AI That Knows Your Code Better Than You Do

00:00:01 Mark Smith
Welcome to the ecosystem show. We're thrilled to have you with us. Here, we challenge traditional mindsets and explore innovative approaches to maximising the value of your software estate. You don't expect you to agree with everything, challenge us, share your thoughts and let's grow together. Now let's dive in. It's Showtime.

00:00:21 Mark Smith
Welcome back, everybody. Almost at full house full complement with the four of us on the show tonight. We're only missing will. You can't join us here is all set to come. And he's had something come up. Hopefully he'll be on the show with us next time. But. What are you guys doing in the wonderful world of Europe and Northern hemisphere and miles away from me in the southern hemisphere? I'm in the middle of winter. Yeah, I just had the winter solstice the other day. And that means you must be on your hottest day of the year, are you?

00:00:52 Chris Huntingford
We we are at here in in sunny Valencia. I think that the temperature is going to touch 34°C today. You know so so it's it's another day. Another day of summer and possibly another random parade through the streets of Valencia. I think we've had one at least one a day. For like half of the days in the last several have been, there's been some sort of random parade or something, so bit busy time for the street cleaners here. Nice. I feel like I feel.

00:01:27 Andrew Welch
Like we were waiting for. Some sort of a story.

00:01:30 Chris Huntingford
Well.

00:01:31 Andrew Welch
But thank you for saying, Bill.

00:01:33 Chris Huntingford
No, no, no. No, no, I I I have, I I mean I mean well, so I did. I did take a video of the of the religious objects being paraded through the street and escorted by armed guards. And I sent it to. I sent it to to Chris but.

00:01:40 Andrew Welch
I. Didn't I didn't. See that? I do love a parade. I'm not gonna lie. I don't. I don't understand why every time I wanna throw like a random parade. You guys are all like, no, you can't do that. And. Then. You know, it's not OK, so when I start requesting random parades, I would expect the answer to be yes. From here on out. Thank you.

00:02:00 Chris Huntingford
What? Just the the four of us walking through the street with great, with great fanfare? Or will we hire a band?

00:02:06 Andrew Welch
No, this is the solo parade you can watch.

00:02:09 Chris Huntingford
So is this Chris? Chris was.

00:02:13 Mark Smith
The last time I saw her parade, the lady was calling out. Shame, shame, shame. Yeah. Yes.

00:02:21 Andrew Welch
I definitely. That's what I'm gonna start calling walking now. Just in general. I'm gonna be like, I'm gonna go on the solo parade. Guys, I'll see you later. What are?

00:02:27 Andrew Welch
You doing going for. A walk? I think that's, I think that's totally fair. Yeah. So go. I'm gonna go into solo parade on the treadmill. And then. Yeah, I think that's I think. That is fair along. A stationary solo parade.

00:02:42 Mark Smith
Last week.

00:02:43 Chris Huntingford
So the Gartner, the Gartner report.

00:02:45 Mark Smith
Well, wasn't the Gartner, it was actually a Gallup report that said that, you know, workplace adoption has doubled in the last 12 months on the use of AI found that interesting going from 21% to 40 odd percent.

00:02:47 Chris Huntingford
It was OK.

00:02:59 Mark Smith
In its use and I don't know about you, but I'm starting to come. Up with very. Creative ways of using, particularly M365, copilot and and with its ability to access all the history and conversations you've had with clients and things like that. In our case, the work we've done for them and being able to extract that and get it into a format to perhaps update stakeholders. And things like that, the time it's saving. Is absolutely phenomenal and Chris, I know you've been working on some amazing stuff for one of our customers in this space and. What you showed me today just honestly mounted my face.

00:03:42 Andrew Welch
It's, you know, you know, The thing is like, this is this is something that I think. We're getting, like, wrong in society, but right individually. And I'll tell you what it is like. I think I've realised what AI does for me, and I've had. I've had a couple of, like, groundbreaking moments, but this project was a groundbreaking moment. I'll tell you why. I struggled to get a lot of the people I needed to on phone calls. OK, So what I did and and it's really hard, right and and a lot of them like the information is siloed in, in different, in different places and things. So you know what it did? I basically built my own custom copilot on top of their dev OPS data, right. And yeah, it's wild. So I took their dev OPS data, which is about. 7000 work items I took the documents that they had in their team. It and then I consolidated and I basically made this repo where I was asking questions. Now I had calls with them really good calls by the way, but it's very fragmented like one guy that's doing integration doesn't know what the other person's doing with another set of integration. But what this let me do was combine that knowledge. So I was basically I had I had my copilot open, M365 copilot.

00:04:45 Andrew Welch
On the screen with my like knowledge repository of all the dev OPS stuff by the way, including custom code. So I was able to execute a full custom code review with recommendations using tools like copilot and my own knowledge. Right. So you have. Sense check, but I basically treat it as like a. It's like an intern, right? So I said OK, I have all this stuff. I have all this information. I can't get all these people on the phone all the time. How do I? How do I work out what is actually going on in this system? Because it's and. And I've done these all this before, right? Like they can get pretty deep. OK. And if I didn't have that information, if I didn't have those tools, I can categorically tell you, the reporter would have generated would have been nowhere near what I got to with AI, OK?  And it's 120 pages and it's 120 pages of productivity information like actual detail around like the physical project being run, the code inside the and I didn't do review all the code by the way, I picked like samples of it because you can kind of pick up patterns and say OK, like this is what it's going to look like. But what was wild is that full project review, full code review. Full tech review based on what was inside Azure DevOps inside some of the documentation that was shared, and then I also took call information from them. So I had a bunch of call notes that I had and when combined I was able to bring this information into a really clear format. Obviously without tons of spelling mistakes because that's how I roll. But it's it's interesting because.

00:06:03 Andrew Welch
I've kind of realized that actually if I didn't have those tools, I could never have done that. I could never, ever, ever have gotten that information. I could never have found the things I found, which is absolutely wild. Now. I've got to give kudos to the customer. They have spent a bit. They have spent a bit of time inside Dev OPS like tracking those work items and descriptions and things, but it's.

00:06:23 Andrew Welch
Ohh like in my brain in tools like that it's still very fragmented. It's when it's brought together and when you're asking at the right things like the prompts I'm doing. Man, I'm not sure. They're probably like 20 line. Prompts. To an extent and and, you know, Anna and I as well. The other day we had a workshop with another organization. We were basically prompting a I live, we had a menti meter.

00:06:33 Mark Smith
Yeah.

00:06:41 Andrew Welch
The customer answered questions on the main meter. We took the information, walked into a GPT, looked for patterns, and and how how accurate were the patterns.

00:06:49 Ana Welch
It was just so accurate, like immediately and then going back to your DevOps work, I think that tells us a lot of things and it tells us a lot of things because not only you get the information about the the projects. So you know, I don't know what sort of features were built, what sort of features. Or stop to mill midfield. Or what sort of I don't know, information they had at the time when they decided to do when they decided to do. Something, but in my opinion the most important thing that you can that you can have with these tools and with AI is actually the human insight. So isn't it amazing to see that, oh, actually 80% of the work is being done is being achieved with this team and in the world of AI? We do have a a junior dev problem and we talked about this before. I I I believe you know and how do you get to see? Well, with that, with those insights, you have actual metrics and numbers around the fact that, OK, so this is the sort of model that we need. So maybe this team should be working, you know, more closely with our junior developers or with another team or like kind of do up skilling. And a bit of culture built like that as well. I think it's really important to bring back all of these AI goals to the components of the team, to the fabric of of an organization, to your culture, to what do you want to achieve as well as obviously getting to the right result for. Your for your project. I like it.

00:08:32 Mark Smith
I just want to say 1. Thing for those people are one. During, you know, is this just a two person project or three person you know working on what we pulled out Dev OPS is over 40 plus developers were on this project. So we're talking about it was sizable and and and and there was a lot of data to work with.

00:08:46 Andrew Welch
It now was. Yeah, 5050 fifty people. But the cool thing was the really, really cool thing was when we extracted the data we had. So this is where it gets insane, right? So we had, we had it in copilot doing, looking for patterns. But the real magic happened where we pulled into fabric so built a whole bunch of custom BI reports on top of it. OK, inside fabric. Using fabric copilot and then we said some fabric copilot were like OK. Show me. Show me patterns that we should be aware of for this. And then it was like it was like, OK, the biggest issue, this is what it told me. I will read this out to you. It said one of the biggest trends we have noticed with this program of work is that roughly 20% of people are doing 80% of the work which means that that knowledge is now siloed with those people. That's what it told me it literally. Gave me the insights. And I'm like, how? How does it know? And then I started then then my brain really started going like, how does it know what a work item is? How does it know what? A bug? Is obviously it's been labeled like. It's just works right? Like what is? What does that mean from a contextual perspective? And then my brain would, oh, should I get this now? It's because because I've been looking at AI as ones and zeros. Like that's, that's how I've been looking at it. Right. And I always you guys know I always put the human spin on it. I'm. We have to have the human elements in this, but now I've really understood it. It's it is actually. It is actually intelligent like it it knows, right? Like it's got. But the only way it knows is you've primed it really well and really accurately with really strong clear data. And that's the only reason I believe that we were able to get the responses. We did because we knew. I knew what the data looked like and I didn't have an expectancy. But when the responses came out, my mind was totally blown. I'm like, this is so powerful and now I've completely flipped the way I use artificial intelligence. I actually can't get through the day without it.

00:10:36 Mark Smith
Cool. Yeah, yeah.

00:10:38 Chris Huntingford
You know, you know speaking, speaking in ones and zeros is how Anna and I need to discuss whether we're going to give our daughter ice cream nowadays, because she's, she's learned, she, she's now rather trilingual, including in Spanish, so we can't speak, but we can't discuss ice cream having in Spanish anymore. We have to speak. Binary.

00:10:59 Mark Smith
Classic is incredible how quickly kids pick up stuff. As in I I was in the same position. I've been spelling words to Meg so because she would understand what Meg the poppy I'm talking would understand what it was and then she just turned around the other day and responded to exactly what we had spelt out in the conversation she had picked up enough to know what we were talking about.

00:11:00 Andrew Welch
Right. Yeah, that is wild.

00:11:03 Ana Welch
Oh my God. And then.

00:11:21 Mark Smith
And.

00:11:22 Ana Welch
That's amazing. That's amazing. Yeah.

00:11:23 Mark Smith
I love kids. So much fun.

00:11:25 Chris Huntingford
And that is and that is the story of how Alexandra is going to pick up a little Swedish because we have started to substitute like Swedish words, including for ice cream, because it turns out that speaking in ones and zeros is not terribly time time efficient for humans.

00:11:43 Mark Smith
Last week we talked about AI optimists, AI deniers and AI rages. From memory, Andrew and we really didn't unpack that. And I actually just happened to go back to it today and and I was filling out a report based on stuff and dynamics mines that we had done, you know, 3 or 4 weeks ago and.

00:12:00 Mark Smith
What are you seeing? What are you guys seeing around these different camps that people seem to be, you know, falling into where we are very much optimist about where the where AI is going? Oh. I feel sometimes I'm a minority.

00:12:19 Chris Huntingford
I think let let let's very very quickly define what what we're talking about. So the AI the optimists. The optimists are, as we you know, self evidently say, you know, optimistic about where AI is going. The existence of AI, the use of AI, we are, I think generally fairly measured, right. There are some that I think are just unbridled but but I like to think that the optimists are putting energy into solving the problems of AI. Not simply complaining or raging against AI, right? Then there are the deniers, the deniers, I think, are folks that.

00:12:56 Chris Huntingford
But you know, you see this manifest in a few ways, but but it's folks who are very deeply concerned about what AI might do to them or to their organization. They, I think, generally sort of feel that either AI is here to shore up their legacy business or. AI is a passing thing, and then there's there's the ragers which I I find to be most frustrating at all of all the. Just seemed to be angry at everything that Microsoft does, or that another vendor does. You know, they several weeks ago that Apple, the report from Apple that came out, which I thought was actually quite funny, these folks held this report up as evidence that AI is garbage, right?

00:13:37 Mark Smith
Yeah.

00:13:45 Chris Huntingford
That even though essentially what the report was saying is that AI cannot reason past the data that it has. Which, well, of course it can't reason past the data that it. Has how are you? Terribly good at that yourself, folks. So I that's the context and that's kind of the quick catch up for Anna and Chris on the conversation, Mark and I had last week.

00:14:07 Mark Smith
Just to bounce on that right and I have become increasingly frustrated with Apple. One, there was the law case where they have willingly given over audio data from your phone and sold it to advertisers and they settled. The court case denying any responsibility for something like $92 million up to that point. They were my second most trustworthy brand. I always considered Microsoft and Apple. These bastions of and tech of trustworthiness and that kind of destroyed it for me. For them. Now they came out the week they came out with. That was their big tech conference and what they released and updates to the iOS software or. Always software, right? I was sorely disappointed and I felt like they had to lead with this big AI story to kind of deflect over the nothingness of what the the bring out a glass experience on icons that which make it super hard to read within 48 hours of that report coming out, open AI released are 03 pro reasoning. Model and inside that Apple report they had all the test prompts that they had run to prove their theory, and an individual on YouTube took all those test prompts run against O3 Pro and it totally destroyed all the findings of Apple. And and and one of the biggest ones is that there's this 3 pin object where you have rings, concentric rings on it, and how many moves does it take? You. To move all the circles from 1 to the third pin.

00:15:37 Chris Huntingford
The the like the tower of.

00:15:39 Mark Smith
Yeah. And it's basically 1023 moves it takes to move that to there. They proved Apple proved it couldn't do it. It couldn't reason that out well. Yeah. Pretty rapidly O3 Pro the the absolute like reason that the whole thing out and and gave the answer in you know in minutes and so it kind of destroyed so many of their test reasons to prove it wrong. We're soon disproved themselves.

00:16:08 Ana Welch
From my perspective as frankly, I believe that the Rookie of the Group, I have to say I don't have, I don't know 20 line prompts that I use. Every hour of every day, you know, so I. Use. Basic artificial intelligence, right. And even with basic artificial intelligence. Which is like. Simple prompts, making sure that I didn't forget anything before a meeting, ensuring that when I when I write an e-mail, I answered all of the questions in the e-mail to like reduce. Turn around time, making sure that my correct calculations are OK. Making sure that my estimate based on previous experience kind of looks OK you know all of that stuff. And I feel like even at that basic level, that AI is starting to ration problems that we would, that we never were able to solve. And what I mean by those problems, I mean, like, human problems like AI is improved on my my relationships with my colleagues. With my with my client. With with everybody you know, because I am able to give accurate information and forget about stuff and also explain things in a way that people would understand. Oftentimes we do communicate at executive level and then it's our job to bring the whole team with us. And that's a really, really hard thing to do. So I guess you can say that AI is rubbish or all you want, but the reality is that. It's just improving our life. Yes.

00:17:57 Andrew Welch
Yes.

00:17:58 Chris Huntingford
What's behind? What do we think is behind the rage camp, right? I I think that.

00:18:05 Andrew Welch
I can tell you.

00:18:06 Chris Huntingford
I think that the optimism camp and the denier camp are fairly straightforward, right? But what is it that makes that is making folks and I'll give you a a concrete example. Right, I've seen a lot of chatter in the, you know, in the in my LinkedIn feed and amongst conference goers at at at a recent event that I was at where folks are furious that Microsoft is pivoting its product direction. Away from some of the more traditional investing in some of the more traditional app dev use cases, which I'll summarize as forms over data and pivoting that more in the direction of. An agentic or a generative approach, right? Like they're beginning to change the way that individuals interact with, with with this technology, right, the user experience, and to me, I I think that that's entirely positive. I don't think that it is Microsoft's responsibility to preserve. Legacy tech simply because you like it, right? But I I I don't. I just fundamentally don't understand this rage that is out there. What, what? What's going on, Chris?

00:19:13 Andrew Welch
Yes. It. It's fear, dude, that the answer is fear. And that's why people are raging. It's very it's. We've seen this a million times. Right like. It's the, UM, it's the whole what was the Ned, Ned Ludd? Whatever the Luddite thing. OK, the the dude that burnt up the whole factory because he didn't want to be replaced by a loop. And like it's it's fear. And the thing that's interesting for me is that.

00:19:45 Mark Smith
The weaving limbs, right and yeah.

00:19:47 Chris Huntingford
I like the way I like the way that. You say loom, loom loom.

00:19:51 Andrew Welch
Loom and it's interesting because that's what's happening now. Like people are I I said this. If you guys remember probably around a year ago we were talking and I'm like there's gonna be villagers with pitchforks scenario here where people are gonna rage and they're. Gonna get passed off, but dude. It's. Yeah, like it's you can rage as much as you want, right? The the trick isn't dynamics. People did this when our platform came out, they were like, ohh, rage, rage, rage. And I'm like, dude, get out of the truck of disillusion and go and learn something new and just pick up the tick and what will happen is those people will become. Full behind it's a sphere.

00:20:22 Chris Huntingford
This is once again, though this is once again and and I I I hate to. Who I I, I I hate to pick on dynamics folks, right, but this is this is where I'm seeing perhaps the most of it, right? It's like these ridiculous claims that I see coming out of this is the Golden Age of biz apps and actual slide that I saw. From a pretty senior person at a conference recently, and I'm gonna try to anonymize this a little bit here, but the golden Age of biz apps.

00:20:53 Mark Smith
Yeah, we shall let. Me know.

00:20:56 Chris Huntingford
And and I I I thought. To myself, as I saw this. What **** are you eating? And I I try not to swear on the show, but like, what are you talking about? This is not the Golden age of mishaps. Satya and Charles Lamanna are out there talking about how Sass is dead or, you know, the the, the the era of of business applications is behind us. And then you see people coming out with the. Golden Age a bit. What is going? On over there.

00:21:25 Mark Smith
You know what? Chris says it's fair. I think it's more than I I think there's other things besides fear. I think that one of the big things that came to mind when you asked that question of me just then was him and *****. From who moved my cheese? Right and and. Well, when the cheese starts running out, they take one or two things. One they go out and find the new cheese, the new opportunity and go after it or others just sit there and just say woe is me. The sky is falling, the world is ****. I have to learn something new and I think fundamentally the word I would use is people don't like change. They don't want to go away from the status quo of what they know. I I totally believe what Chris was saying about fear. That's part of it. But change, I think, is the other people hate change. A lot of people hate change, right. And they don't want to adopt something new because that means effort. I've got to apply myself. I've gotta learn. I've gotta stretch, you know, I become the junior again. And why for ages on my screen I used to have wiped out for life. Where a lot of people don't have that mindset that you do sometimes need to go to the bottom again and start working your way up. And we're in one of those cycles right now. Everybody has a chance to get ahead in this space. If they're exposed.

00:22:36 Ana Welch
Right now.

00:22:37 Mark Smith
To it. If they learn and apply themselves. But then there's a another camp that I've seen, and it's come from a creative friend of mine. Designer. And he. He hates that. It can do what it does so well because.

00:22:56 Ana Welch
You're.

00:22:56 Mark Smith
One people don't need them to do to do what they would have normally come to them to do, because the tech can do it.

00:23:00 Ana Welch
Exactly, exactly. But these individuals and a lot of them and many people in our community and unity in general and like doing the work, do you know what they they feel like? They feel like they're special. So we I I have this. We watch Bluey a lot because we have a three. Year old and then.

00:23:22 Andrew Welch
There's this is the best man.

00:23:24 Chris Huntingford
Bluey is the best.

00:23:25 Ana Welch
There, there's this like episode where this little dog called Muffin. She goes to play with her friends at a play date but. And Muffin doesn't really play nicely. She doesn't share her toy. She doesn't want to learn how to play the game that the other two want to want. Want, want to play? And she's just like a rolly and her dad goes like muffin. What's what's going on? Do you wanna go home? Like what? Happening. And she's like Daddy, but they don't want to do what I want to do. And he's like, OK, great. But still like, like, can you take turns, like, what's going? On and she's like. Well, I'm special and this is what's happening. They don't want to play with other people because they are special. And if you are that special. You really don't like the fact that a tool can create better outcomes than you than you yourself are able to produce.

00:24:22 Chris Huntingford
Yes. I I I think I I just I think that no matter how. No matter how. Special you think you are? And by the way, actually, I would argue that if you are actually special, the thing that makes you special has been your cognitive ability to learn to, to learn the thing. To become the best at at the thing. So go figure this out. Raging, raging is going to get you nowhere, right? Yeah. Raging that Microsoft. Nor any other vendor is going to say, Oh my God, those folks are really upset on LinkedIn, that we're we're deemphasizing our our investment. In legacy technology, maybe we should, maybe we should dial up our forms over data offering again like. It's just crazy.

00:25:15 Ana Welch
The other thing that the the other thing that you know the way just forget is the fact that you know all of this technology is created by really, really, really smart individuals. There is a group of big brains with like a lot of patterns, with creativity within measures like they are able. To like take a concept and transform it into into something totally new. And now with with the help of AI you can do that even faster. So you're you're going to become even less special, you know, by not learning the tool and by believing that actually your skill set is unique, it will be unique. For this particular project and it was you, it was unique until last week. Week because last week Chris Huntingford figured out how to get into your systems and to tell everybody what it is exactly that you're doing that makes you successful. Even with disparate data sets, even with lack of, you know, communication and availability to do calls. Information scattered across emails or excel spreadsheets, or like whatever Chris is able to tell you a story as to how to do that job better. Right. You are no longer specially stunned.

00:26:35 Mark Smith
How soon? You know, for the rages out there. And. And here's a common rage I see. Ohh, you use the long dash. You know in that post or in that word which is called the EM Dash, which the reason it became a a modern phenomena and a hate on AI thing is that. On the keyboard, there's no EM key, there's no long dash key. It's something you put in if you no good. Right. And so same with the Oxford comma, right. And people go, oh, that person's using the Oxford comma. They are obviously using AI to to write all they're using the EM. They're obviously using it. And now I've just gone through a book publishing process and all the way through my editor is going. That should be an EM. That should be an EM. That's. Not the you used a hyphen or a comment. No, that should be you're you're bringing to come? Yeah. And so they're putting the EM in. And so I was like, ohh, what's your thoughts on this and that just like. Just because people are petty about whether they think something AI has created it doesn't mean you throw out good grammar in good English just because the AI is producing it.

00:27:43 Chris Huntingford
So I can say unequivocally I writing I I love to write, I love I I love to write and I love to write multiple forms. Technical, but also fiction. I've published a book. I I I have never ever once published publicly. Content that has been generated by AI 100% of it, and I'm I'm talking down to the LinkedIn Post. I create my own when I post on LinkedIn, which is not daily right. But when I post something on LinkedIn I have written that those paragraphs, right, I can unequivocally say I've never republished. Content provided to me by AI. However, I have for a very long time used the long comma and the the the long the.

00:28:26 Andrew Welch
Yeah.

00:28:32 Chris Huntingford
One dash and the Oxford comma just as a natural part of my writing, so it actually annoys me to no end that this thing that I've been doing for decades, this this grammatical flourish that I've been using for decades now, is associated the the the Rangers associated with AI generated content because it's not.

00:28:37 Mark Smith
Yeah. Correctly.

00:28:53 Chris Huntingford
I can promise you.

00:28:55 Andrew Welch
I will tell you something interesting. I had this conversation with somebody very recently, and they're like, so I like you, Andrew, don't use AI for LinkedIn posts and things, but I use AI in pretty much every other form of writing that I do every other form. Like I published the book, we used AI to do research. We used AI to kind of form structures and things and somebody said to me, oh, did you use AI? I'm like, you should be worried if I didn't use AI. And that's the thing. And this is what I'm telling everyone. I'm like when when we go through these processes. Now, if I interview people, if I talk to people and they're like, oh, we we try and avoid using it as much as possible. I'm like, OK, that's not good because I think in your job you should be using it, not. I think that AI shouldn't remove creativity. That's important. What you do is creative, Andrew. But what it should do is it should aid you in doing your job. It should, like I was. We were able to build 120 page assessment report because we used artificial intelligence to do a lot of research.

00:29:51 Andrew Welch
Review it didn't change the report. Yeah.

00:29:53 Chris Huntingford
Absolutely. When I say that I don't use AI in creating narrative right it it means that I use AI as a research assist, right? But the the creative content is all generated by my brain, not by not by AI. I use AI as an input to my own thinking. Yeah. To gather information. But then the thoughts are my own. Yes.

00:30:17 Mark Smith
But. But that aside, how soon before that just becomes like? You know, like, So what, like like?

00:30:25 Andrew Welch
It is it is, it will. No.

00:30:26 Ana Welch
No, it it is a. So what because and and you know why? Because Chris is right. Some people are like ohh, I try to avoid it as much as possible. We try to do this. We don't really trust it, but we don't. They don't understand, right.

00:30:28 Mark Smith
It will. It is, yeah.

00:30:42 Ana Welch
And then you get some people who are very black and white and also incredibly good at their job. And then these people, very, very black and white and incredibly good at their job, they invent something, right? Something that saves their company a long, long, long time. And I asked it. It happened right there was this person who created an entire structure that saves her company 20 minutes. For each interaction with the customer, that's a lot of minutes. So. So yeah, so how did you, how were you able to create this new product? And she was like, well, you know, I did have from like some of my team got what I wanted to do and others didn't. And they were like, super angry. And even now that they're using the system, they're they're complaining all the time. And I'm like, OK. So what do you do with them? Oh, I just remove them like clear is there. So I just remove them. I just work with others. This is what's going to happen. You're right. I so agree with you, Anna.

00:31:47 Mark Smith
So so very. Very interesting. I was at a I was speaking with a bunch of educators last week on on a project on behalf of Microsoft, and the subject came up of cheating at university to use AI.

00:32:00 Andrew Welch
I'm at Uni and I use AI. When I do my. When I do my modules now.

00:32:05 Mark Smith
And and here is what one of the educators stood up and said to the group. And now we're in a group of 40 educators representing all the universities in New Zealand, etcetera, he said in a student was responding to a question. That had very the student had very bad dyslexia and so he output his ideas and then asked AI to summarise them as the answer. The professor was like, you're cheating. Why? Why you? That's just straight out cheating. And he goes. And so he got hauled over the coals for cheating, right as he got. Pulled up and so he said OK, here was what I gave to the AI. Yeah. And they're like, well, we we can't understand this. It's just, you know, gibberish he goes. Yes, it's my dyslexia. And that's what it looks like. And that if you had submitted that, you would have failed me because you're you can't understand it. The AI interpreted accurately what I was saying.

00:33:00 Andrew Welch
Yes.

00:33:01 Mark Smith
And this is the output and so this person was saying that kid didn't cheat like this is the educator. That kid didn't cheat. But one of the problems it got in universities. Is that individual professors or lecturers, et cetera are going? That's cheating. That's not cheating, because there's no kind of overriding mandate of the organization to go. You know what this person gave the correct answer. It's just that they have a learning disability that didn't allow them to necessarily put in a format.

00:33:29 Mark Smith
That you, the marketer, understood, so they put it in a format they used tool to help them. Like we've had Grammarly for ages, no one said Ohh Grammarly is your you know it put the commas in and the and and the periods in the right place and told you when you're using the wrong there because that's something I do all the time. Is about a GRE or THG IR. Get them mixed up. I just well yes, I've always. I I use them out of context, it corrects that. Shift for me, right?

00:33:55 Andrew Welch
So Mark, I go to uni, I'm at uni, right and I used AI in a lot of my responses, OK, I didn't get hold over the calls though. Boom. At all, because I kind of I kind of used it in a way that I was able to build responses to answers cause I stuck. So I have this thing where I'm very I am just lexic not as not as ******** as you are but like I do. I do miss shape words and sentences but I have a problem of I have the problem of context switching in in the opposite that other people do like I do it. So naturally, and too quickly that like I'll be talking about one thing and then I'll switch to another lane, talk about something completely different. And people are like, how did you make that correlation? I'm lucky to have you guys around because you know that I do it right, but in uni.

00:34:34 Mark Smith
Yes.

00:34:39 Andrew Welch
Dude, they're like, how the **** did you get from that to that? And in my brain, it makes total sense. I call that rigid. Right. And This is why I love doing pre sales cause I can move between subjects so quickly. But the problem is is I don't evidence it. So what? AI's help me do is evidence the movement.

00:34:44 Mark Smith
Yes, yes, yes. Yeah, yeah, yeah, yeah. No.

00:34:58 Andrew Welch
And this is what happened in that document that I did recently, and I'm like, holy ****, now I can see it like I can physically see what other people see now. And I'm like, actually, I don't need to auto correct every time and I don't need to feel ******. I can just go. OK, I can switch between the subjects and it fills in those missing pieces that I naturally just don't.

00:34:58 Mark Smith
Yes.

00:35:17 Andrew Welch
Do.

00:35:18 Mark Smith
Yeah.

00:35:19 Andrew Welch
It's insane, dude. Yeah, it's really insane.

00:35:22 Mark Smith
I love it. Listen, it's been great to. Have you on? The show, folks listening, we appreciate it. If there's something that you would like us to surface on the show, just reach out. Let us know. We are keen to. Get your input. Chris, been good chatting. We'll see you on.

00:35:36 Andrew Welch
Yes.

00:35:37 Mark Smith
The other side.

00:35:37 Andrew Welch
Too, man.

00:35:38 Mark Smith
Thanks for tuning into the ecosystem show. We hope you found today's discussion insightful and thought provoking and maybe you had a laugh or two. Remember your feedback and challenges help us. All grow so. Don't hesitate to share your perspective. Stay connected with us for more innovative ideas and strategies to enhance your software estate. Until next time, keep pushing the boundaries and creating value. See you on the next Episode.