AI Ethics and the Future of Digital Work

Mark Smith: 0:14
Welcome to the Power Platform Show. Thanks for joining me today. I hope today's guest inspires and educates you on the possibilities of the Microsoft Power Platform. Now let's get on with the show. In this episode we'll be focusing on evolving SharePoint teams and AI. Ai is affecting everything in the world we live in these days. Today's guest is from North Furby in England. He's the founder of NovaWorks. He is a lead author and facilitator on the maturity model for Microsoft 365. He's the author of two patents relating to medical devices, which is a little interesting fun fact there. You can find links to his bio and socials in the show notes for this episode. Welcome to the show, Simon.

Simon Hudson: 1:07
Yeah, hey, mark, very good to be here. I wish I was there, though. New Zealand I've not been to New Zealand for just over a year now and I'm missing it badly.

Mark Smith: 1:18
I feel that well, okay, that's not too long ago though that you know, as I have quite a few guests on from the United Kingdom and I know you're in the middle of winter and here I am, like you know, t-shirts and shorts outside at the beach every day. It's a different lifestyle from when I was living in London, and I was fortunate enough while I was there in winter well, in Christmas time to get a snow, which was amazing.

Simon Hudson: 1:43
Yeah, we've had snow in the UK this week, which is pretty rare these days.

Mark Smith: 1:47
But January man, it can be bleak in London eh.

Simon Hudson: 1:50
It can be bleak everywhere in the UK. Yeah, winter I mean it's been two months of just cloud, you know, yeah, you know the light's out winter it's been two months of just cloud.

Mark Smith: 1:57
Yeah, the lights out at 3 pm in the afternoon doesn't come on till nine often and as I'm going to say lights out, the sun's gone, Yep.

Simon Hudson: 2:04
You're absolutely right. On the other hand, we're getting sunnier weather starting now and you guys are going into winter. That's it. That's it, the solstice has passed right.

Mark Smith: 2:16
Yeah, crazy times Food family and fun.

Simon Hudson: 2:17
What do they mean for you? Oh gosh. Well, let's start with my family, because that's kind of important. Um, I've been married to christine, my beautiful wife, for 38 years now 38 years. God, she has a medal, um, so she's sitting downstairs trying to get over the cold that you can also hear in my voice today, but, uh, we've been together a long time and it's still great, so that's rather nice. And then, um, we have a daughter together well, obviously together, I guess. Um, she's uh, 28. She's a beautiful young lady called ella hudson, and she's soon to be dr ella hudson. She's, uh, just submitted her phd thesis in biotechnology and just landed the dream job that she's been after for the last seven years. So, yeah, so my family are kind of important. They're all quite close to me, so that's rather nice. I love it. Food, food. Oh well, I mean, I've got no restrictions on food. I love all foods from all different things, apart from cheese.

Simon Hudson: 3:09
Cheese makes me unbelievably ill, um, but apart from yeah, yeah, no, some people say it's a bad thing. I mean, how would I know? Um, so, yeah, cheese is good, but but foods that I love and I cook a lot. In fact, I've been a lot for a long time. I've been the main cook in the house. We redid the kitchen a couple of years ago so I had an even bigger, more fun space to cook in. So, yeah, and we're just literally this. Last couple of days, we of days, we've started on a thing where we're going to try and be vegetarian during weekdays. Certainly for the next month we'll see how that goes.

Mark Smith: 3:39
We're not vegetarians, but but we want to try and do our bit, so we're going to try to be vegetarian five days a week well, in london, my wife and I did it for a month, you know, and and here's the thing I I was brought up very much a carnivore and on a dairy farm and things like that, and a meal wasn't a meal if it didn't have meat in it, type thing. And so I, you know, for me it was around changing a paradigm that was stuck in my mind that you can't have tasty food if it didn't include, you know, a meat based protein. And so I think, yeah, I'm big on changing paradigms that you inherited through your upbringing that might not work for you now, or that society has fed us that might not work for you Precisely.

Simon Hudson: 4:25
We didn't cover fun, fun, oh well, there's just a list that goes on and on and on. I'm sorry, I'll give you the summary Music, love music. I've seen I think I'm on my 129th band now because I keep track of them all. I've seen I think I'm on my 129th band now because I keep track of them all. I've got a massive music collection of CDs and loads of vinyl and a very good hi-fi and I play guitar and mandolin. I'm trying to learn gypsy jazz guitar, which is really hard. I love my cars. I've got a couple of classic cars, a couple of Beats, actually, yeah. But I also have an EV and we're looking to buy our second ev this year. Games we played dungeons and dragons a lot. We do all sorts of board games. I taught my daughter to be incredibly competitive at games, um, so she beats me. Now we ski, uh, and then I've both spent the rest of the time boring people to death, either about sustainability issues or technology.

Mark Smith: 5:15
I love it. So on my property, my wife and I between us is we've created a brand called High Tech Hippies, which she came up with, and it's because we are trying to narrow one of the things that we've got an acre and a half and I've got a big garden and I'm starting to develop a food forest because I want to eliminate the transportation between the plate and my food source and, um, and we don't have any animals, which was very like. The first thing I did when I bought the property was put fencing in for animals and we've decided, no, we're not going to have animals per se. I mean, we have cats and um, and I'll get chickens at some point for the eggs, but I'm not getting typically I would have got sheep or a cow or something like that on, you know, to fatten up type thing.

Mark Smith: 6:06
And yeah, so sustainability in the last, particularly probably since moving back to New Zealand, has become a massive focus for us from a practical perspective, not so much in our business lives. But how are we? What can we do to change things? You know, I have a worm farm to recycle my vegetables. I collect all my gray water in the property and I recycle it multiple times across trees and plants on the property. You know, if I collect a drop of water I to you know benefit, get benefit from it three or four times before it is um, you know, goes back into the earth type thing. So I'm trying to maximize that yep, and it's really interesting.

Simon Hudson: 6:47
I mean, I got into the whole energy transition part because of my background, um, which we'll talk about in a sec, um, but, um, the energy was where I started and then I realized it wasn't just about energy, it's also about all the other parts of sustainability and diet and the way we eat is a big part of that. But it's just interesting to do Learning to cook just with really interesting, exciting, nutritious meals with no meat. That's a fun thing to do.

Mark Smith: 7:12
So another challenge is the craft to pick up and you just need to look at often the Indian cooking, the amazing flavors that profiles that they can bring out in food and you know, and they're totally vegetarian-based dishes.

Simon Hudson: 7:26
Yeah, about all sorts of Asian food. I mean, there's such a range of different stuff you can do and I'm busy trying to, you know, pick up as many techniques as I can from wherever they are.

Mark Smith: 7:35
What's your favorite feature in your kitchen that you remodeled?

Simon Hudson: 7:40
I mean, I guess the main thing is I've just got an awful lot of workspace now, you know, we've got rid of all the clutter, we just got I don't know eight or ten meters of usable worktop, and so that means I can spread out and get all my stuff prepared and you know, I can have, you know, the dehydrator going in one corner and my bread machine in another and still have plenty of workspace. But I guess the big thing, actually it took some doing, it took a couple of chefs, I know, to convince me to switching from a gas hob to an induction hob.

Mark Smith: 8:10
Wow, interesting, interesting.

Simon Hudson: 8:13
So we're a fully electric home, now without gas Wow.

Mark Smith: 8:17
I've just been approved to build my home our home on the property here, which has been laborious. I left London in 2019 and the architecture for it had been started in 2018 and we've only just got to approval. It's a long process and one of the things in the kitchen that we're doing quite different than most is putting a commercial underbench dishwasher in. Oh yeah, okay, because they're much more efficient than your normal home dishwashers.

Mark Smith: 8:47
I did not know that and they turn around like typically I've got where I am now a dish drawer and it takes an hour 15 minutes to do a standard cycle. A dish drawer and it takes an hour 15 minutes to do a standard cycle. A commercial dishwasher takes five minutes to do the same. High-level sterility, cleanliness, that type of thing, yep, without that running for an hour 15 minutes. And so it's just like and it was from an architect they said listen, this is probably the game-changing feature we recommend in new builds is that you pay a little bit more for a commercial dishwasher. It's still small, not a big industrial machine, but they run quite differently than your home user unit.

Simon Hudson: 9:28
I've never heard that before. I'm going to have to go and investigate that now because it's interesting. I want to see what the energy profile looks like, but also what's the water use look like.

Mark Smith: 9:36
Yeah, let's move on and talk about your career. How did it get to where it is? What are you focused on now? What's the big bets that you're making from your career at this time?

Simon Hudson: 9:48
Well, I'll give you the kind of, you know, the bullet point version, then we'll expand on that. I mean, you know, the really short version looks like physics, medical device, corporate entrepreneurship, retirement. That's my career in a nutshell. Not exactly an obvious career path, however, yeah, I studied physics, did undergraduate and postgrad in physics and bit of chemistry Rubbish chemistry, don't ask me chemistry questions and then I spent 15 years working in the medical device industry globally, which was really cool. I mean, I learned a lot, got to do some fantastic things, you know, traveled the world. That's where my patents come from.

Simon Hudson: 10:21
I designed two new. I designed a lot of new products, but two of them were patentable. I made a difference. That was a big thing. I designed products that helped people, and so mostly it was chronic wound dressings. So these are the horrible wounds that don't tend to heal, usually in compromised, uh, older people, and I designed things and sold things to try and make that better to manage, better for the patients, better for the clinicians. So, and make the wounds heal. How long ago? Oh gosh, um, that's probably 18 years since I left that industry, something like that, maybe 20.

Mark Smith: 11:00
And you know, patents interest me and it's something that I feel that I need to do at some point. But I just feel like, is it a long, drawn out process to get a patent, or is it a relatively straightforward? And from the UK perspective, if that's where you did it, I mean, you know.

Simon Hudson: 11:43
New Zealand patent law and the UK patent law are very similar because my daughter's becoming a patent lawyer, so I know these things now. Oh, no way. Yeah, yeah, the key thing is, you know, to invent something which is genuinely patentable. That's the first step. And there's a whole bunch of things, but it needs to be an inventive step that an ordinary person knowledgeable in the field wouldn't automatically come up with and which hasn't been done before or talked about in public. So that's your short version. So, um, you know, if you can come up with a great idea that is patentable, then that's your first step. Getting it patented usually can take six months to a year if it's a good one. You know there's a whole bunch of searches to make sure that no one's going to challenge it. You know you do a bunch of stuff, or your patent attorney does a bunch of stuff, and then the patent office themselves does a bunch of stuff, which is mostly publishing it and waiting to see if anybody comes back and says this is infringement or we've already got that idea or no. It's just being obvious. But it can be fairly quick.

Simon Hudson: 12:41
The the hard thing is then protecting the patent. Um, so anybody can come along and challenge it and go. Actually, you know we've got something. We either want to work around patent, in which case you challenge them, or they've got prior art that they come back and go. Actually, we need to constrain the patent or have it struck off. So there's more money. It's more costly to protect a patent than it is to create one. It's a fascinating area, though. In the UK or maybe there's more patents that come out of Great Britain and I guess what we used to call the Commonwealth heaven knows what we call it these days without treading on toes than everywhere else in the world put together, which is kind of cool.

Mark Smith: 13:22
Yeah, yeah, I worked at IBM for three years and they used to be the top patent producers and just to recent times that they have now dropped out of that. And I've spoken with people inside Microsoft and they get patent awards every time they get something patented for Microsoft and often there was with the MVP program and stuff. Why is there so much secrecy around everything? And it's because of what you said before. If too many, they can't patent something. If too many people, if it's leaked down the public domain or anything like that, yeah, it's not too many people.

Simon Hudson: 13:53
If anybody says oh, I've already heard about that, that's it, Patent's dead.

Mark Smith: 13:57
Yeah, it's sub 50, isn't it? It's like a very low number, it's none.

Simon Hudson: 14:03
It's none.

Simon Hudson: 14:03
It's none If you go and talk to your mate down the pub and go, I've got this great idea, and then he gets challenged by the patent lawyers and goes you heard of this, it's in the public domain, it's none, which is kind of scary. So you're absolutely right. I sit on the Microsoft for Startups Advisors Board so I talk to quite a lot of startups through that and through the university stuff that I do, and one of the top things I tell them is you know, don't tell your mates about it. If you think you've got a great idea, you can talk to me because we're under NDA, but don't tell anybody about it. Maybe your partner maybe, but make sure he or she knows and not to talk about it.

Simon Hudson: 14:44
Yeah, great stuff. That's so interesting. And of course, the purpose of patents is not to hide your invention, it's to put the stuff out there so everyone knows how to do it. But you get the benefits of it for a little while. So basically, you're giving mankind your brilliant idea in return for about a quarter of a lifetime's benefit from that.

Mark Smith: 15:04
Yeah, I've gone through a trademarking process, so I know one. I had to go through multiple iterations because it was not only just a word trademark, it was also a design trademark. And you know, the first iterations were considered to be too infringing on others because it involved the silver fern, which in New Zealand is, as you can imagine, a big deal. And so I finally got it through. And then, of course, where I really spent the money and the 10 plus years of having it is the infringement. Yep, Somebody comes along, sets up a new company and decided to use the whole phrase in it. They've not done any research or anything, and of course it flagged to my attention and then I had to lawyer up because it's only as good as you can defend it right.

Simon Hudson: 15:47
And so, yeah, yeah, it's a really interesting space. The whole intellectual property field is fascinating and probably deserve a bit more attention from the tech people, not just in terms of protecting what they do, but actually building some better tools for doing intellectual property management, and AI is both a problem and a solution in that space.

Mark Smith: 16:07
I like it. Let's drill into that. Tell us about in the discussion we had before. You've been involved in SharePoint, teams and AI. Just let's take those three broad headings. What are your thoughts on them, your experience with them and your thoughts going forward with them?

Simon Hudson: 16:25
Well, I guess you know SharePoint's where I'm going to start, because it's literally where I started in my move into tech, because I haven't got a tech background. You know, I think that shocks people sometimes. Here I am as an MVP and they go so which technology? No, I've never been in IT, actually I just started IT companies. But SharePoint for me was really interesting. I guess I've had my hands on SharePoint since Team Server, so pre-WSS, pre-sharepoint 2001. Because of its ability to do information management, I mean the way.

Simon Hudson: 16:58
I think my science background in particular lends me to think about things in terms of information architecture and structures. Sharepoint and also, slightly before that, probably Lotus Notes, lotus Domino, were the first tools I had hands-on that let me start to structure the information stores I wanted to do whatever the heck I was doing in my career at that time, and so then, when I left the medical device industry, I ended up working for an IT company running their medical division over in beautiful York, and they were doing nothing with SharePoint. They were doing loads of stuff with content management tools, big, expensive tools like Vignette I mean proper expensive, yes, yes, I remember Vignette and I said well, this SharePoint thing looks like it can do a lot of what you're doing, but for like 10% of the cost, which didn't actually please them because they figured they could make more money out of Vignette. I don't know if anyone buys Vignette anymore. I I don't know if anyone buys vignettes anymore. I think they may have backed the wrong camel there.

Simon Hudson: 17:52
But I got more and more into looking at SharePoint as a tool and then realized and this was really the core of me becoming an entrepreneur realized that all the dozens and dozens of companies and organizations I was talking to when you drilled right down into it, what they needed was basically always the same. There was a core sort of knowledge and information structure that everybody has. It's about sort of, um, you know, centralized content management, you know document center, about departmental activity. But there's two halves to that the internal department and the outward facing part of the department. There's all these structures I saw appearing time and time again, and every time I spoke to the tech guys in the company I was working for and others, their solution was throw a bunch of developers at it, do some complicated pointers. Time again, and every time I spoke to the tech guys in the company I was working for and others, their solution was throw a bunch of developers at it, do some complicated pointers, waterfall processes and build a brand new one from the ground up. Now this is crazy. They all have the same needs. They may not know it, but 90 of what they do is the same. Why don't we just build a core that has that 90 and spend the rest of the money on actually doing the bit that's unique to them? So they didn't listen.

Simon Hudson: 18:54
So I left in interesting circumstances and started my own company doing exactly what I'd been saying. I set up Cloud2. We built a product called Hadron and Hadron embodied all of that, the information architecture that I'd been working over for five years or so, and we think that was the very first SharePoint-based what they now call an internet in the box. We call it a solution accelerator, but we think it was that. There's just one other company that maybe has equal claim to that, and we did that. We mostly focused on building that for the NHS because, man, they needed stuff At the time.

Simon Hudson: 19:28
It was the average time, because we looked at the research to build an intranet was two years upwards of quarter million pounds. Wow, wow, two years, quarter million pounds, you know, half a million new zealand dollars um, we were doing it in 14 weeks and about a hundred thousand dollars amazing. So. So that was that was my thing and there's sharepoint developed. We kept reworking the model and doing more things, but for us it was always really about the information management and architecture and the collaboration and not really about the traditional comms wanting internet where they can put some things that no one's ever going to read. So that was the heart of my story. And then, towards the end of my time at Cloud2, because I guess I ran that with my business partner for 10 years or thereabouts and I had an exit plan and then, about 18 months before I exited, teams came along. Yes, and at first I went. I don't really know what this is Teams it looks like a not very good version of that other product that I've completely forgotten the name of now.

Mark Smith: 20:32
That was like a stream of consciousness sharing Link originally, and then Skype.

Simon Hudson: 20:36
No, a competitor product. I can't even remember what it's called, it doesn't matter. Ah, yes, yes, Slack. So that's the one which I hated. I hated and I thought the Teams was just Microsoft's bad attempt to build Slack. And then I had this road to Damascus moment and I go oh, but you can store content in it and you can expose that content through tabs and channels, and you can structure channels within team areas and you can actually build an information architecture around collaboration. And that lit me up enormously. So I can build information architecture for collaboration, not just for content.

Simon Hudson: 21:08
And at that point I suddenly became one of the biggest voices of saying you know this Teams thing, we really ought to be doing it. And I built out a couple of models for Teams solutions in the same way that I had for SharePoint, and started migrating the whole concept of an intranet into being embedded inside Teams, so you wouldn't have to go to SharePoint for one thing and something else, you just go to Teams. All your stuff would be there, including your email and your calendars. And that model still continues. I think Microsoft are catching up on the idea now. But the idea was your entire workspace is inside Teams, yeah, Except when you have to spawn out to go and do something else.

Simon Hudson: 21:45
So we did a lot with Teams and I've continued doing that a little bit since I retired. So I left Cloud2. I shut down a second company, a medical company that I started, so I now just have a small consulting company. We still do a bit of work helping people get on board with teams, and all the people that I'd helped that took my advice. When that thing that happened a few years ago, that pandemic-y thing happened and we weren't allowed to walk outside the house, all the companies that had taken my advice just went and switched to remote working without missing a beat and all the others ran around with their heads on fire. So I was really quite proud of that piece of envisioning that we had Now introduce AI. Yeah, so I mean I guess we're coming up to two years into AI now, aren't we?

Mark Smith: 22:33
November 22,. Right, no November 23.

Simon Hudson: 22:36
Yeah, coming up to two years, you know, since you know, sort of human-readable large language models came along and changed the profile and that's been really interesting. Yeah, I played around with AI a lot to begin with. I spent a fair bit of time talking about the benefits, allaying people's fears about the fact that you know they probably won't ruin society overnight. It almost certainly won't take people's jobs, but it will take the jobs away from people that don't learn to use AI as every technology in the history of mankind has done. You know, if you don't adapt and move on, you will be affected, but there won't be a loss of jobs overall. And I did quite a bit of stuff looking at the ethics and morality of AI, and still do so. I have an ongoing I'd like to call it a project. That's far too grand a title, but I have an activity to try and find out how we can build ethics into AI If we don't have it at the moment.

Simon Hudson: 23:28
The moment we build ethics around AI. We put these guardrails up that qualify the stuff we ask it, so you can't ask it to do nasty things and qualifies what it puts out so it can't tell you nasty things. I want to build it into the heart of it. And the reason is at some point very soon we've seen it in the last few months agents are coming. We saw agents running inside self-driving cars already.

Simon Hudson: 23:52
These agents are going everywhere and we need to understand what their moral framework is. And those moral frameworks need to be definable in a way that we can say you know, you guys in New Zealand you've got a slightly different view on things than us laggards here in the UK have, and we're certainly a good deal different to the ones over in the US. I don't think I want to trust my car to use an AI built with US morality. It's going to run some people over that I don't want to run over. So I'm really interested in that. Ethics of AI, not in terms of how do we use it, but how do we make them ethical, because if we can get that right, maybe our AIs can be the better version of us.

Mark Smith: 24:30
Well, yeah, my specific thoughts are very, very utopian in the future of where AI will go and even on difficult things like organizational bias that has been built up over potentially decades. For then why couldn't we have an agent, a bias agent, that explicitly goes out and quantifies that and cleans it up and, you know, helps enact new policy? So we don't just employ 99% males that are white, that are you know, because that's what we've always done and therefore we're modeling off what we've always done and that's how the you know that bias becomes institutionalized. So I do agree with you One thing that struck me recently with you know, copilot and the Microsoft Graph and people you know using it, and they've been collecting data For the last 10 years we've talked about one of the mega trends that are going to affect the world is the proliferation of data and it's funny now because it's almost like we don't have enough data and the need for training in the future, and I think models will evolve with this. And then there's synthetic data, of course, coming in now with AI generated.

Mark Smith: 25:48
But one of the challenges is that humans make mistakes and no matter how hard we try to stop humans making mistakes and it was funny, I was on a call yesterday and someone said what do I need to change in what I'm doing? And I'm like, I'm not good at spelling. It's part of my dyslexia and no matter how hard I try, I can't see that I've spelt something wrong. I can't see it, and so if I'm going to go, I need to correct that mistake. I've got to understand that it's a mistake in the first place before I can correct it. The problem is in institutions and organizations that have SharePoint teams, et cetera, micro mistakes have been made all the time. You didn't put the correct T's and C's in a document, you didn't adjust the numbers in a spreadsheet, you dropped something that's changed the calculations, all these little incremental and then you copy that document and now there's 10 different proliferations of that mistake and you amplify that across thousands of employees, potentially an organization.

Mark Smith: 26:53
And then we go hey, ai, learn our organizational data and let's use that as a grounding point. And then we go oh my gosh, ai is hallucinating. Is it hallucinating? It actually might be just presenting back to us massive organizational mistakes that is human to make. And so, therefore, I think that you know the ai story inside any organization is a data story, and although we've been collecting data for years. There needs to be an amount of triage on that data to and then also narrowing down when we have an ai use case, to make sure that we're not giving it the ocean to operate on, but really we're giving it just the data set that is pertinent to the AI task, rather than we're going, hey, let's just give it access to everything and let it learn and find the patterns, because the problem is it's finding the patterns of error and thinking that it's truth.

Simon Hudson: 27:51
Absolutely, and people like you and I and those within the community have been banging on about the importance of data governance since AI first became a thing a year and a half ago. Because if you've got bad data in your organization and, let's face it, almost every organization's data is rubbish you know they've got some great stuff in there, but they've kept it mixed in with all the rubbish Then you're going to get bad results. It's a garbage in, garbage out model on steroids. So we've definitely got that problem and I think maybe one of the greatest initial benefits of AI is going to be getting people to sort their data out. Yeah, so I'm a silver lining kind of a guy. Maybe people actually go well, maybe we do need to do that. That's something they've been telling us to do for a decade.

Mark Smith: 28:35
Have you seen any tools that orgs can use that can potentially help them with this?

Simon Hudson: 28:41
I mean, there's loads of tools. At the end of the day, it's still a human judgment thing and it's a process thing. You know, obviously Microsoft have got tools like Purview. I think that that isn't really going to help with the legacy data issues or governance issues. I think it's pretty good at doing some things about looking at the quality of the data, putting in effective lifecycle management stuff all that content management where.

Simon Hudson: 29:04
I started 30 years ago Doing all that stuff. Right, there's no shortage of tools, but someone's going to have to, in each organization, bite the bullet and go. You know we're going to implement policies now, and I'm not talking technology policies, I'm talking organizational policies. You will not keep your rubbish for 30 years. You know you're going to get rid of it after six months unless it's subject to a retention policy. You are not going to create duplicates and if you do, it's a disciplinary offense. You're not going to do all these things that we've got away with because we had too much story space and people couldn't find the time or wouldn't make the time to clear up. So I think there's some stuff there.

Simon Hudson: 29:43
However, there's a big caveat to this, which is there is an argument to say that progress is made through making mistakes. Yeah, so there's a little bit of me that wants to get rid of all the dangerous and and stupid mistakes, but a bit of them. It says we still need to be allowed to have errors because the errors make us go. Oh, that's interesting. I wonder what that's telling me. Or or that chart I just saw. I know someone's just told me it's wrong, but what would it take to get there, and so I think so, yeah, evolution thing, because evolution is a process of errors and some of them are big and non-viable and some of them are small and progressive. So I think that we need to find ways to allow error, but to make sure that we govern that which is truly dangerous and accept that AIs are certainly no better than people but we'll probably talk about this more.

Simon Hudson: 30:38
But this concept and, I think, microsoft, to their credit, because they get a lot of stuff wrong, but to their credit this idea of a co-pilot, something that sits alongside you, you work with it, it works with you, you correct its mistakes and, man, it makes a lot of those. I've been doing some coding today and it just makes mistakes, but I'm smart enough not to be able to write the code but to be able to go. That's not right. And then it corrects me when I get stuff wrong. So that co-piloty my buddy who works with me, my shadow, that has lots of merit and allowing ourselves to make mistakes, but then correcting each other is how we learn. Have you watched Silo? I'm five episodes in, literally started at the beginning of the week.

Mark Smith: 31:23
Just because you use that phrase shadow, and of course it's very yeah, I love it as in it's actually enhanced my keenness to get back into a bit of sci-fi, because of where we are, you know, in the world and oh well, I'll tell you what you know, going off off piece for a moment.

Simon Hudson: 31:41
Go and find all the books there's only three at the moment by stuart turton. It's sort of sci-fi but it's cleverer than just sci-fi. I'm a big sci-fi fan but stuart turton, top author, start the Seven Deaths of Evelyn Hardcastle and if you don't like it, I'll give you your money back.

Mark Smith: 32:00
So here's the thing I don't read a lot of novels, but I got into Dan Brown and the reason I don't get into them? Because my life goes on hold, because I've got to keep reading and that means all through the night. I just read and his most recent book, which was published maybe four years ago five years ago, okay, it's all set in Spain and how much of a predictor of AI that book was is off the Richter scale. Mind-blowing compared to where we are now. Off the Richter scale mind-blowing compared to where we are now. His ability to mix the real world with where things could go just honestly amazing.

Simon Hudson: 32:43
Yeah, there are some really interesting visionaries out there. I mean, you know the number of books that get published a year. Someone's going to get it right some other time. But I think Dan Brown's really interesting and he raises some interesting positions in his books. Are they great literature? They don't need to be. They tell a good story, great storytelling and they present some interesting ideas.

Mark Smith: 33:01
Yeah, I see we're already over time. This has been such an interesting discussion. I'm going to close with this. We start 2025 where the conclusion of 2024, we have been marketed to heavily by all major players that agents and agentification and agentic is going to change the entire world in 2025. And I believe that. But I have become more reserved in my thoughts around that of recent times and I suppose my reservedness has come about by my expectation of a gentic is that something will act autonomously after I press go, and my simplest example of that is my dishwasher. I press go, it's going. That is my dishwasher. I press go, it's going to clean my dishes. I don't have to go in and intervene most of the time and I'm going to wake up in the morning and there's going to be clean, dry dishes that can be put away in the cupboard and you have a high degree of confidence that it will do that.

Simon Hudson: 34:12
That's a key thing.

Mark Smith: 34:13
Yes, a high degree of confidence, right? I feel that with current ai right one, it has this uncanny ability to give up when it runs out of steam kind of thing. It doesn't have a what I'd call a heartbeat monitor on it. You know it doesn't go, it's it's still ticking over. Um, I've started doing my own gpts where I give it a books. I've taken all my kindle library over 300 books and what I've done is removed, drm off it, and because I'm only using it for myself, not public, you know and so then I've used it to create a intellectual property position accepted as in. So it allows me now to query it because what I found is that I the way my mind, mind works, I often have quotes in my mind which I have read and then I cannot find them and I want to attribute to an author, but that's not available in the public domain oftentimes. And so you know, one of the books I quote a lot was Brad Smith. He wrote a book a couple of years ago at Microsoft you know the head of legal counsel in Microsoft and there's some quotes that he has on data, and I needed to get to those quickly, and so what it's allowed me to do is drill in and get those.

Mark Smith: 35:24
But I've just finished a book called Nexus by the author of Sapiens and Homo Deus, and incredible view because it's like it's taught it was only published six months ago, so it's like up to date around ai and the effects, and he and he talks about knowledge in there and how knowledge has been used over time to and. But knowledge needed technology to distribute it, to make it effective. Right, the, the printing press, the telegraph and like the rise of communism and how that worked and how having networks of people that knew information that others allowed massive control. And his concern was that now you know, any knowledge in the world required it to go through a brain to repurpose it into something else. Right, you couldn't have a book write another book. It needed a brain to consume that book, write a different view, etc. And and advance it. Uh, ai is potentially going to change that paradigm and and change how information can transfer. Um. So, in other words, the communication and the technology are combining more and more and so it just gives this, really.

Mark Smith: 36:39
But I went to that book, put it through in person and said, listen, just give me a quote per chapter. And I said no, I said give me 21 quotes out of it. Well, it took 21 quotes from the preface, even though I directed it to do the whole thing, and it was like, oh, let's take the preface and get you. I directed it to do the whole thing and it was like, oh, let's take the preface and get your 21 quotes out of that, which were not great, not what I was looking for, and so that's, when I go back to the engendic world, I'm like there's still some problems to be solved here for me to sit and forget and have confidence that I'm going to get an output that I'm after. What are your thoughts?

Simon Hudson: 37:16
Yeah, well, I think you know Microsoft and everybody else, you know they have to create the market. That's what they're doing now. You know we're two years into AI and, yeah, and we're only just beginning to see the promises that they made in March 23, you know begin to be real. I mean there's lots of excitement about it. Some of the tools are great, but they're not without their flaws Let me be kind on that. So I think you're absolutely right that right now there are dangers of letting the AIs go and do whatever they have been told to do, because we may have told them wrong, we may have told them right but they might just not be up to it. So I think you're absolutely right to be concerned. I suspect, pretty confident actually, that the kind of things we're going to let the agentic I really don't like that word, but let the agentic technologies get their hands on are going to be the non-critical systems to begin with, and they're going to be done if they're done right. They need to be done with human oversight to begin with. I mean, this stuff's all in the Microsoft Maturity Model piece on AI anyway. But you know that need for human oversight and to work confident and put in place human-like processes. I talk about this every now and then. It was a guy from Capgemini, actually, that stole his idea, but he said AIs that take the job on of a person need to have the same controls a person would have. They need to have a manager, they need to have a set of objectives, they need to have performance reviews, they need to have things that are done to them if they perform badly. And so I think that that mindset of absolutely let's use agents where they're appropriate but not just make it a tech thing you plug it in or leave it to its own devices, but treat it just the same as another member of staff that needs to live up to their salary and expectations so I think there's something really powerful in that.

Simon Hudson: 39:05
However, there's a flip side. I did a session with do you know Bill Ayers? Oh, the name rings a bell Bill Ayers, longstanding MVP of the UK. Definitely someone you should get on your show. He's a fantastic, fascinating guy. Bill and I did a session back in Texas in April on AI doppelgangers.

Simon Hudson: 39:27
So because I'm intrinsically lazy although my dowry says otherwise, I'm intrinsically lazy. I want an AI, Simon. I want a doppelganger of me that can do some of the things that I can do. I want it to know the stuff that I know. Going back to all the points I've made earlier, I want it to have the same ethical base as me, so when it makes decisions and takes judgment, it does it in the way that I would. I want it to be able to take some of the load off me, um, but I I also want it to be able to free me up yes, being constantly on call to people so they can do the thing. Let me go off and live this wonderful life that I'm probably living already, to be honest, um, but I'd like to be able to do that. Yeah, do that earlier. So I want those kind of agents. I want my ai to have enough agency to relieve me of the some of the drudgery, letting me go on and do more, not, yeah and of some of the drudgery.

Mark Smith: 40:15
The Edmund life Edmund.

Simon Hudson: 40:17
Yeah, and just some of the turn-handle stuff, because I need to spend my time doing things which are genuinely more creative and insightful. I need to be reading those hundred books and going oh that's interesting, because that book intersects with this one and here's a brand new idea that comes out of that. The AI can't do that, but it can. Let me have the time to do it. So that sort of part of my dream is embedding our personal morality in the machines, getting them to learn from us as individuals.

Simon Hudson: 40:45
Being my buddy that can look after my affairs when I'm asleep or drunk or enjoying myself Could be all three of those at once, who knows? Or enjoying myself Could be all three of those at once, who knows, yeah, and letting us move towards that more utopian world. That, I think, is actually a genuine possibility. You know, we're A hundred percent. We might be living in the. You know we might be the last generation that has to buy energy, because we're so close, if we have the will and the political control, so close to having, you know, having unlimited energy for all of mankind if we don't screw up the world in Cleveland Zero-cost energy.

Mark Smith: 41:21
I believe that's definitely going to be one of the solves that we're going to get to, and that excites me because we saw what energy did to the last 150 years and it was fossil fuel energy then, and it came with a price, but the invention that came from it was mind-blowing right. So I think you're on the money, Simon. I think we could go on for hours. Thank you so much. This has been such an enjoyable conversation.

Simon Hudson: 41:51
It's been an absolute pleasure, mark. I hope this reaches out to people in the way it's meant to and I look forward to it. Are you going to MVP Summit? Absolutely Good, I will see you there as soon as I book my air tickets. I'll look forward to it. Lovely Cheers, simon. Good, have a great day.

Mark Smith: 42:07
Hey, thanks for listening. I'm your host business application MVP Mark Smith, otherwise known as the NZ365 guy. If there's a guest you'd like to see on the show, please message me on LinkedIn. If you want to be a supporter of the show, please check out buymeacoffeecom. Forward slash. Nz365 guy. Stay safe out there and shoot for the stars.