Episode 99

Matt Bryan on The Transformative Power of AI in Hiring and Data Analytics

TIM: We are live on the Objective Hiring Show with Matt. Matt, welcome. Thank you so much for joining us.

MATT: Hey, thank you for having me.

TIM: It's our pleasure. And where I'd love to start is by hearing a little bit more about yourself. Who is Matt? Who are we speaking to today? Who are we listening to?

MATT: Yeah, so I am the director of analytics and insights for Samsung ads across Europe. But I think the main thing for the audience of this podcast is that I've been hiring analysts for 10 years and working with analysts and data engineers and data scientists, all with the goal of saying, how do we take data and turn it into meaningful value for businesses? A whole bunch of experience working with analysts, hiring analysts, but really understanding, as an analyst myself, how can we make a difference?

TIM: Where I'd love to start today is hearing more about your experience as a candidate. Is there anything that springs to mind in your own experience in getting roles that was a little bit unusual, either for a good or bad reason?

MATT: Or, as a candidate, so straight away, sorry, in my head, I've got the hiring manager hat on, and let me just quickly go through that one. I had a great experience. recently, where someone just did their research. We were taking them through a bit of the hiring process where they have to pitch to us, where they have to show that they can interact with clients, and they started using, basically, everything I'd said as part of a keynote speech just a few months ago. And they'd gone onto YouTube, they'd done their research, and they were talking the same language as us. If you want to surprise someone in an interview, I hope that's something I can say about myself when I've gone for roles. But yeah, I'd also say something that I found surprising, and it's a good thing to prepare people for. I've been through hiring processes where I've had eight rounds, seven, eight rounds; sometimes you've got to be prepared for this stuff. Data is such a fundamental part of a lot of businesses, but there's a lot of uncertainty when you're hiring managers; quite often everybody wants to talk to you, and they're a bit unclear about what they want, and it can lead to these multi-round interview processes.

TIM: The first tip is a great one and maybe one that's almost overlooked now because if you're a candidate and you were Googling, like, I don't know, job tip research, this kind of stuff. It'd probably be inundated with AI-based content and tips, and here's how you apply en masse, and it's a numbers game, and blah, blah, blah. But the fundamental basics of, yeah, doing your research, knowing who you're speaking to, and knowing your audience are probably more true now than ever, you could argue.

MATT: I think you're absolutely right. It's all for me; it's about being able to talk the same language as the people that are hiring. If you can do that, and quite often it is a case of doing your research, you'll put yourself in a fantastic place compared to other candidates. Every hiring manager is looking to say, Can I see you fitting in this team? Can I see you doing the job tomorrow? And if you want to achieve that, do your research. It will pay off. And in an online world, you're going to find loads of videos about that business. So it's a possible thing. It's achievable. Do your research.

TIM: And what about from your own experience on the other side of the table, as a candidate? Have you ever had any, yeah, any processes that were just a bit unusual, any interview experience that you were left a little bit scarred by perhaps, or anything really good?

MATT: So I think we've all got war stories. I think the most common one that everybody can probably have a thought about or have some experience with is actually when you just get no feedback. When you do all of that work, you know I'm saying do your research, you do all that work, and you go through a process, and then you find out, oh, you didn't get the job, have no real feedback, why? I think that's the classic challenge for candidates. We put so much work into just the application, and you know your CV might get rejected straight away, but writing that CV takes time. And then you get in a process, and you have interviews, and maybe it's stage after stage. I think, as a candidate, and everybody's probably got a similar story, when you've gone through that investment of time and effort to not get any feedback that's going to help you for the next job you apply to, I think that's a problem in the hiring process. And that's something that all hiring managers really need to be conscious of. People have invested a lot of time into this. We need to make at least the kind of value exchange of it, make it fair.

TIM: Yeah, for sure. And I feel it's really depending on where you got to in the hiring process that would dictate what type and quality of feedback you'd expect. I don't think anyone at the moment at the kind of CV stage would expect anything other than a yes or a no. But if we're talking about several rounds of interviews, surely you should get some closure out of that.

MATT: Completely agree. Completely agree. Yeah, it's, I think there's many ways you could argue it. Some people say to me it's a respect thing. Almost. I've spent an hour with you. We've had this conversation. Just, even if it's that first round interview, give me something of, Hey, you didn't meet the standard, I'm afraid, or whatever. But yeah, the days of zero feedback I think we've all been in that position where we had the interview, and there's zero follow-up, not even a no, you didn't get the job. Yeah, I've been through those experiences. Thankfully, I've been through other experiences where someone has given me the role. But I think we've all had that kind of you didn't get a role, but you just don't hear anything back. And that's something I think, just as hiring managers, we need to hold ourselves accountable and make sure that doesn't happen.

TIM: I can recall one myself right at the start of my career where I really needed a job. And I got into the final stage for this financial analyst role with a financial modeling business. And I'd met the partner. He said to me, You've got the job. Don't worry. Just come in tomorrow to meet the MD. Have a chat with him. I traveled. Two and a half hours each way on the train to Sydney to meet this MD. Met him in his office, had a conversation of not more than 10 minutes, mainly centered on football. I remember because he supported Tottenham Hotspur. So maybe this gives you a sense of his character. It

MATT: I'm an Arsenal fan. It's never going to work.

TIM: Yeah. Exactly. I'm pretty sure I might have ripped slightly on Tottenham or made some reference to Arsenal at that point. So maybe that's where it all got derailed, but I never got the job from that point forward. It was a 10-minute conversation, nothing to do with work. And somehow I missed from one yard out effectively. I still don't know why to this day, other than, Oh yeah, you didn't get the job. So I would love to know, because maybe it had nothing to do with me. Maybe it was, they pulled the role, they had budget changes, they'd gotten a recommendation from someone else, and hired them. Like it could have had nothing to do with me. But of course I would internalize and think, Oh shit, what did I do wrong? Where could I have changed the outcome? But with a simple phone call and a one-minute conversation, that could be. And I feel like if you're like a senior leader of a company, you've got to at least be able to make those tough calls, which nobody wants to do. Nobody wants to reject someone. I know it's difficult. It's probably the last thing on your list of 100 things, but I think we just have to.

MATT: Yeah, and I think there's the, ideally there's some feedback that's actionable for the candidate. But even when it's not necessarily actionable, because sometimes it's the, Hey, you narrowly missed out because someone has a very specific experience or something. But even letting them know that, hey, we think you're great. It just so happens there was one person who's got this perfect skill set for what we need. But you are great. Even just that, as opposed to nothing, helps. helps the person reconcile what's happened a bit. But ideally, actionable feedback where they can help to progress their own skill set is the ideal.

TIM: That is the ideal. And I personally hope that AI in automating, simplifying, and improving the hiring process, which I think is inevitable, will continue. Maybe start to unlock some of this because it won't be left in the hands of humans who are forgetful and reluctant to make difficult calls. Storing data in various systems is not even a way to transmit it. I feel like I could. Improve that a lot, which we will get to later in the conversation. One thing I want to ask you about, though, is when you think about hiring, what is hiring success to you? Because I feel like that's something we almost gloss over. Like what? How do you think about this for the roles you hire?

MATT: What does hiring success look like? What a great question. I view analytics as a highly creative role. So what I mean by that is sometimes you have roles where it's very process-driven. And an amazing candidate versus an okay candidate has the potential to be maybe twice as good, twice as productive in these very process ones. I view analytics as a times ten. The difference between someone's okay and someone's great is a times ten in impact. I've seen many analysts that, hey, they're doing some work; they're delivering what the stakeholders are asking for. It's alright. It's okay. And then those that are able to go, say, Hang on, understand what the business needs, where the impact will be, spend their time, come up with something really creative, and drastically change the business. It's such a creative role. Hiring success for me is about finding those brilliant people that understand the business, putting them in the right role, and then seeing that impact on the business for me. Really hard one to quantify. As such, but you do see it. I think we've all seen it from time to time, wow, because analytics, data, insight, or whatever label we want to put on it, is a relatively new field, to be honest, over the last 20 years. And because of that, huge headroom to change the business. We're about almost, I view it as we're about change. We're not about reporting. We're about saying, hey, we've got this wealth of knowledge that almost didn't used to exist. Thanks to the collection of data and cloud computing, the things that allow us to do scale. But we have a change role, then take that, push it into the business, and create some value. Yeah, I will always say I think the role of data people is not just about our technical skills; it's about our business skills; it's about our change management skills, because it's such a creative role. I don't know how many people out there talk about being a kind of data as creative, but for me, the potential for us to make change is absolutely huge.

TIM: Yeah. I can't think of many people to whom I've described data as a creative role. What makes you think it's because I understand that the kind of change agent aspect—what makes you say that it's a particularly creative role?

MATT: I think data analysis is creative because quite often it's looking into unknowns. It's looking for patterns that no one's seen before; it's looking for meaning that's never been seen before, and it's about those linkages of almost sometimes really abstract things. And at the same time, having to really leverage some deep skill sets, whether that is your analytics capability, your data engineering capability, or your data science capability (soon to be AI), very much, but there's a whole bunch of things you need to link up. It is really creative. And there's, in the right environment, we can really have analysts, data folk, thrive and make massive differences to business. So for me, yeah, data is a creative role.

TIM: I couldn't agree more when you describe the almost exponential difference between a great and a good candidate. And I think it's quite similar in software engineering as well that you have the kind of concept of the 10XR. developer, where it's, yeah, and I think you described it well, the difference between someone is going to pick up a ticket, code it, ship it, fine, but like someone is going to step back and think, Hang on, is this even the right solution in the first place? Not really. Is this even the right problem in the first place? Not really. If we did this instead, then we would go down this whole other path. And those people are lifesavers. I interviewed someone last week who described it. This is, oh, there's three levels of—they called them level three thinkers. So it's like the first person would notice a problem in the business and ignore it. A second-level person would notice a problem in the business and tell someone about it. The third-level person would notice a problem, fix it, and then tell people they fixed it. And when that was described to me, I was like, Oh my God, I would love to hire a company full of those people. How much better would my life be?

MATT: I completely agree. I tend to use a slightly different analogy, and I tell my analysts I want them to be data consultants. And we use this kind of waiter-versus-doctor thing of, the business comes to you, and if you're in the waiter mode, you'll say, What do you want? And they say, I want X, Y, and Z. And we go, Great, I will do X, Y, and Z for you. But it makes the assumption that the business actually understands data, understands what we can and can't do, and they're fully across it. And they're not fully across data because it's not their job to be. It's our job. So we need to be the doctor to say, almost, So you've come to me and asked for X, Y, and Z; I'm not going to say. That's nice, but actually what you need is A, B, and C, and I can tell you why you need them, and I can tell you the benefits of them and the drawbacks of what you're asking for, and I can tell you how this is going to make everything better for you. So I really want them to understand the business and work with them, but actually it's their job to say, Here's what we're going to do. I think that works at that kind of like level three we discussed, but yeah.

TIM: What do you think then of the idea that with the development of the progress of LLMs, maybe in six months time, the idea of writing SQL or Python from scratch would be just ludicrous, that you're going to be prompting an LLM? It's going to be doing the coding with, let's say, near-perfect accuracy, hypothetically. And then suddenly there's this kind of chunk of the technical skill set that an analyst or a data scientist would have had that's now redundant or taken away. Do you see the blend of the skill set of, like, a data professional changing to become, let's say, a little bit more consulting and a little bit less technical? Is that where you see it going? Or am I off the track?

MATT: No, I think you've definitely got something there. A few things that are fundamental. Every data person out there needs to lean in to AI, to LLMs. If you do not do that, you're going to find yourself left behind in the next kind of year to three years. And that probably goes for most areas of business. A few fundamentals. We saw Google last year; 25 percent of their code was written by AI. That's going to shoot up massively this year. You'd have Amazon saying AI wrote code equivalent of four and a half thousand man years. These things will only go up. Currently, humans, us, are playing the role of them checking that code. But more and more, AI is going to take over tasks we currently do. Now the question is, is it going to be substituting our effort or enhancing the things we do? Now hopefully, if we lean into it and we improve our skill sets, then it will just be complementing the things we already do already. The other day I used AI to help me out with a bit of complex SQL. And I thought, ah, I can really think about this. And it'll probably take me half an hour of playing around. But I was able to have a quick conversation with an AI, and within a minute it had some code. I ran a bit of it, went, Oh, that's not quite right; you've missed this. Ran it again, and in five minutes, instead of probably half an hour to an hour, I had the answer I needed. It's there to complement the work I'm doing, to speed up the work I'm doing. It's also not just technical when it comes to AI. There are so many day-to-day aspects of it. Even if we were to go back to interview feedback as a hiring manager, actually that feedback, I can use AI to just quickly dictate and transcribe my feedback. The reason managers don't do it is because it takes a long time. But if it's a minute of work and it's really easy, we can make things better. Managing my communications plans or writing emails can go from big, onerous tasks to something really simple. There's a technical aspect for all Datafolk that they need to lean into and understand. There's just everything else we do that could get better. But the candidates that don't lean into this are going to have problems getting hired because they're not going to be as productive as their peers. So we all just need to do it, and we need to do it quick.

TIM: Is there some sense, especially maybe for candidates or people who've been doing their profession for a long time, they're like a master of their trade? And I guess, particularly in the West, we would have a tendency to utter the phrase, I am an X, and X is my job. And so we have this kind of, our job is tied up with our identity so closely. And there are some roles where the task is the name of the job. Like, I am a coder. If you're a coder and you say you're a coder, and then in six months time, the concept of coding is no longer done by a human. Then I feel like people are going to have to go through a very quick mental shift. To go, I'm no longer a coder. I'm a problem solver, or I build products, not code. Is that going to be a challenge? Do you think for some people?

MATT: I think sadly yes. There are some roles where there'll be a lot of consolidation. So where businesses had teams doing some stuff, it will drop to a few people. There will be roles possibly that even completely disappear. We're entering what is arguably one of the biggest; some people say it's up there with the Industrial Revolution. Who knows? But let's see in five, ten years time where this takes us. I think it's going to take us somewhere very different from where we are today, and whilst I have no idea where that is, to go back to your original question, people that are currently saying, I do X and X is my role, I guarantee you role X is going to be very different in 10 years time; you're going to have to embrace it, or I think you may fall foul of roles no longer existing.

TIM: And at least in the short run, as you say, AI is at least for many things going to make us drastically more productive. So if I'm like a software engineer and it used to take me a thousand hours to. Do one thing, and now I can do that in two hours. Surely the value that I'm delivering is drastically higher. That's to my massive advantage.

MATT: Yeah, completely. And I think also then you'll start to look and say, like, how do I measure outcome? And it's probably the value that I can deliver, so it may be, Hey, I'm judged less on my technical coding capabilities and more on the outcome that I'm able to deliver, shaping AI, but understanding the business as well to get the outcome that everybody needs. So yeah, it might be that we are lifted a bit more out of the detail on some things. But I think we'll always be, where the humans exist, should probably be that business context. And the individuals, I hate to say politics, so let's call it the social interactions of the workplace. But AI is not going to take that away. You're going to need someone still to understand those personal interactions and how they work around the business. And that's always going to be there.

TIM: And so will that end up being like again? Trying to see what we're getting to, will that end up being a relative? A larger proportion of someone's job is, at the end of the day, just dealing with the people side of things rather than the technical side, whether it's the coding, the dashboard building, the analytics, the modeling, or what people used to think of as the data job; now it's no longer that.

MATT: This is where we start to get really matter, right? So in the framework of what we currently know and understand, yes, you could imagine a lot of that technical work shrinking, and therefore the role is this kind of very social side. However, this is the bit where we get to rethink stuff into a new future because I think there's going to be a whole world of new things that open up, some of which we don't even recognize right now. Yeah, there's a big, bold future. I'm almost saying I don't know. I don't know. I just know it's going to be a very different place.

TIM: Yeah, there should be so many new roles created. And the way I think about it, at least from an analytics perspective, is I don't know if we had some way to say what are all the decisions we made in the last year in this massive business like Samsung that were, I don't know. A million decisions were made, from tiny ones to big ones. How many of them were perfectly accurate? How many of them were data-driven? How many of them did we even have any data for at all? I would have thought probably a tiny minority because there's so many; we make so many decisions each day on a whim, quickly. There's so many things we're not even measuring at all, even in really data-driven businesses. So there's just so much upside for making, I think, better decisions. If we collect the right data, how to system to make the decisions, et cetera, et cetera. So we're not even touching the sides at the moment. I don't think

MATT: And here's the other thing about AI. It requires quality data, right? You have to feed data into it, and that data needs to be of quality. So there are so many kinds of governance and oversight and engineering roles that will still have to exist, and certainly from an LLM perspective or whatever the next thing is on our path to artificial general intelligence, that there's going to be massive industries there for data folk.

TIM: One domain is definitely hiring itself. Because I don't know about you, but I feel like hiring has been done the same way for a long time: it's a job ad on a massive job portal application form CV screen with a human bunch of interviews. Reference check offer that is This is the same as when I started my career, as it is now, really, but I think there could be massive disruption in rethinking that and not necessarily going step by step. And we used to do this with a human. Now we're going to do it with AI. I feel like we need to almost start from the ground up. I think, what are we even trying to achieve? What's the end goal, and build some kind of new system from the ground? What do you reckon?

MATT: Yeah, I think so. I think so. I certainly think that the process of CVs itself is going to be over at some point. You can, we've, there's been studies out there with just two hours of interaction with an LLM, of answering various questions; you can get to the point of creating a clone of yourself. That, with roughly about 80 percent accuracy, could answer questions on your behalf, like survey-type questions. And then if you take this perspective of, okay, so we're going to create AI agents that can effectively mimic, to a 100 percent accuracy, but to enough, things about us, I can absolutely imagine almost having a screening call with AI; that's absolutely a thing that could happen. Like, initial screening is a conversation with AI, and it's drilling down to certain bits of your capability against the role. And that could almost, to a degree, replace that first CV screening phase because you can scale it almost infinitely. There's so much I think will change as part of that hiring process. What are your key takeaways from AI in hiring?

TIM: Just think there's so much upside, and we could get rid of all of it. What I assume will happen is it'll be like in stages. Where, so there's, I don't know, the current HR tech that's used that's 10 to 15 years old normally, kind of an ATS layer. I feel like there's going to be like a new wave of that where it's AI native rather than a system that's built for humans to click buttons on a website that's almost like a sequel wrapper. Now it's going to be some kind of more AI-based thing. I hope the CV is going to be gone, but I feel like the first generation is going to be, oh, an AI CV screener and then an AI interviewer, an AI reference checker, like just a replication of what we're currently doing. But then maybe after that, there's going to be something more holistic that's thinking from the get-go. One issue I feel like. That's plagued hiring for so long as a lack of data, which is what you've just described, would solve that problem because if you had these AI agents that could then generate the data on the fly based on what they already knew about the person or the company, that would be really clever because that would get rid of the lack of a data issue that we currently have because it's, oh, here's someone's CV. Great. What does that tell me about them? Like seriously, it's a nothing document. We

MATT: Exactly. Yeah, completely. And the fact that, arguably, AI can be really flexible in the questions they ask, like starting off with a core, sensing a strength or a weakness, and diving into it. It's just how fast this technology has moved is almost mind-blowing, to a degree. Of thinking about where we were five years ago to where we are now. It's just amazing.

TIM: on that One thing that's come up repeatedly in my conversations in the last couple of months has been thinking about, you know, What do we even select or hire for now if the skill sets are changing so quickly? Maybe some skill sets are going to be redundant soon; technologies are evolving so fast you can't necessarily just hire for tool XYZ because those tools could be replaced in a year or two years or something. A lot of people have been talking then about hiring for adaptability, hiring for an ability or willingness to learn. Is that something that you would value? And if so, in general, how do you think about hiring in a time where just things are changing so quickly?

MATT: Yeah, exactly. If a big thing that's changed for me has been if you asked me two, three years ago, what does the modern analyst need? And I'd say, you need a data science toolkit. And that would have been my answer every time. I don't say that anymore because I think almost that data science toolkit is going to be replaced and irrelevant in a few years time. Because, as you mentioned, an LLM can probably get you there pretty quickly. What we look for is a few things. Number one, I'm generally looking for client skills. And what I mean by that is, can you talk to clients? Can you talk to stakeholders? So that's one of those key things. And then do you have the technical skills to back it up? And quite often with that, it's a Venn diagram of a unicorn in the middle. And I'm super conscious of that. It's really hard to hire people that can do both things in the data field. I wish it wasn't, but it's a real challenge. I can either find people with client skills, but actually you find that they're really not that strong at the analytics. They don't have the technical background. Or really strong technically, but can't do the client skills. And the problem is that 10 times 10 of creativity, that for me is those two things. Now, on the technical side, I want you to have a technical discipline. I want you to show that you can be correct and thorough. But then there's the, Are you willing to learn? like you said. Because whatever technical skills you've got, that's going to be refreshed time and time again over the next kind of three to five years. And if you can have that willingness to learn, and you've all got those soft skills to really then drive home whatever you do and push the change into the business. That's what I look for.

TIM: I think because, yeah, if you have the, let's say, if you have the technical genius who can't communicate the value of their work, can't convince anyone to change anything, then they're just producing work that. can't deliver the value that it could, then I guess you've got the soft skills communicator person who can't actually execute what they're talking about. So they would need help, but you get that a crossover is the deadly combo.

MATT: Yeah. And also, on the bad side, I've seen people with the soft skills that are incredibly influential and able to convince people, except their technical skills were so bad they were convincing people of the wrong answer. And for me, I think that's the worst one of all. I'd rather someone be completely ineffective than effective and pushing wrong things into the business.

TIM: What about the creativity aspect you were talking about before? Is that going to somehow become more important or less important? Do you see any change in the pattern there?

MATT: I think the principle of it will always remain. So the creativity will be, hey, we've got a whole new bunch of different skills, different resources out there, of LLMs and everything else to come. It might be that it goes from being times 10 to times 100. I don't know, but I think that the creative aspect of piecing it all together, understanding what's possible, and then melding that with business needs is for anyone out there looking to understand how they develop themselves. That's where you need to get to. It's about helping businesses understand how you can leverage these new resources and the resources they have. to really help drive the business forward. That's the challenge. And I think, yeah, as you say, that creativity aspect, never going away, may even become more impactful.

TIM: I think you're right. And people have been talking about them. Let's say the concept of the idea of a ten- or one-person AI-driven company that's worth a billion dollars Maybe it's possible now, and it's only really through that kind of leverage tool of AI that if you know the AI and you can program and you understand the business domain and you can sell, then maybe you could become a, yeah, a unicorn company just as yourself one day, and that, yeah, certainly requires the creative element as well. It will be fascinating to see what happens in the next few years with this kind of new wave of AI-first companies coming up. I can't

MATT: Yeah.

TIM: them.

MATT: Yeah. But, once again, it's for everyone in the data business. Every single one of us, no matter what position, needs to be thinking about how we're going to upskill ourselves with AI. And I don't mean that in terms of how am I going to build amazing new products. Actually, just even our day-to-day. How are you going to automate your day to make you more efficient? The people that do it are going to do well. The people that don't do it are going to fall behind really quickly.

TIM: One thing I had heard recently from a couple of companies doing one large analytics consultancy, another large tech company, was they said to their entire workforce, this is like a thousand to 5,000 people. Stop your day-to-day jobs, stop working, and just spend the next, what was it, like a couple of weeks implementing AI, Chachapati, Claude, or whatever it was, and just think about what everything is that you're doing all day, every day, and just see which ones you could automate away, which normally I think is the wrong way to think about it, because it's like a solution looking for a problem, but maybe in this, profound change. You need to carve out that time to really sit down and go, Wow. Like, this tool is changing quickly itself every day. B is such a profound change. It's worth it. It's not like it's an incremental improvement to something that might help. If only I could get rid of this entire thing I used to do. I wonder if people need to, yeah, carve out a bit of time just to sit there and think about how they could use this amazing technology.

MATT: And I don't think everybody needs to do everything immediately. We've, with my teams, said, We, you have to do this; you have to go away and learn about what the business implications are for AI. You want to go think about the use cases to understand where the industry is going. And then I want you to think about your day-to-day, and I want you to just pick one or two applications of it. And the point being, I think there's a muscle memory thing of you've just got to start; you've just got to start doing it and realizing how useful it is. And then very quickly, when you get in the habit, you're going to come up with idea after idea. And the other thing is, this is still really new technology. Even if you go to ChatGPT, it was only in December that they had a really good conversational AI. Prior to that, we were all like typing these prompts that were really a bit annoying and difficult. Now you can just have a chat. Where's this technology going to be in two years time? None of us can predict. I think it's the, You don't have to solve everything; just do something now.

TIM: That, that's a good one, actually, especially for people who tend to get overwhelmed easily. I certainly get a sort of fear of missing out kind of vibe a lot. You go on LinkedIn; it's like AI this, AI that, and you start to think, oh, yeah, this is all a bit too much. But yeah, it's good just to. Boil it down to one or two things. One thing I've used Chachipiti for recently was two things, actually. One was some of these plants in the background here. I was repotting one, and I noticed something funky-looking in the soil. And I did the video check, and it told me it was a mold and that I had to do a few things to fix it. So that was great. And another one was, as a language tutor, I have some post-it notes around my apartment with words from Russian into English. And I just got it to give me a quick test. I had a quick language test with it. Had a conversation. It was speaking in Russian to me. This is for free. This is pretty amazing stuff.

MATT: Completely. I think just as a learning resource, right? We've never had anything like this. You've got these models like DBT and Claude that have been trained on almost the entire wealth of information of the human race, right? This stuff has never been achievable before. And so quickly you can shape it to say, Hey, I want to learn about X. And you're going to get a great set of responses. You can tailor it to say, Okay, dive into this thing; I need to know more, or make it broader, make it deeper. But just using it as a learning resource, it's amazing.

TIM: Have you had any recent experiences that almost switched a lightbulb on in your head to go almost like, really pro AI?

MATT: Yeah, for me, it's the OpenAI when they released the kind of latest piece of conversational AI. I always found it really hard writing prompts. And then getting these kinds of answers back, and then you'd wait for it all and go type this thing. When we could change that to, it's now a conversation. And I can talk as a human; I can slightly ramble; I can add pieces in as I go. Somehow, it pieces that all together into something meaningful. It understands and gives you an accurate output of what you wanted. And then, or even if it doesn't, you can really quickly correct it and then get to that. That was for me that kind of human-level interaction, rather than writing what felt almost like code. I could just talk. And at that point, that was the light bulb moment for me of, Oh wow, this is going to change everything.

TIM: Yep. I've also liked how they've done a pretty good job at really humanizing, if that's a word, humanifying; it's probably the one with the voice, and it's a real kind of level of empathy in it. Like again, when I was doing my little Russian lesson, it was encouraging me and it was like politely correcting my mistakes, not being too harsh; it had a real nice balance to it. What, one thing I've also found great. is that in general, asking for feedback on things, I find like I'm very receptive to hearing feedback from AI, probably more so than I am from some humans. judgment. been a nice use case.

MATT: That's great; that is great. Absolutely. I think the great challenge we've got. There are so many use cases for AI. It's the kind of thing that's unknown right now, but we'll figure it out as part of the journey. I just say to everyone, when was the best time for you to start learning AI? Probably a year or two ago. When's the next best time? Today. Just start with one small thing, just do something, and then keep that muscle memory going. And you'll be in a drastically different place in a year or two's time.

TIM: Have you ever read the book Atomic Habits?

MATT: Yes. Yeah.

TIM: It's a great book. And yeah, that's where the author certainly recommends he's trying to get a new habit going or get rid of an old habit. You just start with something small. Don't overwhelm yourself. You don't have to go on a marathon. Just go for a hundred-meter job. That's better than nothing. And it's the AI, isn't it?

MATT: I think it's the same with AI. But I think the real reason it works better than most things is you get this benefit back straight away. It's not a case of, Oh, yeah, I went for that run; I felt pretty bad; maybe I'll go tomorrow, but I'm aching with AI. It seems to be just reward.

TIM: Yeah. Yeah.

MATT: reward.

TIM: Yeah. That's a good point. And it's just, yeah, so readily available, pretty much free. Why not give it a crack? And yeah, the options are almost endless. And I think it's going to have a big improvement on hiring. I think it's got a. A big ability to make things a lot fairer personally because the way hiring is done at the moment is so subjective. I feel like it could really help to make it more objective. I don't know if you've seen any of those studies around CVS, where the only thing they change is the name on the CV. And it basically reveals huge racism and discrimination that I feel like, with the right trained AI, we could eliminate that in theory.

MATT: I've got to agree. There's so much concern about AI bias, but hey, the practices we currently have are probably extremely biased, and the research shows they are. Let's test it. Hopefully we find something much better.

TIM: Yeah, exactly. If the starting point was panacea and bliss and some kind of utopia, then I'd be fine. I'd be worried about AI, but in the context of hiring, I think it's, yeah, a market that hasn't really changed in decades that is pretty flawed, just to share one specific metric. In Australia, if you were to apply to a job with a Chinese first and last name, your odds of a callback are one third less than if you applied with an Anglo-Saxon first and last name. That is just a rot. That's fundamentally unfair. And this is like comprehensive studies done by the University of Sydney. So if we can chip away at that problem, surely that's a great use case of AI.

MATT: Yeah, humans, human bias artificially limiting the talent pool. That's good for no business. We want to get the best talent we possibly can. And yeah. Let's break down those biases wherever we can.

TIM: Matt. It's been a great conversation today. I've thoroughly enjoyed it. We've covered a lot of ground. We've certainly hyped up AI a lot. I hope I could deliver in the next few years because what if I Yeah, thank you so much for joining us and being such a great guest and sharing all your insights and thoughts with us today.

MATT: Thank you, Tim. Yeah, it's been an honor.