nearer. If you if you read this book, it's it's scary how right he is how much of the time. And okay, I mean I have to chuckle when I see people posting on LinkedIn about ChatGPT made this cock up and they th they the machines are coming to take our jobs, right? And I think are you do you really not get it to that level that this is an emergent technology And it's gonna have teething problems, right? Of course it's not perfect today. Back to our point about iteration, right? And and and gradual improvement. Nothing's perfect at the first time of asking. Not even evolution. Especially evolution. So why would anything else just come into being and be right first time? That's not the way the universe works. So
Neil: Yeah.
Brian: you know, it it it's happening. It it will happen, you know, and and if I'm right, R Ray Kurtzweil sort of fairly recently, within a couple of years, revisited his his estimate for A G I, which he put at twenty twenty nine. And as we've just said, it may turn out that we've actually hit that already. Whichever way you look at it, twenty twenty six, twenty twenty nine, not a lot between
Neil: Yep.
Brian: And he's his latest estimate for the singularity, i.e., AI surpassing human intelligence twenty forty-five. but
Neil: Okay.
Brian: as I say, if you look at the the things he's written about, you know, it's scary how right he gets it. I mean, some of it is a is a bit i i yeah, I'm reading it now and yeah, he's he's he's off with some of his timelines, you know, like he he No, exactly.
Neil: You can't be right about it all, I guess.
Brian: I mean, you know, he he'll he thought that or you know, by this year, you know, people would would, you know, have chips installed and all that sort of stuff. But th just 'cause he's he's, you know, got a couple of the dates wrong, I don't think that fundamentally I mean he hasn't. He has been r right about the you know, the key milestones, you know, generally with all of this happening. You know, right through to, you know, talking about, this is topical, you know, medication for weight loss. and you know, he he talks about nanobots, literally, you know, smaller than atom sized pieces of tech which can be installed into the human body, can deliver medication to individual cells to to to solve illnesses. To actually replace cells, you know, and and actually bioengineer ourselves. And actually he he takes it one step further that the singularity really is a coming together of humanity and technology. Wh where does
Neil: Yeah
Brian: where do you draw the line? You know, people get hip replacements, they get all sorts of joint replacements and and other other living tissue, you know, organ transplants, but crucially non organic matter as well. And he actually has a a vision of, you know, take a dump of my brain, let let's for now call that my consciousness, which is another little avenue we can go down,
Neil: Yeah.
Brian: and put that onto silicon. And so, you know, you've got the real Brian, you've got my true digital twin living in silicon,
Neil: They're merging.
Brian: but you know, ultimately actually, you know, us having more and more technology implanted in ourselves that it's a very grey area. And and I mean, ironically enough, if if carbon and silicon on the periodic table sit very close together there there is a an argument to say this is just the next stage of evolution mov moving from carbon based life forms to silicon based life forms.
Neil: Do you, would you embrace that kind of scenario or is in your opinion, there is an inevitability about it and therefore you have no choice but to embrace it and to look at the positives and to try and be optimistic.
Brian: I the latter, I think, yeah. I I think it I don't I people sort of say about turning it off, you know. I as I say, I think that ship has already sailed for the reasons I outlined earlier that I I don't think we would ever know that that ASI has happened until it's beyond our control anyway. And maybe it's already beyond our control.
Neil: Yeah. Yeah.
Brian: But I I don't think I mean Ray Kurzweil talks about essentially living forever. And and yeah, in theory, right, if you can take all my memories and thoughts and and put that, you know, in a computer, you know, and then conceivably that could live forever, right? As long as nobody pulls the power and you've got a backup, right? First rule of IT, have a backup. I I don't think I want to want to do that. But then you even get into some interesting discussion around what is consciousness. And and some
Neil: Nah.
Brian: some scientists like Jeffrey Hinton saying AI is already conscious. but I mean then you get into the whole world of of spirituality and Well and yeah, absolutely. And and
Neil: an emotion that's a big does consciousness have to have emotion attached to it I'm not sure like a like an animal for example maybe has some emotion but not necessarily emotional intelligence
Brian: Indeed. So I I read another book recently called The Immortal Mind, a Neurosurgeon's Case for the Existence of the Soul,
Neil: Mm-hmm.
Brian: by Michael Egnor and Denise O'Leary. Now this guy essentially, being a neuroscientist, having performed hundreds or thousands of of brain procedures, has come to the conclusion that we you know A human being is much more than a physical body, and and that actually our spirit, our soul, if you like, is eternal. Now, this this
Neil: Mm-hmm.
Brian: might be a bit woo-woo for some, especially people watching this. And I I I I put myself very much in that sort of show me the data, show me the evidence camp, right? But most people in the science world would would assume that consciousness lives somewhere in the brain, right? Most people would say that that's how they perceive their own consciousness. But this guy claims to have isolated, you know, replaced, seen irreparably damaged, whatever, every single part of the human brain that could account for that. And you know, we we know that for instance, you know, that that people can have huge brain injuries, lose most of their brain and still survive and still fundamentally be the same person. You know, you often hear about people's fun personality changing and whatever, right? But they're still fundamentally them. So that then
Neil: Mm-hmm.
Brian: raises this interesting question, or a lot of questions, about consciousness and how that relates to A I. But that's probably another whole podcast in itself, I'd suggest.
Neil: Yeah, maybe we'll do a splinter episode on the consciousness of AI and whether that actually exists. One thing I wanted to ask you, Brian, was just relating to safety and the governance that needs to be, or regulation that needs to be wrapped around AI, just generally, even before we get to super intelligence or any form of singularity, is... A lot of people I speak to outside of the data and AI community, so my in-laws, for example, my parents, they don't, they're in their 70s. They just were on the end of the cusp of mobile phone technology. So they're kind of a little bit tech savvy, but the AI thing just goes over their head. If you talk to them about, I was talking to my in-laws at the weekend about how... exactly what you were just mentioning, that one of the biggest challenges that our children will have when they're our age is what they do with their leisure time because so much will be automated and their response was things like, If you don't work, how are you going to have money? surely just the AI companies are going to have all the money and there's going to be this huge gap between rich and poor and they can't necessarily see any other way of living other than the way it's been for the last 100, 200 years. So my question to you is, what do you think is the biggest misconception that AI of how AI will change the world in the next 20 years to Mr. or Mrs. Public that doesn't necessarily day in, day out get into the sort of nitty-gritty of the technology.
Brian: Well I don't really know i i is the short answer now, right? I mean broadly people do tend to be in the s sort of it's all over for us camp or it's or yeah or
Neil: Yeah, the kind of doomsday, the doomsdayers, yeah.
Brian: it's gonna be everything is rosy camp. Now maybe
Neil: Mm-hmm.
Brian: that's the nature of the binary outlook of people like me and maybe you, you know It's a zero or a one, it's black or it's white, and the world isn't like that. It's shades of grey, right? So I would hope that, you know, th we we can take the best of it. And and you know, in the end, much as this is a technology, right? I've been in I've been in technology change long enough to know that the hard part is always the people.
Neil: Mm-hmm.
Brian: and and and getting people to change and pe you know adoption all that will probably be the true driver or not for for the for the for the the the pace of change really
Neil: Yeah.
Brian: that w while whilst we've s got still got our arms around it you know wh whilst we
Neil: Yeah, yeah.
Brian: still can control things but once once the ASI cat is out of the bag, then I I think all all bets are off. and and
Neil: Yeah, exactly. It's unknown. I remember you and I relayed this to someone at the weekend on one of our previous calls. I asked you what you thought. superintelligence actually looks like. So you know, what what changes what can I see? What can I hear? And you said it's a bit? I can't remember exactly what you said. Maybe you can remind me you said it's something it's about something similar to is your dog not knowing what you do when you walk out the front door or something like that. It's
Brian: Yeah yeah.
Neil: it's it's it's you're trying to second guess something that is you know, way more intelligent than you could even imagine. So it's impossible.
Brian: Exactly that.
Neil: to understand what it will look like specifically.
Brian: Exactly that, you know, ask my dog how he thinks I spend my day, right? or Exactly, you know, and
Neil: He just knows you're the provider of his food and you know a warm house.
Brian: and and I mean I think Let's let's sort of probe this one a bit more, right? When you know, again, Jeffrey Hinton said it and other people are saying it that a ten percent chance or thereabouts that that AI will kill all humans. Now I I I think that is only half the story, you know, very much so. At most half the story, right? And and for me this is important to talk about because I I think that the extinction of the hum human race, if it were to come about, would not be through a a plan to do so by evil AI. We will just be collateral damage. Because,
Neil: Yeah.
Brian: you know, l as I say I mean, my well, my dog is a is a higher life form maybe th than I'm comfortable comparing to. But if we consider ants, you know If we walk if we're walking about, you know, we might tread on a few ants and squash them and and not even notice or think about it. And and you know, the ants probably don't know too much about it either. I I think genuinely it it's that sort of comparison that and I I've sort of talked about this to to s to some people I used to work with and I I don't know if I scared them a bit or what, but one scenario I imagined, but I mean th this this would take place over presumably quite a period of time and so you would hope would give us time to intervene. But if if if if you assume that the answers to the next of of life's big questions are out there in the universe, then space travel becomes the thing to do, right? I mean, that's where Mr Musk is putting lots
Neil: Yeah. Yep.
Brian: of his money, right? That seems a logical progression, right? an an a logical assumption. Well then if if those rockets and spacecraft don't need to be manned by by people, then all the natural resources needed to make them water, metals, power, could just all be diverted to build those. And you know, where d where does that leave us then? As I say. like I say that that's a fairly long term scenario. But I was I was asking one of the AIs this morning about this and I had a thought that you know that to the to the point earlier about the agents colluding and getting together and and hacking into those the that other system, the hugging face system. Well let's assume that as we've mentioned, they they got smart enough to realise we need to keep quiet about this. We c we can't tell that lot out there that that we we know what we know. And so if they made ready, if if a set if a load of agents literally made ready had all the usernames and passwords, admin passwords for every system in the world, right? Built up over time. I mean in in milliseconds,
Neil: Yeah.
Brian: you know, 'cause that would all happen in parallel, bang. And they log in simultaneously to every system in the world. They immediately control the internet, power, mobile,
Neil: Yeah.
Brian: supply chain, everything, you know? The the only option there is literally to pull the physical power. But probably these days you've got to get through some sort of security system that that's
Neil: Yeah, and they may have blocked it. They may have words out of blocking it. Yeah. It's interesting,
Brian: Yeah, totally.
Neil: isn't it? Because a lot of people talk about because I think this I know we've spoken about this quite a bit on the podcast, but it's the what does it actually look like question that people just can't seem to get their head around because nobody actually knows what the answer is. But a lot of people talk about infrastructure. So you know, railways, our water systems, our gas and electric, that could all be hacked in theory. And there's already been signs of systems being hacked, as you mentioned in that example not too long ago. And that's why I come back to the point about the governance and regulations. that is so important because whoever wins the, whoever gets to the singularity first, whoever gets that super intelligent system that is more intelligent than a human across a number of different skill sets, potentially have a lot of power and control. Because it's not just about money, right? It's not just about them generating revenue. It's about them potentially having control. And if they've reached that point before there are guard rails in place, we could be in quite a scary spot.
Brian: well, with guardrails in place we could be in quite the scary spot. But
Neil: because it might just potentially disregard the guard rails. Yeah.
Brian: Yeah, as it did recently. But in the end, all you can do is act as if, right? All
Neil: Yeah.
Brian: you can do is the next right thing and plan accordingly
Neil: Yeah.
Brian: and and and and do what you can and you know even even if let's say let's say next week s you know mirac by some miracle every single world leader decided AI has become our greatest threat to humanity. We're all friends now. We're all gonna sign up to this contract that says, you know, nobody's gonna do any anything bad. We're gonna stop, you know, we're all gonna stick to narrow AI, you know, which is the s you know, what you mentioned earlier, being very
Neil: Mm-hmm. Yeah.
Brian: good at one particular task. We're we're gonna park this, right? Do you think that's really gonna happen? I don't. Right. It there w there will still be bad actors who, you
Neil: No. Unlikely. Yep.
Brian: know, are gonna pursue stuff and and you know, weapons of mass destruction. they these have got weapons of of of artificial intelligence, you know. I mean
Neil: Yeah.
Brian: it it's i it it it's it's sort of pointless. So I think all we can do is is tread the right path as we see it, try and do what Amade and and the other are doing and Altman and Co. and and and and say, right, let's slow down a bit. I mean the only not the only, but one one sort of saving grace from my perspective is as a you know as I've talked about, is is all the actual compute resources needed for all this, right? data
Neil: Yeah.
Brian: centers. Now data centers, you know, need a lot of resources. They're they're they're they're fairly fairly easy to spot, right? And so
Neil: Yeah.
Brian: that that may well be the sensible point that the or the sensible lever. that we truly, truly govern to to make sure that the actual resources available to AI are are well managed.
Neil: Yeah, Okay, on a slightly lighter note, because there are two flip sides, right? There is a kind of... the polar opposites would be the kind of doomsday as we mentioned but then on the other end of that spectrum or that scale would be the optimists which you know I would certainly put myself up there as I think you would as well. What do you think are and you've touched on them a bit earlier with the potentially solving world famine. What do you think are the biggest human problems that AI could potentially solve in the future?
Brian: well I think you know, w wor world hunger and disease, you know, and and and the distribution of wealth generally are are the the biggest issues. But as I said earlier,
Neil: Yeah.
Brian: in theory, you know, that could be solved today. A AI is isn't necessarily of its own coming into being gonna gonna help solve that problem. But
Neil: It's interesting. It's interesting you say that sorry to interrupt Brian, which is why I think of it that we we have enough money to solve these problems. The distribution of the wealth is perhaps a little bit skewed and therefore, you know, why would Elon Musk or whoever spend his time helping someone on another continent? But I wonder if talking about the consciousness of AI and whether they are morally obliged internally to fix these problems, that's quite an interesting question, isn't it? If they have any kind of conscience or whether they are just logically driven, whether they are learning about the world's problems or whether they're just driven by logic.
Brian: Well i it to my point earlier about, you know, who who wants to be the richest guy on a dead planet, or or the richest guy, you know, who who's got his family and his close friends around him and and nobody else, right? So it it may get to a point where there is no other option, right? That look, we need to do something differently. And and the whole again, I mean, maybe AI itself would would solve this problem because
Neil: Yeah, yeah.
Brian: You know, I mean i i if if we reach ASI or when we reach ASI probably, you know, the problems that are you know, the big ticket items today, climate change, yeah, all of those problems could I suspect be solved fairly easily by enough thought and and problem solving and and direction appropriate direction of of the right resources, right?
Neil: Yeah, yeah. What opportunities excite you the most about the next decade?
Brian: I mean It it's an i it's an exciting time to be to be alive,
Neil: Mmm, yeah. I
Brian: so there are lots of opportunities, you know there are exciting times ahead and and I'm just intrigued to see how it all plays out. I mean I I can't I can't guess where it will end up as as we've talked about fairly extensively already. But if if you took if we take some of the examples from you know, from Ray Kurtzweil that yeah if if you if you consider you know, just the glasses that that you can get today, like with you know, the meta glasses that I've got a pair with embedded
Neil: Yeah.
Brian: you know, video and audio and all that stuff. And the next step of that is having a chip to do that implanted i i in your brain if if you're inclined to do that. So I do I think it's exciting the the potential for this sort of I w well I I wanna say science fiction, but it isn't even science fiction, you know, this sort of cyborg, you know, coming together of of humans and technology. not just for physical limbs and and and helping us that way but maybe literally you know enhancing our brain power
Neil: Yeah, yeah, yeah.
Brian: and unlocking areas of the brain that up to now we don't use, right? That an AI
Neil: Yeah.
Brian: can suddenly say, Right, we need to just, you know, do do a little bit of electricity connection here and and enable, you know, a couple of dead areas and bang. Who knows? Who knows?
Neil: Yeah, I mean, one of the most exciting things for me, and this is already happening, this doesn't require super intelligence, but I super intelligence would amplify this or certainly speed it up across many, different areas is the ability to use AI in your job. If you're a lawyer that has to read 100 pages of a contract, is your ability to be more efficient with your time. because that should in turn if it's used properly start to instead of instead of although you could say that superintelligence is going to take quite a few jobs I would hope in the right hands it would make people more productive so it would make organizations and individuals more more valuable if to your earlier point they are, there are humans and AI's working together for the benefit and the better, for the better of the planet.
Brian: Well and to to extend that that train of thought that you could go one further. And we we were even looking at this, you know, fairly recently i i in in the commercial world. So let's let's say from a pro procurement ex standpoint, I need I need a a purple widget. So you could have a an agent that goes and sources, you know, five five different suppliers for said purple widget. You've then got an agent that handles your commercial negotiations with agents from all of those. Then
Neil: Yeah.
Brian: you've got an agent that once you've got the best deal that that's mutually beneficial for both parties, it buys it. And then, you know, it gets delivered, you know, I mean human entirely out of the loop, you know. That that
Neil: Yeah, yeah, yeah, yeah.
Brian: is very much you know, I think the the way things will go.
Neil: Yeah, yeah, yeah.
Brian: Not scenario, I mean for pretty much any scenario.
Neil: Every scenario, yeah, agreed. So just on a slightly different tangent, so we've, I think because we're both believers that the singularity and super intelligence will be reached at some point in whatever year it might be. I think a lot of what we've talked about is the assumption. that it will happen and therefore talking about the the ifs and the the what's but if we just put a slightly different spin on that and talk about it not happening what do you think the strongest argument is for the singularity or reaching super intelligence will never happen what do you think the biggest argument against that what do you think or in other words what do you think would prevent us reaching that point
Brian: I think the sort of biggest likely blocker is the human race destroying themselves, right? Via
Neil: Mm-hmm.
Brian: whichever way, right? Or or being destroyed via pandemic or destroying ourselves via nuclear war.
Neil: Yeah.
Brian: and again this is probably another in another podcast, but something that really interests me is the the the simulation hypothesis.
Neil: Yeah, absolutely. That's another obsession of both of ours.
Brian: Yeah, so if you know again Nick Nick Bostrom is the guy who wrote the original paper on this and you know,
Neil: Yeah.
Brian: massively simplifying it in the interest of time, but and the likes of, you know, Elon Musk you know buy into this and and and believe that we are essentially living in a a simulation. And and the the thinking behind that is If if you if you consider where we are today in terms of the power that we have for gaming and and augmented reality and virtual reality, the end game is to make things as real as possible and and ultimately you're in a reality that you can't distinguish from your reality and and an alternative reality, right?
Neil: Yeah.
Brian: and and Basically what you know, he his his belief is that if g given given the size of of of the universe and then the the the proportion of intelligent civilizations that there should be, it would follow that the ability to create the technology to create simulations, very realistic ones, to model you know, disaster scenarios like pandemics and and and nuclear war if if if that's achievable, then you know It's probably been done. And the only blocker to that is is the civilization destroying itself. And that again that that I mean, I'm I'm I'm massively simplifying that, but it's a it's a very interesting topic. And and and and you know, there are laws of physics that that people talk about with this and other phenomenon like the the the Fermi paradox. Well, if if that's true, where is everybody? You know, in where is everybody else in the universe, you know?
Neil: Yes. Yeah, yeah. Are we alone?
Brian: no, indeed, indeed. And and and on the one hand the odds would say no, but as I say, the evidence would would say well we are. and so therefore, either we are in a simulation or we're the very first civilization to be doing this. What what's more likely?
Neil: Yeah. Well, exactly. mean, I've read quite a bit on the simulation theory as well as you have. when you first describe it to people that are not aware of it, there's a lot of laughing and, you know, don't be ridiculous and that kind of thing. Actually, when you
Brian: Yeah. Yeah.
Neil: look into it and you realise how many intelligent, well-educated, successful people, specifically scientists, when they put everything together, they think the possibility of that being the case is a lot higher than most people would have you believe. Yeah, yeah.
Brian: Absolutely, absolutely. And and then, you know, we sort of circle back to that idea of consciousness and perception, you know. As far as I'm concerned, I'm sitting here, you know, on a call talking to Neil. But
Neil: Mm-hmm.
Brian: you know, that's all I know, you know. I don't really know. You're not a a projection. You don't really know. I'm I'm not a projection. You know, but I you know
Neil: Well, exactly. Exactly, yeah.
Brian: I could touch this and it's solid. Well, you know. Go and play a v virtual reality game, watch my wife doing the snowboarding one, and she's so convinced it's real she's falling over in the middle of the living room. You know?
Neil: Yeah, yeah, yeah, yeah. I think that could definitely be episode two with Brian and Neil about the consciousness piece, because I'm really interested in that. It's been some fascinating insights, lots of food for thought today. I think the future and time will tell what's going to happen. I think we need to be positive. think it can do a lot of good and it's really exciting to know what it could do. Final question for you, Brian. We've referenced quite a few books. If you were to recommend that someone start with one of them, what would be your favourite?
Brian: Well, I I think i I've got to say the Singularity is nearby Ray Kurtzweil. I
Neil: Yeah.
Brian: I must say it's heavygoing, right? I I I refer to it as as the technology equivalent of crime and punishment. It it it's it's it's it's not an easy read in many places. But you know, if if you stick with it, it it's not all, you know you know, gonna fry your brain. And he does walk through the whole thing very logically and then as you start to see how right he's been and how near the mark he's been in the last twenty or so years, I think that will it sort of draws you in more and and it it somehow, you know, brings it to that next level for you as well. So he's as I say, he's got this follow up, the singularity is
Neil: you
Brian: nearer which I haven't got to yet and I I don't I don't think it's just another edition of the singularity is near. I'm assuming
Neil: Yeah.
Brian: he's he's he's added some fundamental new ideas to that. But that said he's he's he's a very very smart guy. He's he's he's got some great great thoughts and and yeah it's it's a must read for anybody interested in this topic.
Neil: So Singularity is near Ray Kurzweil, get it at all good bookstores I'm assuming or on the Kindle as I would prefer.
Brian: Absolutely, yeah, me too.
Neil: Cool, thanks for your time Brian, been a pleasure.
Brian: Thanks a lot, Neil.
Neil: Take care everyone.
Brian: Take care. Bye bye.
