[BONUS]: Embracing Nuance and Complexity in the Age of Data Analytics – [FULL INTERVIEW]

0:02
So hi everyone, Paul Sweeney and I got into the end of our conversation the other day. And we were really just chatting already for an hour. And so we decided just to cut it off then for the main interview, but we also decided then to just keep the cameras rolling. And so this clip here is welcome to the second part of our conversation, where we talk about much wider topics than just AI and how that’s affecting us in the collections market. But also talk about much broader topics aside, such as what artificial intelligence means, the wider world, the earth environment, and it’s all gets very similar philosophical with the universe. So we thought we’d create that in a bit of a bonus clip for everyone. And so here it comes. And please enjoy. I’m gonna leave it recording. I know we’re massively over time. So where I was going with the machine thing is, if you’ve got a few minutes now, if you’re an ant, if we’re an ant, yeah, if you’re an ant, are you aware of the wider colony that’s actually going on? And you could say, as a human, I could just think I’m an ant. And I just interact with other brands, very philosophical here. But you could say, Look, if I’m a person, am I aware of almost like this larger being that’s almost like going on, which is like human society? And are we and is technology just another aspect of that larger society that that’s going on? And so that’s what I mean by always, almost like getting embedded within it. Somebody voted for Philips? No, no, I don’t remember then. No, it’s, I’m super comfy. I just wonder,

1:18
okay, just to take the analogy. So there’s a book I started reading, and then it threw it against the wall. But it was a super intelligence was the name of the book. Yeah. And so the analogy was, okay, so that ant colony is a form of super intelligence. But an organization is also a form of super intelligence, because an organization is sensing and responding to the environment around us, and sending information and generating bodies of knowledge. And therefore it has a form of intelligence, which we don’t think about. So sometimes the company knows stuff that we don’t, and it tells us stuff, and then we act on it. So we are in like larger bodies. To go back to the ant example, there’s another assault, the ant example seems to be driven by pheromones. So the system that we’re not understanding is how a pheromone system works in the insect world. And what’s communicative and coordinate values are, and what patterns they seem to be repeating. So they, it’s an example of your, your kind of your example of we have 20 things that we can manage in our brain at the one time, and we can’t deal with ultraviolet light, infrared light, we don’t see it, we deal with this spectrum. And it’s the same with the perceptive world and the acoustic world and the sensory world. We don’t process all the information around us. So the ants are dealing with a whole system of communication, that we don’t even monitor, we don’t even perceive what that is. So we a very good and annoying thing to do is to always try and find out what the mechanism is. Because nothing happens by accident. There’s always a mechanism by which a thing can occur or something happens. So ants don’t coordinate with one another, through some form of superpower. There’s a system a process, a mechanism that allows them to communicate, we just haven’t really investigated it enough. And we didn’t have the tools to do it. And there wasn’t the economic reason to figure out why answer working like that. So the other one, that probably interest you is, I don’t know if you’ve ever listened to or read the book, I think it’s the secret life of trees.

3:40
I’ve heard about I haven’t read it, it’s a joy.

3:43
listen to the audiobook and kick back and take it in as a story. But the basic tenant of that is, trees communicate with one another. And they do so through the root system and through the fungal layer that sits over so there’s a mesh of funnels, and it actually transfer signals from one tree to the other through that fungal mesh. And a tree that’s not doing as well, because of its situation like it’s not in the light enough or it’s not creating enough sugar, the trees around it will send it sugar to give it a boost because they’re living and sharing those signals back and forth and going that tree short and if one of those trees gets attacked by an insect it releases a layer of resin to deal with the insect. But it also sends that signal to the other trees around to say you should start creating the resin too because these insects are in the area and they shouldn’t get landing on you. And final point is if you’re in a forest, the air and particularly deciduous old forest, the air around you stick with resonance from the leaves. So as your As they say, forest bathing, which is going for a walk in a good old forest, yeah, you’ve come out of it feeling relaxed. Yeah. Because you’ve just been soaked in this neuro, like it’s it’s an indoor toxin, but it’s a neurological effect on you, it actually is a mechanism that works on your brain and calms you down, and you feel less tense.

5:21
And you wonder if even with this, this, almost like super intelligence, or with some of these AI type models, being able to understand all these parameters, if like, we actually understand the how we’re actually part of that system to a certain extent. And even the human being, you could say, the human beings, like at a cellular level, a cell in my finger here is not aware of much more than the cells around it. A model for a human being could be almost like a coral reef. So I do diving as an example. Right? So are we like a coral reef where we actually think we’re one thing, but actually, we’re actually just a combination of lots of different cells. And it’s this magical? Yeah,

5:54
we don’t even count the viruses that live in us, the bacteria that we need to process everything, our biome in general. And I think the weight of viruses in our body is actually more than the weight of the cells, there’s more of foreign matter in us, then there is us. Yeah. And again, the concept of person as an ecosystem, is enormously complex. And we’re only scratching the surface, as to how it all works together, like we don’t really understand it, like we really don’t, is when people start talking about AI acting as a human. And with the psychology of a human, I go, Oh, I’m delighted that you’ve cracked the psychology of individuals problems, because cognition is a black box to a still, like little lights go on when we do the cognitive scanning. But cognition itself, very poorly understood.

6:55
I was on a Institute of Physics webinar yesterday, I think it was and they’re talking a bit about some of the stuff, you’re talking about how you can, you can clip all the different connections, and you can do something like 80% of the connections. And actually, the model still works, which comes back to the example you’re using, but then also then talking about brain and understanding around how the brain works. And so these ones like neural networks can help us understand the way the brain works with the relationships, but also can even be used to understand for example, disease pathways. Yes, we don’t understand much about the brain will allow us to do that then actually help solve problems like brain issues as an example. Because intellectually, we can’t understand it. But we need the computers to help us do that. Very interesting kind of linkage to what you were talking about.

7:34
Yeah. And the just the human conceit of the way. So the way we structure how we The goal here is, oh, if the machine could think like us, it will be smarter, right. And so we think that the way forward is symbolic reasoning. And there’s particular approaches to design. But what’s happened is that the large language models and the probability based just sheer it doesn’t matter if it was what’s the next word, it’s probability based, the ability to do that turned out to be so much more powerful for machines thinking. So we’re trying to model how a person thinks I take that and put it in the machine. What actually worked was let the machine do machine thinking, and it will get a different outcome are different way to that outcome. But it ends up in a lot of the area you would have ended up with, which are symbolic reasoning. Yeah. And it, I think it’ll raise some interesting issues. I was at a presentation in 2009. On iPhone data, so some phone data from a US carrier. And what it did is they tracked all the young people in New York on a typical day, and they then pattern it over New York. And they went, look, look, here’s where they are is an amazing go. Wow. Okay, what now? And this is something really interesting happened. This is we can tell up to like, I think it was 85%. At that time, we can tell you, we can tell where you will be at seven o’clock tomorrow. That’s amazing. How can you do that? This is because people do the same things over and over again. Yeah, I’m only going to go to two places on a Friday. It’s like that bar or my sister’s house. Yeah, that’s what I do in a Friday. I have all the options you could do globally, fly to Paris, dive in the ocean swimming. Yeah, you could have went to the pub, that’s what you actually did. And I have a feeling that when you see the data, you’ll go, oh, we all kind of behave in very similar patterns. And it’s not that complicated to predict what we’re likely to do, like, our brains and our emotions are structured to get us to act in certain ways. And maybe, sorry,

9:52
I was gonna I was gonna, sorry to interject, I was gonna say, but that’s one of the models around the brain isn’t that the brains is actually a prediction machine because we’re not physically computing those stuff. And we’re using right predictions to say, this is what I’m going to say next. And I wonder what it’s actually going to teach us around. Actually, by reflection on, we think we’re different than we actually are. And the way you actually work might be actually very similar to the way we’re actually building these models.

10:16
That’s Yeah, it’s interesting. The it goes, I think back to there’s this idea of authentic selves, and what is it person and an authentic individual? And personhood? Who are you? What is that little round thing in your brain, the homunculus of you, and we all have an idea that we’re in individual consciousness and that exists. And then you look at any neurologist or cognitive scientist, and they go, Oh, that’s not real. And you go, Oh, what you mean? No, that’s not real. That’s, that’s an effect. It’s not a thing. That’s its effect. So okay, let’s go from there.

10:57
I just try and resist the reality. While I believe it’s the case, at least anyway, because it helps me get through the day, to a certain extent,

11:03
there is a phrase and I’m gonna say it incorrectly, but the general tenor of it will be correct, let’s dig up the word. It’s some felt, doesn’t unfelt is that sensory environment and sensory net that you live in, as a beast, right, and the relationship you have creates the sense creative, the feeling of looking the sound of hearing, you’re creating this reality as you’re moving through it. And everyone’s is different. And every culture is slightly different. Because they have different contexts, they have different environments. There’s a way of being in the world, that if you think about it, as an unveiled, it allows you to go okay, it’s, it’s this co creation space between you and your environment, and cognition and sense, and experience, and that processing of all that together, ends up creating this entity that you think of as you are your consciousness. And it will be really interesting, eventually, like Copernicus, thinking that that the Earth or the sun did not revolve around the earth, the earth revolved around the sun. And guess what, there’s different planets out there. And if all that’s the case, then maybe there’s no God, maybe this whole garden of Eden story thing that we had was, and then God started, a Christian got started die off, and that kind of thing started to fall into the background, and you had the enlightenment. And I wonder, what would happen if we will, in 20 years time, we’re confronted with actually all our thing about the individual and human consciousness and the gift of consciousness. None of its real folks, here’s the data, here’s the, here’s how you can think

12:51
could be just like the earth becomes a system doesn’t. And we have to look after getting environmentalists here, but we have to look after the trees because we interact with them as much as they maybe the earth becomes maybe then the universe becomes a system. And it just, it’s all just one big system. And although we don’t have the arrogance to say that we’re just, we’re individuals as part of that, but actually, we’re just part of this bigger system. And almost like, if you can model it, we can assume the human race or earth could do some amazing things. That’s the

13:18
project of the university thing I’m involved in advising that that’s the kind of goal is to change the thinking structures, so that people can bring it into life and make change happen, because I think, like you only have to hear you never get changed in the NHS, which is just so nuts unchangeable. The only way to do it is to level it all you never get change in politics, because you’ll never get change. And it’s because change is so hard to make happen. And what you need is people then with tools and the resilience and the levers and the supports that helped make change stick, and then bring on the next change. And the only way to do that is to think systemically is to think systematically in systems wise. And to be systematic in your change. I think like the nature of the challenge that we’re faced with is I think some people get it, and the people who get it are really depressed. Because it’s yeah, we’ve just burned the planet, there is no coming back from 1.5. I don’t believe 1.5 I think it’s absolutely going to go to 2.5.

14:21
That’s accelerating as well. I think it’s really scary. What’s great.

14:25
It’s like if, if you’re looking at any of this and following any of this information, it’s hard to be worried about other stuff.

14:34
There’s nowhere else to go. Even I went on a on a long haul flight was the first long flight I’ve been on for a long time. But it’s so easy at different places. And you think I can be on the other side of the world in an afternoon you think like but there’s nowhere else to go. I can go anywhere in the world, right? But there is we got to be looking after the stuff that we’ve actually got. We’ve got to be looking after and burning all the stuff to get there. Right and it’s like we just we’ve got to be thinking about how do we fix the system to keep where we live habits But really, I think

15:02
we’re on the same hymn sheet here. It’s just how do we how do we make change happen? And that’s the hard bit.

15:08
But I do think your points around the number of parameters, and can you use these new analytical techniques to understand much more complex systems than we’ve ever done before? Yeah, I think it’s fascinating. And that even in terms of I mean, in terms of tech, things like coding, how do you code more complex systems? But also then, like brain pathways, environmental pathways? How do you do that? How do you change behaviors? How do you be persuasive for people, I think is fascinating. Chris,

15:33
let me just give you just two examples. This is from yesterday, like yesterday, some guy went to chat GPT. And he, within a couple of hours, created and with no coding just by instruction prompts, got it to make a virus computer virus that could enter your computer and exfiltrate your files. And he did that. Just creating it with Chad GPT, no code, no experience, created like a world class virus, and just went, Holy crap. Yes, those ways that the system tried to stop them do it. But he was able to break down the tasks and then recombine the tasks so that it got done. So certain amount of kind of smartness. And now a few are, and I’m sure the really smart people at Google and IBM and all the rest of them have been on this for years. They know that they have all the data from the medical experiments from the medical files in the hospitals, and they can look through your last 20 scans year long and maxed out which 20 scans are your blood and match that with some other Apple data from your wrist to go, actually, we should pull, we should call Paul in and just give him a blood test on virus B, because there’s something on his lung there that that was different than yesterday. And it was only able to do that, because AI was able to cross barriers and figure out that thing could be religion attributes, yes, so many things. And take that for all pictures of all things in all situations. And then go, what next? What could happen? And it? I know that sounds so much like everything, but how do I get another example of that? This is like not this is not bananas territory. But if you take actually this is this is another example. It’s real world example. They’re able to take pictures now off the Vinyard and go, the drone flies over the venue or back and forth. Okay, it’s arrow, it’s solar powered, has no exhausts and stuff like that. So it’s taking pictures all day long. And then you’re analyzing the picture. And it’s going, because it’s turned it on, I think it’s ultraviolet, or infrared. I can’t remember which one, and it goes, oh, there is a insect pest on that particular vine. And it comes down and just saps that and goes back up. And what happens is you have a pharmaceutical free Vinyard. It’s like there’s no pharmaceuticals used at all in the production. And you go, Oh, that’s interesting. So solar powered, it’s a drone. It’s used to monitor something that didn’t get monitored before. And it’s able to take an action. That means you don’t need pharmaceuticals, in your pesticide control, and you go, that’s not something that I would have said is going to happen. That’s not something I would have predicted. And you go, okay, maybe this stuff could go really interesting, really fast. If everything can be like a vision systems can look at a field and tell you what’s going on, and can alert you or give you actions or tell you things. Because one of the things that I was looking at yesterday, someone wants to talk about facial recognition systems. And so they were talking about people, and facial recognition and your right to privacy and stuff. And I said yeah, what about animals and said, What is this facial recognition systems work on animals? Yes. What do you mean it’s in Ireland as a company, and it knows that this cow because it recognizes its face and colors, ate this much and went over there and exercise that machine that yeah, so Okay. So what is it lots of interesting things happens. Once you can identify the thing. You go, that thing moved, stay there, that thing tends to like that corner. Don’t know why this thing tends to eat at this time. And not that time. I don’t know why. But now I’ve got some data and I could tie it to a an object approach. send everything and do the next iteration of why might that be happening? And so the idea of sensing monitoring, I think that’s all very much going to be visual systems predominantly. You can monitor like voice, noise, auditory, but it tends to, and it’s good. But I think that it’s like a quantum better when it’s with voice when it’s visual data, I think visual data is going to do a lot more interesting things than we understand right now. I think it’s going to be that’s going to be real, observable world, like what changes in the world from that?

20:43
I don’t think we’re that far away from presumably, you’ve got these augmented reality glasses as an example. But there’s an app as a Cornell University got an app or ambered ID, I think is a Merlin bird. Yeah,

20:53
got it. Yeah.

20:54
And it’s fantastic. I absolutely wonderful. Whenever we’re going out for a walk and it clicks, at least the bird you’re hearing, right? So I’m learning the bird. It’s not that far away from having that visually, were you wearing the glasses, and you’ll point out the information you’re interested in. So what plants there are others wild garlic over there, or this is the bird that’s happening in the background? They’ll give you a history around what’s happening in the back. All the information is available, I can go and look it up on the Internet. Today. It just takes a long time to do it. But the processing power to be able to do it instantly, almost like agile watching and who’s people’s what people’s names are and what their background is. It’s really not that far away at all. I don’t feel

21:30
so I I don’t think so either. I don’t know. And the way I conceptually think about it is it’s really mediating our relationship with the world. Yeah, because you’re now able to connect to it and understand it in ways that deepen your relationship with it. But it’s got it, that’s how it has to be framed. Like, like I was out for I live by a river. And I’m blessed with the fact that I can just walk in here, the birdsong and my lunch break and stuff like that. And I really make a point to doing it. And I’d be there now and that my sometimes my wife is with me, and oh, that’s the that black, that’s the culprit. Because McKnight squeaky noise, says he, that’s the cool thing. Because that one there, that’s a thrush. And she goes, I’d love to be able to tell all the birds apart. And I said, it’s practice. Since you’re just you’re listening, you go and look it up, you go, that was the one you recorded, maybe use the app. But what happens is generalized Birdsong, now becomes individuated. And you can name them and you can position them and you can relate to what’s going on, and you understand it a little bit better. So your relationship to the world is deepened by understanding it a bit more. And I think that that hasn’t acted on the world in some way. Like you haven’t changed things. But it’s that thing about facilitating the relationship between the person and the world? How does technology do that better? So that it, I think, was for grounding the individual? It’s like, how can we make it so that you better meet people, if you’re able to better listen to them, you’re able to better be better with people. One of one of the guys I live around here, he works in a deprived area, trying to get people involved and engaged and doing things. And he said, the biggest problem they have with violence is impulse control. It’s just no one’s got the ability to stop and think about consequences about what’s happening and control them. Because they’re all dealing with some level of trauma. Whatever it is, anger is some form of trauma, it’s some form of thing that’s happened to you that you are no longer able to control your own temper, you just haven’t developed the skills to do it. What if something was able to help a person understand that they were in a pre anger situation? A poll, this kind of conversation tends to send you into a spiral of anger. Now, all that is right. I’ll be back and I’ll talk to you in 10 minutes, I need to put out the bins and you leave the conversation. And then you come back. And you go, cuz most behavior change is actually the setup. It’s not the behavior. It’s the situation before the behavior happens. So if you’re able to sense that, then the behavior doesn’t happen, then you can start doing the cognitive kind of work to say, Did you recognize the situation, did you? So it’s, I think those soft areas of sensing and then helping people understand the world that they live in, will be very non disk RIP won’t be like followed around by big robots. It’ll be like a watch or something telling us something that just

25:06
reflect on like, how much would how much I knew 30 years ago, let’s say, like pre internet versus how much you know now. And the ability to watch stuff on YouTube and stuff I know about me getting into all sorts of things. I’m not an expert and all this stuff, but just like so much more rounded and better educated. Maybe you get take that further ahead, and you have this extra insight that’s actually happening, how much more do you need the cognitive load to be taken off you to be able to do things better, you just suddenly you start to know about birds, and how the systems interacting in a way that we didn’t before. Fascinating 100%.

25:37
And if you take I took a poem from I think it was Seamus Heaney, and I put it into the web chat GPT. And I said, analyze this poem, using the ideas of syntax and vocabulary and something else in related to history and race. And give me an analysis of how that happened. And it was really good job, they did a really good job. I said, Wow, it’s really quite good. And then I put something in, I made up. And I said, this is a poet named whatever. He’s done this in relation to the Seamus Heaney poem, do an analysis of the Compare and contrast kind of thing. And they came back and said, Oh, this author took that view. And this author took that view. Well, that’s really good. Now imagine if you took to go back to our business and say, hey, I want to take two loan agreements, and put them into the chat. And let’s say, compare these loan agreements and tell me which one is better and why. And you go, this one is better for this. And that one’s better for that. And so okay, if you were dotted out of that, and you wanted to do XY and Z, how would you have written that agreement? And this is the agreement, I would have gone with you go looks like a better agreement. I’m gonna send this agreement back to the company and go, actually, this is the agreement I want to sign not the other one that you have. And you go, Okay, we don’t operate like that. This is our agreements, I think, yeah. But the minute people can do TripAdvisor for whatever it is, you do, it’s I don’t like the policy changes. I don’t like the things that you’ve done here. For whatever service it is it all that’s going to change, like, How many times have you clicked except on 230 pages of an apple change of services agreement for how we’re going to deal with your data on your iPhone or your laptop, if you have the ability to just get that and send it to chat GPT and go, tell me what I should be worried about in this document. And it goes, the major issue in this document is they’re no longer storing your data in the EU. That’s not acceptable. What should I do to change that with this particular agreement? Oh, you have to go to here in here and change the settings. Okay, that’s do that then. It’s going to be interesting.

27:58
I like the idea of so was talking about debt advice earlier in the week, and almost like, how do you present different personas to be persuasive for people to not get into debt, this is pre collection to not get into debt. And you might be chatting to someone down the pub might be different than someone who might be going to a sports field, or they’re with their kids, or whatever it is. And we all have different personas, what might be persuasive for me, is not persuasive for someone else. And how do you have an almost like an agent that can change the way they interact with you to get the outcome that actually better for that person? I think that’s what that kind of stuff is fascinating.

28:30
It is fascinating. So I gave you seven words and said these seven words, describe humanity. What would you think? Do you think this is rock solid? Or would you go? I think we should investigate this a little bit more about how we think about personalities, and how that’s framed and how scientific or not that is, all you can do is say I get better outcomes. When I do this, and this,

28:58
this welcome to that you can use the data by test these seven different scenarios, I get better outcomes based on these attributes here, and it gives me a better outcome. So that’s it. Yeah, it’s it’s based off the data why

29:11
I get hung up on this is I used to be an academic, and I never truly lost the habit of thinking. So that what happens is when you get down into the into definitions, and say, Why is this definition here? And who defines it? And how strong is that? It usually comes down to a thing called What was it called? The internal consistency of the factor. So if you do factor analysis of something, the factor itself will have various measures of how specific it is in measuring that thing, like its internal consistency. And it has to sit if you’re using that factor has to be tied to the exact questions that was used to use that because that’s how the things are built. So, when someone comes along and says, Okay, company size is related to their ability to do x, which would be a typical academic framework, you go define size, how did you define it? Why did you define it by number of people and not number of outlets, or turnover or revenue, or like you’re even definition of size wasn’t nailed down to anything in particular, and no basis, in fact, under it. So, with regards, I will come back, there’s one other point I’m making, they’ll come back. If you look at the repeatability of all the experiments in cognitive psychology, there’s some bananas number, like 90% of all the experiments conducted in cognitive psychology, were not repeated. So they have no basis in science, none of them. And the whole area got back close to being abandoned around five or six years ago, because the theory base was so weak in the in the in the science of what they’re talking about. So when I hear about ours is based on psychology, and it’s based on echo Oh, that’s really interesting. Can we get into what theory of the mind you’re using? What theory of the psychology you’re using? What personality types? Where are you getting your research base from compared to what was that decision, and people will go, actually, I don’t want to be in this conversation anymore. Because this is spiral, I thought it was a good idea to do the quick, it’s friendly people versus formal people, emotional people versus rational people, Irish people versus Italian people make categories up as you go along. And so I’m very skeptical of how scientific these things are, I think the most solid base of what’s going on with you in terms of your cognitions, etc, is the actual biomarkers. biomarkers are like observable, you can say, that happened that happened, your voice, your voice, your eyes rise, very hard to I suppose

32:11
human beings, we tried to simplify things I was when you explaining our thinking is very Victorian, actually, it was it’s worked. It’s worked very well in science, hasn’t it? We try and simplify things to understand them. My background originally was physics, right? And like, the deeper you drill down into something, you realize that nothing really exists. Because it falls apart, the probability, the end, right, and it gets, at what point do you want to have to group things? And is it helpful or not to get there, but but maybe maybe these these new mathematical models will actually help us really understand the next level of complexity, now, you’ve gone

32:44
straight to where I was going, which is, it’s, if you just threw away those concepts of personality all together, and just say, this is all the data I have about this person. And then let the model figure out, okay, these are the parts, this is what happened. And then you can let it go completely free, because there’s things that it could do that you don’t want it to do. What I think what you might, I liked it. So they found that I think it was a certain type of man, I think it was like 55 to 65, or something like that was consistently late for NHS appointments. And at a certain type of person, or they’re not reliable today forget. So they whatever. And they went in, they looked and then some Brightspark went, how did they get to the hospital, public transport? Why were they late? Public transport, let them down. They couldn’t get there. If you had two bosses, one needed to join to the other to get you there on time, you were screwed, there was a percentage of times that the transport was going to let you down. And therefore you were going to make it It had nothing to do with your personality or your whatever, whatever, whatever. But if if models end up having access to all the data around more and more data, it will make the inferences I don’t think it’s ever gonna have access to all the data, like across all systems and all life because that’s just not the way that’s going to go. I think that we’re going to see if you ever hear this. There’s another concept called context collapse. context collapse is was observed in social networks where I’m on Facebook and I’m friendly, happy polar lives here in a living in in my area. Twitter is interested in these kinds of things. And LinkedIn is this is my professional work life. And this is going to talk about all three are different parts of who I am. And there’s separate contexts in which I want to manage those relationships. But what happens over time, is they start bleeding into one another. Like I start posting on Facebook about what’s happened my company You

35:00
go on LinkedIn more getting more social as an example, right.

35:03
And so what’s happening is the context of each of those things starts to collapse. And the attenuation data actually drifts. And what happens is what I’m seeing a LinkedIn loose with predictability as to whether I’m looking for another job, or whether I’m looking to hire or whatever, because I’ve polluted the clarity of the context with all this other data. And so that, that context collapse, I think, in all sorts of data happens, and what will happen, I think, next is the MLMs was just find a new path, through the data set are certain things that just tend to work, these are the things that you tend to offer. And I don’t know, if I think that they I think it will give hyper specific paths for people to get things done. The question is, then, how do you govern that? Like, how do you make sure that you’ve got control over those processes, and that you don’t, you don’t get people to do one bad thing so that they end up with a good thing. Like I used to have this example where when Amazon came out, and it had the Amazon screen, and it could see you, and actually what it could see is you and your room and your posture, and what it was judging was your mood. And it was great. lexipol happy, Paul sad, Paul, very sad. Posture can tell you an awful lot and stand up in the morning and ask Amazon does this suit me? And it goes, yes, that color that tone that cut does flatter your whatever and whatever. Okay? Because, hey, Amazon has a special offer and chocolate cake today. Would you like chocolate cake? Yeah, I love chocolate cake. Bindung are later chocolate cake arrives my door. I eat chocolate cake. I’m happy 10 minutes later, go shouldn’t have really eaten chocolate cake in enough ways as it is. I should be getting whatever. Hey, Paul, would you like to look at our gym equipment? Yeah, I suppose a better look at gym equipment. Right? Yeah, I’ll order some gym equipment be good for me. Hey, as an AI, the path of offering a depressed person chocolate cake and selling them gym equipment works? Yeah, why don’t we just do that? So okay, that’s probably not a good path to, to have out there. So I think that the idea of like the categories that we’ve lived with and how we think about, like, I think the opportunity is in breaking those and looking deeper into what’s happening, and figuring out just much more nuanced way. What’s going on. And we are examining messages from people who are, like in difficulty, as I’m sure you’ve seen in many different instances. And people’s lives are complicated. And there’s a lot going on. And if you’re offering someone yes, no, maybe, then that’s not really helping. That’s not really getting into something with someone and figuring out how much they need to or want to engage with you. And maybe not everyone needs to engage with you. Maybe people just need to have a firm line and understand that they need to keep it. But the opportunity seems to me to be to not keep the categories that we always have to not keep the culturally defined things that we inevitably have that we’re blind to most of the time anyway. And just use the opportunity to be much more attenuated, much more considered about what might happen.

38:45
So losing that reductive analysis that we use, we’re reducing it down to the factors and maybe just taking it up a step. I really appreciate you taking the time. It’s fascinating. It’s good to chat to you as well as Paul, thanks very much. I really appreciate it.