Why Your First Hire Isn’t an Engineer Anymore — Michael Batko, ex-Startmate CEO

The first hire used to be an engineer. Michael Batko thinks that’s now backwards — and that’s just the start of how AI is rewiring the way you build a tech team.

In the first episode of Building Tech Teams, James MacDonald sits down with Michael Batko — who spent eight years as CEO of Startmate backing 240+ startups, and now runs the AI-native venture batko.ai. They get into what actually changes when the cost of building drops to near zero: why go-to-market now comes before engineering, why the 10x engineer is finally real, what “AI-native” means for a person and for a company, and how to lead a team through it without torching your culture.

It’s a practitioner-to-practitioner conversation with real opinions and real scar tissue — hiring for attitude over skill, the rise of the AI chief of staff, the squeeze on middle management, and why your hard-won experience is the one moat AI can’t copy.

About the guest

Michael Batko spent eight years as CEO of Startmate, where he backed 240+ early-stage companies, and now builds AI-native tools for founders and businesses at batko.ai.

Building Tech Teams is produced by DayOne.

Listen to the episode

Transcript

Auto-transcribed with AI.

Michael Batko [0:00]: Fire 5 to 10% of your team every year, otherwise you’re not being ambitious enough, which is actually really hard to do because nobody wants to fire people ever.

James MacDonald [0:06]: I love that take. It’s a controversial take. It’s brutal.

Michael Batko [0:09]: Yeah, it’s hard. 10 desktops open and every single one of them is a Claude Cowork conversation working on 10 different problems at any given time. That is kind of like my current status quo of how I work.

James MacDonald [0:22]: There’s a different article every week about legal profession won’t be there in the future, the accounting profession, the recruitment, any white-collar roles.

Michael Batko [0:28]: And honestly, I actually believe in all of that as well. It is scary and it’s like, if all those types of jobs will be fundamentally shifted and changed, if not non-existent in the future. The only positive I’ve got for somebody who is listening to this right now is like, at least you are listening to this right now and you are literally thinking about that already.

James MacDonald [0:46]: Hey, welcome to the first episode of the Building Tech Teams podcast where we’re talking to technology leaders across Australia about how AI is affecting technology teams at the moment, the different changing nature of roles, team sizes, team dynamics, how AI is really affecting this going forward. On today’s episode, we have Michael Batko, former CEO of Startmate, currently AI CEO of batko.ai. I’m so excited to have this episode because Michael brings a unique perspective, having seen hundreds, if not thousands, of startups and scale-ups. He’s seen what’s worked and what hasn’t when companies are building their teams. He’s also now working with businesses up to enterprise level on how they’re changing the nature of their teams, how they’re implementing AI, and what the flow-on effect of that is. So really keen to get his insights in how AI is affecting technology teams now and into the future. Welcome, Michael.

Michael Batko [1:39]: Thanks so much.

James MacDonald [1:40]: The reason I was really keen to have you on as our first guest is there’s probably not too many people in Australia that have seen as many early-stage startups and those key first hires, where companies go right, where companies go wrong, So, really keen to get your initial thoughts on what is that first key hire look like if you’re a startup founder? How do you go about building your team?

Michael Batko [2:00]: I mean, yeah, for context, for 8 years I was running Startmate where we invested in like 240 companies essentially super early days and then saw them scale to hundreds of people. But yeah, the first hire, like I said, so, so crucial. And like it used to be the most common question you get is somebody has an idea and the first hire is like an engineer because they need to build it and so on. And that was really the case for the last like 10, 20 years kind of thing. Whereas really now with AI, and we can dig into this much, much deeper, is getting an MVP and a prototype off the ground is just so easy. The cost of execution, the cost of building is just so low these days that actually the first hire, because the cost of building came down so low, it’s actually go-to-market and customer acquisition and eyeballs, which is now expensive or like the really hard thing to do. So I actually think my take on this one would be the first few hires now have completely flipped into like you actually want to prove that there’s even customers there and actually hire go-to-market, whether it’s sales, marketing, business operations to get that off the ground. And obviously all of them AI native just to, and we can talk about this part as well, but actually you probably don’t need an engineer until a bit further down the line. So it might be a little bit controversial here.

James MacDonald [3:09]: No, I like this and I do agree on the flip side of that. To the software engineering community who sometimes get a bit butthurt sometimes, let’s say that, on what is their role in the future. Are they still important? I definitely believe they are. There is a lot of startup chatter around, hey, I can throw something together on Lovable and I’ve got a business. I’m now a software engineer, loosely. At what point do you then need to hire some proper software engineering to make sure we’ve got a scalable platform, to make sure it’s secure, architects, built right? Where do you see that coming in? Because I do genuinely believe the role of a software engineer, senior software engineering is still vitally important.

Michael Batko [3:49]: 100%, no, no, for sure. And I think that there’s exactly that misconception of like, oh, I just can vibe code my way to like actually making a billion dollar company or whatnot. And that’s definitely not true. That being said, that kind of zero to one is possible without engineers. But then once it goes above the one where you’ve proven you’re onto something, your product has expanded. And actually I’ll give you a real life example, which is Batcode.ai, which is the company I’ve started about like 6 weeks ago. I started Vibe Co. with another Claude. I’m still just by myself just doing it. I haven’t touched a single line of code. And what it is, is actually solving the problems of founders from ideation all the way to exit. Anything from like you want your pitch deck assessed to introductions to VCs to writing a great press release. Like I’m boiling the ocean, like 20 or 30 kind of problems that founders have along the whole journey. But as I’ve started building it and I’ve been building it now for 6 weeks, the products that I’m building are actually production ready. They’re awesome. They’re getting great feedback. But what I’ve only just realized like 3 or 4 days ago is that I literally have started running out of capacity. And that’s not even my own capacity, but actually what I can deploy as a website. Literally, it’s telling me that Vercel, which is what actually puts the website up live, is running into build limits where I literally can’t deploy it up anymore because it has become a beast of 2,000 pages, all SEO optimized with way too many API calls and stuff. And that’s the stage at which I’m going like, all right, I think I need to upgrade myself on like the more dev skill types of things. And then in the future, actually, you do need that dev to essentially be the orchestrator of like, all right, what does everybody build? Let’s bring it together. Let’s put an architecture around it. Because with AI, you just bolt on things, whereas like somebody to actually think about the holistic picture, that’s what you can do with AI if you have that insight. But most people don’t because they’re not engineers.

James MacDonald [5:34]: Yeah, correct. And I do think that the 10x engineer that everyone was hiring for 5 years ago, the rockstar developer, that 10x engineer now does exist because if you have you look at, you and I become a little bit dangerous with Claude Code or other tooling out there, but somebody that actually understands software engineering with the new tools on top, that’s where those people are at an even bigger advantage, especially if those people are also ideas people or have some form of commercial now so they can come up with an idea and build something.

Michael Batko [6:03]: 100%. Like, I totally agree. Like, um, it’s almost like every person now, like, to your 10x engineers, every person has like now received a 10x because they can now optimize their own role and build and get things live. But now the engineer actually becomes even more important because like those 10Xs across all those people you’ve hired or you’ve got in your team, they’re all going to come crashing and burning when you try to put it all together. But now the engineer becomes even more important to actually bring it together under one hub, make sure it works, cybersecurity, et cetera, et cetera. And like without that, like you don’t, like you can’t actually scale beyond a certain point.

James MacDonald [6:35]: Yeah, I completely agree because there is a lot of talk when talking in startups at the moment, most people go from zero to one or ideation to let’s build a proof of concept MVP, test if I can get a customer. But you’re still looking at enterprise-level software companies still have teams of hundreds of engineers. Tons of them. They’re there for a reason, right? They’re not there because, hey, Canva have employed all these people and they’re stuck with them. They’re just producing more and more output, right?

Michael Batko [7:02]: Well, it’s going to be an interesting one. Another opportunity which lots of people have been hitting me up with, but I personally don’t think I’m the right person to solve it, is all of those larger startups, scale-ups, but also corporates. And the question they keep asking me is like, how do I actually get AI into our production without breaking the whole thing? Because you have got hundreds of people, you’ve got thousands of customers, you can’t just throw it in there and just hope for the best. I can do it right now because I’m by myself doing it, having a bit of fun. But they’re literally asking for like round tables to compare nodes and how to actually do that. And that’s actually quite an interesting, difficult challenge. But engineers become super crucial there to actually make it work because you can’t just throw something into production and hope for the best.

James MacDonald [7:41]: Yeah. Yeah, I think the big challenge from what I’ve seen, what I’ve read is that Building AI native from the ground up is significantly easier because you’re building AI native from the start. Trying to make a company AI native that was started and has been built traditionally is a significantly harder problem.

Michael Batko [7:57]: Like, it’s culturally like an interesting one to think about as well. And, and the other one is actually from like, uh, who is the orchestrator of it all to be responsible to bring it together? That’s— I feel like that in itself is going to become a massive role of just like the person responsible for bringing AI tools together in the company. I feel like that in itself will become a role in the future and engineers are incredibly well placed to do that.

James MacDonald [8:21]: Yeah, and it’s also part change management, right? It’s not just a software engineering skill, it’s not just change management skills, somebody that’s got change management background but also reasonably technical.

Michael Batko [8:31]: Yeah, it’s actually an amazing role. Like it’s like the technical side of things but also like the change they can have in an organization, the impact, plus the people skills required. It’s actually a bit of a unicorn role, which I genuinely feel like it’s going to be one of the coolest and the most fun roles.

James MacDonald [8:45]: Also one of the most important. That conversation has come up at least half a dozen times in my past 2 weeks.

Michael Batko [8:50]: Oh, I can imagine. Yeah, that’s going to be a— I’m so excited for that one because it’s almost like a CTO, but actually like CAIO, I guess. Yeah. Yeah, the impact that it can have is actually so cool.

James MacDonald [9:01]: I agree. Well, at the start you mentioned AI native. So whether it be the software engineers or go-to-market, sales, everyone else, how do you describe AI native if somebody’s not AI native at the moment? Let’s say your past 6 weeks journey has been quite significant, right? Yeah. Is AI native just a mentality? Is it a learning? How does somebody go from where they might be at at the moment becoming AI native?

Michael Batko [9:25]: It’s probably like the individual and there’s the company itself. Like the individual, the journey usually is something along the lines of I use ChatGPT, and I get into ChatGPT or Claude. Then the next one is kind of one of like, oh, I’ve created myself a skill or workflow or something like that in it. And then you kind of start seeing the magic of something keep looping through. The next journey then, which is actually difficult to take, it’s not actually difficult, but like it does require that kind of like almost like inspiration from someone else, is creating yourself almost like an AI brain. And that’s almost like the first layer of being AI native. For yourself as an individual. What I mean by that is distilling all of the information about yourself in one spot, whatever the one spot is. It can be a GitHub repository, it can be a Notion page, it can be a massive MD file. But like scrape the entire internet for your name, throw in all of your insights, meeting notes, emails and stuff. Suddenly there’s literally like a digital replica of you. And now once you’ve got that, that’s actually where the magic starts to happen because now you can start plug in all those workflows and all the cool things that AI can do and literally have like the digital replica of yours do that incredibly well. That’s kind of like individually, like how you can really unlock yourself. And now the next step then is actually the hard one for organizations to do, which is how do you scale that across individuals? Like as an organization, that’s usually where you want to bring somebody external in. We do that actually for companies as well. That actually again, like requires that kind of orchestrator that we were just talking about, somebody to actually navigate of like, let’s create that company brain. But then cybersecurity becomes really important. Ring-fencing information becomes really important. Who’s got access to what? But once you’ve got that, that’s really your first layer to making a company AI native. To give you a real-life example, I just started the company last week and we’re literally from day one, we’re just like, we’re going to make this AI native. And the instinct is replicate everything the way we did it before. You’ve got a Slack channel, you’ve got different things, you’ve got one-on-ones, you’ve got your Notion templates to do around your one-on-ones, whatever it is. Instead, the first thing we literally did was throw an agent into our Slack channel, like using old tools like Slack, but actually throw in an agent. And the jobs we get, we throw in one agent, which is literally our chief of staff, who then navigates every task we tag it in into like the 7 or 8 other agents that we have. One of them is literally like scraping Slack and writing our company history as we go, like from day one. The other one is kind of like seeing any errors and fixing them up straight away. The third one is literally like writing out lessons learned and stuff to navigate as a team and so on. And that’s kind of like what it means to me of like being AI native from day one. We’re just literally like 10x-ing ourselves to actually just like have agents literally like work for us.

James MacDonald [12:06]: Yeah, and I think that the hierarchy of what used to be the hierarchy of people, you’re seeing that build out in the hierarchy of agents where you do have agents managing agents and agents handing off tasks to different agents. And then you’ve got QA coming back, sitting over the top of the engineering agents and making sure you’re QA-ing it as well. So it is that same philosophy. And I think what’s really interesting, you see a lot of young people out there at the moment that might not have worked in a corporate environment saying, “Oh, I can rebuild your entire marketing team with one agent.” Where it’s not actually quite true if you don’t have the context on how different people come together in teams, how different processes work. I think having that business context or the commercial context, engineering context, and then being able to use the right tools puts you in a significantly better position than just trying to replicate something you don’t quite understand.

Michael Batko [12:55]: 100%, yeah. So there’s two parts there. Like the first one is kind of like the agent structure itself. The other one is kind of like the business understanding. The agent structure is quite interesting because to be honest, that’s actually how I started. Put in one agent which did like 100 things and that’s actually like how you get started and I made all the mistakes. But throwing in one agent doing 100 things keeps breaking all the time and it’s so bloody annoying. So it’s almost just like have one agent do one thing and do it incredibly well because that’s how you do it. The other one is actually the whole agent management, which is quite interesting again. Because then I had 5 agents plugged into my inbox scraping my inbox for separate things. And then they also kept breaking or kept giving different context. So then it’s again like restructuring your team. It’s like, “Right, kill those agents, make one agent actually catalog my inbox and then delegate to all the other agents.” And actually that smart engineering, that smart kind of architecture, you either learn as you go or you have somebody with, and this is the second part, all the business context to actually make good decisions from the get-go. And actually the thinking piece, the strategic piece is like, because execution is so cheap and building is so cheap, is like the thinking becomes actually game-changing when so few people do it. And including myself, because in day one you’re just excited to start building.

James MacDonald [14:09]: Yeah, correct. I think the real challenge comes at implementing AI into enterprise or into bigger companies as well, because as you said, individuals can use it. Individuals can— my, my call game changed as soon as I built the MD file and all about myself and actually had that context to point back to for everything. Doing it at an individual level can work and is easier to do. Replicating that across the business, implementing AI into traditional businesses that do have more important guardrails, data privacy issues, significantly harder challenge for a company at the moment that’s looking at bigger organization. I know I need to do something with AI, that’s all I hear about at the moment. How do you— first 2, 3 points.

Michael Batko [14:50]: I mean, yeah, there’s a couple of things there. Like the first one is kind of the one of like incentivizing the team and education. There’s another kind of like blocker, which is often busy work. And there’s like essentially bringing in the external help, which is probably like the 3 core points there. The first one, which I recommend to any business, startup all the way to enterprises, it’s just incentivizing the team to just play around. And playing around, I genuinely mean playing. Like so often it is just around getting over that friction and the hurdle to just get started.

James MacDonald [15:19]: Real quick one on that enterprise organization play around is that everyone, you want to download ChatGPT, I’m going to download Claude and we’re going to upload client information. Yeah, yeah, because I think that’s a real— I, I think we’re gonna hear some wild cases of client IP, company IP, data breaches being uploaded into open source LLMs. It’s coming. It’s wild how much is that happening.

Michael Batko [15:42]: Yeah, really good point actually, because I’m— so actually back to the previous topic that we just talked about, the layers of like what an organization needs, like that organizational brain that I talked about is really important, but it only works if you’ve got the AI policy right next to it. And I 100% agree, you actually need that. Otherwise the organization is just scrambling. Nobody knows what they can or can’t do. And you really want to nail that down and be really specific with stuff, what is allowed or not. Otherwise you can run into a lot of trouble there. But what I actually meant with playing around, I literally mean it. There’s team meeting, everybody gets a prompt and throw it into Gemini or whatever it is and just come up with the funniest image. Stuff like that. That’s what I mean with playing around. Because you just want people comfortable with actually starting to use it. The problem, and this is actually the problem for any organization, including myself at Startmate, which was innovative. We were working with all those startups, I was literally going on the AI roadshows and learning from founders and bringing them together and giving talks about AI. Even for me, actually to use AI was really hard just because your day job is so busy. All of us have crazy to-dos, just like work way too late and so on. And to take that extra hour or whatnot to actually play around with AI, it always feels like you’re behind because you are. But then also like, you know you should do it, but you don’t have the time. And that’s actually the way you can break out of that is essentially incentivizing your staff then with like using those types of tools. So I have seen organizations literally put up prize money out there, actually pretty serious money as well, which is allocated by the senior team and so on. On the other side, actually bringing in external parties as well. As I mentioned though, we do that for companies where we literally come in and we interview your team and like, tell us about all your problems and issues, and then we build it for you.

James MacDonald [17:25]: Yeah, this is a really interesting point. It feels like everyone wants to get some quick wins with AI, and I feel like if you break down some of those little quick wins with AI to start with, you can actually build some momentum within an organization. A couple of those quick wins normally take some mapping because I think the easiest use cases for AI in any size organization— what is the current process that you’re doing? Let’s map that out properly. Where can we apply AI to make it more efficient, right? But it does require that mapping part to start with. So having to do the work first on top of what you’re already doing is going back and mapping what you’re already doing if you don’t have all those SOPs already in place. So it’s like one step back to take two steps forward.

Michael Batko [18:04]: Totally. I mean, that one’s interesting, right? Because like when you say it like that, it feels like a huge task. Yeah, I need to take all my workload now, I need to map it out. I’m just gonna be like, in a traditional context, it would take me days because I’m like, this is so annoying. Everybody hates kind of like documentation, right? And maintaining documentation even worse. But, um, the interesting one there is once you’ve done it a couple of times and now with AI, actually process mapping becomes really easy. And the absolute upgrade there, which I just want everybody to get onto is dictation or a tool called Whisperflow that I personally love.

James MacDonald [18:38]: Oh, it’s changed my game.

Michael Batko [18:38]: It’s absolutely game-changing, right? Like literally on my first day using AI, I walked into an office and the guy and his colleagues started talking to the laptops and there was just me like typing away being like, I am so slow. And they were just like talking to AI and If you write an email, you couldn’t just do that and talk to it and write a great email. You actually do need to type. But if you give AI instructions, the more context you can give it, the better. So that’s kind of like the first thing to like process mapping, which is like suddenly it becomes so easy because all you have to do is talk about your job. The other thing you can do is just throw in lots of examples of what good looks like as well as templates. And AI does a pretty good job at like actually mapping that out. And then it’s almost just about you asking good questions to Claude, being like, ask me great questions about this process and just keep talking to it. Honestly, within like 10 to 30 minutes, you’d map out a pretty solid process for pretty much anything out there. And then optimization on like the actual building off the back of it is also so fast. The first time is scary. The first time is scary, but it’s actually so much around like asking good questions for Claude to ask you questions back.

James MacDonald [19:42]: I agree. I do think the idea of either having an AI champion internally or bringing in an external partner is a really smart idea for organizations to actually think about how do we start getting some quick wins, because trying to turn a traditional organization to AI-native organization is very difficult, and it is a bunch of stepping stones. Having the right things in place, but actually just getting some wins along the way and then doing more and more as you go, rather than trying to overturn your entire organization, become— we’re AI-native.

Michael Batko [20:12]: Totally. Yeah, it’s, it’s just such a scary process, right? Like, because everybody feels behind, so they feel like they need to now throw AI into their main product, but it’s actually not around throwing it into a end product. It’s literally to your point of like the quick wins. It’s even stuff like, hey, help me prepare before the meeting. Like, hey, help me write my emails better. All of those types of things give you the confidence that, hey, this is actually saving me time, this is working. And as you go through those, like you just get more and more comfortable, confident, and you actually start AI enabling your business. And then you actually internally want to build the AI tools to actually run your business itself. But that actually comes way later down the line. It’s actually almost like the empowerment around everything else around it, which comes first.

James MacDonald [20:54]: So we go back to the building teams part. Touch one more on the startup side because I think you would’ve seen as many startups start and fail, start and succeed and scale. How do we go about still in this age of AI, how do you go about building those? First, you’re a founder, technical co-founder, you got some product market fit. I wanna make my first 5 hires, first 10 hires. What does that look like if you are looking at all of your experience in successful scale-ups How do I go about that?

Michael Batko [21:21]: Yeah, I love that question. I mean, so this is back to the point earlier. Traditionally it was always just like hire a couple of engineers to make your product kind of live and it then takes you like 3 months. And the problem, the mistake that every single startup out there always has made is like build it, then they come is kind of like the classic quote here of just like you just build a product and you hope for the best and no customer actually ever shows up. And then you go back into like, oh, should have probably done some go-to-market and marketing and sales kind of thing. I felt like the way it’s flipped now is two parts here for me, actually. Like the first one is probably a bit of a, like a personal lesson in life and bias also as well, which is attitude over skill. That is kind of like my personal hiring philosophy where it’s actually not around the person who can do the job already. It’s actually about the person who just loves your company, what you’re building and so on, and will just go out there really hard and just learn everything about it. Because like skill itself, you can always upskill somebody, they will learn on the job. Whereas I’ve seen founders do that mistake all the time, be like, oh, I need X skill set. They hire somebody super senior into the role, but then like the job description changes 6 months later and you’re stuck with the wrong person kind of thing. And that’s kind of like one thing. The other one, back to our conversation here, is you actually want people who are experimental, who love just shipping things fast, and that kind of like asking great questions and curiosity and kind of like bias towards action is actually what you want to hire for. So it’s almost just like hire for that attitude rather than skill and hire somebody who just loves shipping, for example, with AI. And you’re literally just like building that AI-native company from the get-go because suddenly every single person is just not afraid to just like touch AI tools, get something out there, experiment. And suddenly the experimentation now is off the charts, like compared to like doing one experiment every 2 weeks, you can now launch so many different ideas and things. So that’s kind of like hiring from the beginning is like my, again, like personal bias, like hire junior, let the team grow with you and just like unleash them essentially.

James MacDonald [23:22]: So hiring for that personality, that skill set, the ability to learn, because I think that tooling is changing so quick. You could be, I think what’s quite funny at the moment is AI engineers and a lot of companies hiring for AI engineers who are essentially software engineers or data engineers who maybe using some AI tooling, or you can go quite deep, ML engineers. But the whole idea that these AI engineers have 15 years more experience— there’s a very, very small percentage of people who are genuine AI engineers who have 5+ years experience, right? 10, 10 years. Like, you’re talking the minute few. But people are paying overs because you put the word AI engineer in the thing. If you’re a larger organization and you want to, you want to start to hire for these type of roles, you want to become more AI native, you want to ship faster, How do you go about that from a software engineering perspective?

Michael Batko [24:09]: Yeah, that is a fascinating one. I think there’s two parts there. Like the first one is like, what is the actual skill set? And the other one is actually like a deep AI engineer itself. The skill sets itself is actually not even the one of, like a year or two ago, we threw around the words of like prompt engineer and like prompting and stuff like, right. And getting really good at it. Back then AI just wasn’t quite there. Then you actually had to design a good prompt in order to get that.

James MacDonald [24:33]: Which has changed significantly.

Michael Batko [24:34]: Which has changed so much. Like now I just talk to it, give it as much context as possible, and actually it’s really good at distilling it. So the skillset’s really changed. So you don’t, even if to use AI itself, you don’t even need to be an engineer or have that skillset anymore. So that is actually quite interesting from a job description perspective. Where AI engineers really do come in with the deep experience and the 10, 15 years of experience and actually understanding the models themselves, the maths behind it and so on, which is actually really that is a really unique skill set, which is incredible. Agreed. But that is actually almost as if when AI is at the actual core of your company rather than enabled. And most of like 99% of the companies are like AI enabled. You don’t need the deep AI engineer who understands all of it, the inside out. But if you are the company, you absolutely need that. Yeah. But they’re also like pretty high salaries from what I gather. Yeah, they’re very high salaries. Especially in San Francisco.

James MacDonald [25:23]: Yeah, because you’re talking the minute few. But as you said, that’s if you are a genuine AI business as opposed to AI-enabled, which is becoming essentially more productive, more efficient, quicker outputs, better testing. You’ll build better product because you’ll be able to test that product with your market quicker. You potentially—

Michael Batko [25:42]: I mean, you can build your own models, and there’s actually like, there’s, there’s a moat in that as well. But it’s almost just like, are you that business or not? Then you need to decide, and 99% of businesses are not. And then you essentially want to go into like the skill set of experimentation, curiosity, and asking good, great questions.

James MacDonald [25:58]: Yeah. The size of technology teams in general, big trend at the moment, uh, a lot of layoffs out there in software engineering or technology companies in particular, not only software engineering roles, it’s happening in sales, happening in product. There’s also a counter argument to say companies have just hired too many and they’re using AI washing as an excuse to let some people go, make some redundancies. I’d love to get your take on where companies are at. What’s your opinion on what’s happening? Is it genuine, the company’s becoming more efficient with the use of AI, or is it, hey, we’re taking this opportunity to look at our profits, reduce some headcount?

Michael Batko [26:36]: Honestly, like all of the above is probably the answer of just like on the one side, like every single company always hiring is hard. You should always pretty much like, I think there’s that rule of thumb of like 5 to 10% of your team every year kind of thing. Otherwise you’re not being ambitious enough and like, not holding the bar high enough, which is actually really hard to do because nobody wants to fire people ever. So it’s always like a good excuse to essentially be like, all right, cool, let’s cut back the team which is not performing. That’s, I think that’s often the argument of AI washing, which totally fair, like most companies actually should do that on a more regular basis.

James MacDonald [27:08]: I love that take. It’s a controversial take. It’s brutal because hiring people is brutal.

Michael Batko [27:14]: Yeah, it’s hard, like actually like getting it right and like honestly, if you are If every single person you fired is still there, like you’re probably doing something wrong because like you can’t have 100% success rate of hiring people.

James MacDonald [27:27]: I’m in recruitment, it’s impossible.

Michael Batko [27:29]: Yeah. So it’s like you actually do need to prune, I guess is the other way kind of like to say it. So that’s kind of like the realistic take there. On the other side, I do actually really think that AI has that massive effect as well of just like actually rethinking how people should be spending spending their time. And I actually do think the teams will get smaller and should get smaller. That it comes with a massive culture change because you need to upgrade your team on AI in the first place and actually believe that they can do it. But for example, Anthropic actually released a bunch of great articles about their team. It said a really real change that they’re making is the ratio of product managers to engineers. Previously, I think it’s like 1 product manager to 5 engineers and you orchestrate that team and stuff. But now that the engineers can like produce so much more work so much faster to a higher quality, they actually quasi become product managers themselves already. So now the product manager job kind of gets squeezed as well, which then means that like you have so much more pressure there. So what I’ve seen, at least heard, is that you now have one product manager of like 20 or 30 engineers because you manage bigger teams because you can actually do it, which just makes your job more exciting as well. But it does actually— there’s a real change there in product management.

James MacDonald [28:41]: Yeah, I agree. In all of technology, that product manager role, as many as— or as much as any other role, has changed and will continue to change more significantly in two ways. But product manager becomes more technical themselves, but then the software engineers can take on more of that, you know, product management, traditional product management type role.

Michael Batko [28:58]: But then on the other side, the product manager role actually becomes more interesting because back to the other point that we mentioned, it was the one of like actually doing the thinking and the strategy ahead of time and the architecture and the structure of it is actually a really fun part of the job, which you now get to do. And now you get to do it with 20, 30 engineers rather than just like 5. And like, the, the velocity of shipping and stuff like is off the charts as well. So hopefully that job becomes more fun even though it becomes more scarce.

James MacDonald [29:23]: Yeah, I agree. Same from a design perspective as well. I think, I think that, you know, there’s a lot of, uh, talk, oh, I’ll just throw out Nanopadana too and yeah, I can do design, I’m now a designer. I got, you know, I can use Lovable, so I’m a designer, you know. I can look at UX, I’m a designer. Whereas I feel like it can go both ways. Both everyday people become closer to a designer, but designers become more technical and actually bring their ideas to life significantly quicker as opposed to I’ve designed this, I’ve architected the right way, and then I’ve handed it off to my engineering team and it’s not come back the way I actually designed it. So I think it can go one of both ways.

Michael Batko [30:00]: That’s super interesting, yeah, I like that. I haven’t thought about that part yet, but I totally agree there, ’cause from a design perspective, they would be really frustrating of like designing something and the engineer goes in and there’s like this cycle of weeks where like both of you just frustrated each other. But to your point of like totally designers become more technical, can handle a way better scope design to an engineer and actually get it live much faster. So the velocity of that almost like feedback loop, that’s actually super exciting.

James MacDonald [30:26]: Yeah. So if you talk about hiring team, you mentioned hiring people is challenging and it is one of the most challenging things. I work in recruitment day in, day out and you don’t get everyone right for sure. You mentioned you hire for like personality traits. Talk me a little bit more through that. Hiring a senior level, senior level role, let’s say we’re hiring a really important role in your company. How important is the personality traits, what that person brings to the table versus what they’ve done in the past?

Michael Batko [30:52]: Yeah, I mean, you would know this so much better than I do, but I can tell you all the mistakes that I’ve made in the past and kind of talk about some of the scar tissue. But for me, I call it literally like attitude over skill. And attitude is like on the one side personality traits, But also just like the fire for the company itself, like that passion itself. And that is something that you can never teach. That is something that you can’t even just like instill in your staff that easily because it takes time. But if somebody comes in and they genuinely care about your mission, the vision, like, you know, they’re never going to stop. They’re absolutely obsessed with what your company is about. Then it’s almost just like create a role for them in a way of just like throw them at problems and they will figure that out with the hunger that they have. And that’s actually the kind of like most beautiful hire that I’ve ever— like, those are the best people I’ve ever hired. In particular, if they’re actually juniors. And this is maybe my other bias, which is like junior people over senior people and then letting them grow with the company. So much better also just to promote internally because people have so much more context and you can know them and trust them. Senior hires, I don’t know how you do it because it’s really hard. Like, getting a senior person right is like—

James MacDonald [31:59]: I completely agree, but I think it’s the same thing. I think the motivation is probably the biggest question that you can get right or wrong in a hire. Understanding what the person going into a new job, what are their motivations? Because if they’re motivated by money and money alone, going in a not-for-profit is not gonna work for them, right? And if they’re super passionate about an industry though, and you’ve got that right role for ’em, to your point that they’re gonna thrive and they’re gonna continue ’cause they want to solve this problem ’cause they love that industry, but they’re not as motivated by money, like that industry-wise is the better role for them, right? So it’s like money doesn’t solve everyone’s problems. Industry, Lifestyle, people might not want to commute, spend time with family, different stages in life. So I think the motivator part is the most important part because, as you said, skills can get you at least 70% of the way. And if you can get that with the right attitude, you’ll learn the rest. It’s like taking a 50-point job description, you know, 10-point, you’re like, okay, you really don’t need all 10 of those points. What are those 3, 4 core skills that you need? Let’s nail that. And then get the right person with the right motivators in. And those are the people that I tend to see succeed.

Michael Batko [33:06]: Yeah, I love that. Actually, to build on that, there’s another kind of core value that I live by when I hire is the hungry over the proven. And I think especially in the senior teams, that is really relevant because it’s so easy to have had a— so easy to have had a long career and have lots of achievements. But do you still have that hunger to essentially keep going? Because are you going to rest on your laurels and just be like, I’ve done it before, I’m going to have a cushy job, I’m going to delegate everything to my team?, or do you really have that hunger to like take the next level, which speaks of that attitude again, which you spoke about as well.

James MacDonald [33:37]: I mentioned one part there which just opens up a conversation. Senior level people, you’ve mentioned juniors, you like hiring juniors with the right attitude. Seniors have got the experience and there’s this middle ground as well, right? There’s a lot of talk at the moment in you’ve got that senior level experience, we’ll go more software again, you understand architecture really well, you understand security, you understand data privacy, how to, uh, how to, you know, set your data in the right ways to actually enable AI tools to sit over top. Senior level experience. You’ve got the juniors, which are— don’t have a great deal of experience, but they’re AI native. They have been Claude code 10 hours a day for the past, you know, period of time. You’ve got this middle level, and there’s a lot of talk at the moment in the media around are they the people in the most danger, uh, the juniors are going to become more efficient, and you could probably go from a junior with next to no experience to a mid-level engineer reasonably quick now, or at least quicker than you could in the past. Yeah. How do those mids then, you know, either get themselves in those senior roles? And then again, from a senior perspective, do you— you can’t rest on your laurels and just say, hey, I’ve done all this before, this is how software was built in the past, this is how it always gets done. You also have to lend in the tool. So be keen to understand your understanding or your perspective on which roles are in most demand, which roles have the biggest challenge lying ahead?

Michael Batko [34:55]: Yeah, I’ll give you two kind of like almost like conflicting views on this one. Yeah. The first one is the one of, honestly, the hardest thing about running any company, and I think most people would agree with me on that one, is just people problems. Yeah. So, like actually people management is the one skill which never goes away anyway. Like actually being a great manager, really hard. Like takes a lot of experience, like it takes a lot of empathy and scar tissue as well. And like you become better over time. So, like that is actually where management and middle management as well as senior management need to be incredible at, where almost no level of AI enhancement can actually get you there. So I do think that skill set is actually super unique. Whatever else AI is going to replace, it’s something that it’s actually really, really crucial. I’m going to conflict with myself there with my next statement, which is what I’ve actually seen other teams do, which is some of the most impressive AI-native teams, is go from having the classic kind of like, like one manager and up to 5, 6 direct reports, which is most of us capping out at like 6 or 7 maybe direct reports, to actually having an AI chief of staff manage a team of now 15 or 20 people, which is, I mean, quite interesting. So I’m really interested of how that’s going to play out because essentially what they’re doing is still having their personal relationships, but it’s now 15 or 20 people, which is crazy. Like, there’s a lot of direct reports, but having AI as a middle layer now to actually do all of their kind of like what are the top priorities, how did you perform data collection, to then surfacing it to the manager itself for that manager to be across so many different things and actually almost like flattening the whole structure of the organization and empowering everybody to go fast and further, which is quite interesting. I’m really curious how that’s going to play out. I don’t exactly know yet, but in theory, there’s actually some beneficial things to it because like that, even the flattening of the structure itself, is interesting. On the other side, the kind of like human element is actually lost because like often to your manager you actually want that kind of like direct connection to actually, um, I don’t know, like a personal relationship really as well.

James MacDonald [36:57]: Yeah, if I counter that on one side, if you’ve got— if you used to manage 6 people, a lot of your one-on-ones— did you do this, how are you going against your tasks, that’s all done by AI now, right? You don’t have to wait for your weekly one-on-one for the manager to know how’s this person progressing, what have they shipped this week, etc., etc. Let’s say that’s all taken away. So that’s getting done on a daily. You’re getting a daily update, what’s been done, how are we tracking, what tasks have been done, da da da. So take all that part away of it. So we don’t even have to talk in our weekly one-on-one about, you know, what’s being shipped or what. We can just talk about how you’re going, what are you motivated by, what challenges are you facing, how can I enable you, how can I unshackle you? And you get that personal part back by ripping the crap parts of the one-on-one away.

Michael Batko [37:41]: Yeah.

James MacDonald [37:41]: So if we can— if AI can do that and make that a more efficient process where you walk into that one-on-one, you both know what’s been achieved, what hasn’t been achieved. You go in there with all the information already there, dashboards are done, whatever. Then it goes back to that more person-to-person relationship, which I completely agree will be the difference between companies succeeding and not. And I think it also ties into a lot more companies moving back to some form of hybrid face-to-face relationships, in being able to build relationships. Whilst AI makes everything more technical, I feel like there’s a big push to get back to some more human and human relationships, building team, building culture.

Michael Batko [38:21]: Totally. To counter that though, like, that’s interesting as well to, to like build that point out, is, um, that takes a lot of emotional energy, right? It’s like, yeah, sure, like you can take out the top 3 priorities and have it perceived against your job and stuff, which is almost like the easy part. Like it’s actually the human part, which is so much harder, right? Because like taking on all your team’s like sure wins, but also like challenges, people with like the family getting sick and all those kinds of conversations, actually really, really difficult on people managers. And like, that’s kind of like interesting as well, because how that’s going to play out. Because like then sure, you’re going to have more direct reports, but then the emotional energy required for every single one of those one-on-ones is actually really taxing. That’s a tough job. I mean, middle management is hard.

James MacDonald [39:01]: I agree. I’m going to go two different points on this. First, for the— if you are a middle-level manager, you’re somebody on your way up in your career and you want to continue to go down that people management route, I think what’s interesting— it used to be hard to go down people management route to get more upside to grow through an organization, whereas the last 5, 10 years, I think there is really a place to grow from a pure technical perspective. You don’t have to manage people, which is great. Of, because people management is not for everyone. But if you do want to go down that people management route, communication is obviously important. Empathy— are these skills that you think are innate to people? Do you think there’s skills that people can learn?

Michael Batko [39:40]: Yeah, fascinating. I’ll actually approach this one from a slightly different angle, which is, um, a challenge which I literally heard about yesterday from, from one of my clients, where we come in, we build AI tools for you, and the question kind of came up of just like, do we draft email replies, for example, which are customer-facing for them? And the answer was like, yeah, of course, like we send the same template to a billion different clients. But then that’s external. And then next conversation was like, what about internal kind of communication? What if you kind of like let AI draft kind of like, hey, happy birthday, hey, this is a milestone, hey, this is like our end of week summary and whatnot. And actually in that case, the answer was like, we absolutely don’t want to do that because communication then changes into like, well, all your communication is now AI-driven, your company has no culture, it’s got no flavor, it’s got no uniqueness about it. And that is actually quite an interesting dystopian future almost of like, to your point of like communication is actually really important, it is unique, it makes the organization, which is really the people, like what it is. So almost like that’s the part of the company which you never want to lose and a skill set which you absolutely need to hone and get better at. And back to your question, like, is it innate or can you learn it? I think some people are naturally just like better communicators than others, which are often people who approach life with curiosity, have grown up in families as well where like arguments and debates and stuff are encouraged. This is actually quite interesting kind of like dynamics that I see there. On the other side, it’s definitely around something you can learn as well, like exposing yourself to the right people, having the right manager above you as well to stimulate the kind of communication. Communication, getting better at written communication in particular, there’s actually courses and everything you can do. But even just, this is probably the best tip I can give anyone who wants to get better at the communication. Every single day, think about whenever you get an email, whatever somebody has a presentation, whatever somebody says to you, and you notice this, it was really good. It’s almost like write it down, capture it somewhere and ask yourself, why was that good? And even just a little like daily almost rhythm of just like, why was this email amazing? What did I love about it? And then just capturing that is almost like a sure way to just skill yourself every single day on great communication.

James MacDonald [41:48]: I really like that. I’m gonna go the flip side again. You mentioned, uh, management’s hard, taxing, emotional part, really taxing. Do you think that just more leans into teams hiring more agents rather than more humans? So instead of if I’m building a software engineering team, I’m building a team of agents, yeah, rather than building a team of people. I started using Pulsia piece of software built, one-man company. And his whole philosophy is around— he worked for Travis at Uber, and I think Travis’s Uber was like one-click, uh, one-click car ride. He’s one-click business, and he’s got— it’s a one-person business that builds one-person businesses. So he has agents working for architecture, agents on the infrastructure, agents working for QA, da da da da. His whole philosophy is to build a team of agents for each company. Yeah. That’s one-person businesses. Yeah, but how does that apply even to bigger businesses? Are we going to go to the point where as a middle management, I don’t want to deal with the people issues, I don’t want to deal with sick days anymore, I don’t want to deal with that emotional toll? Yeah, instead I’m just going to hire a bunch of agents.

Michael Batko [42:56]: Honestly, that’s a fun conversation because I literally had this conversation with somebody a day or two ago. And from my personal experience, again, because I feel like I’m tapping out out. So that’s kind of like an interesting one. But like, where is the limit? It’s kind of like an interesting question there. But overall, I would say yes, there will be a squeeze on teams. Like, you should actually automate a bunch of the things which people don’t have to do anymore. So I do actually think teams will get smaller. Where’s the limit is kind of like an interesting one. And there’s like all those headlines around like a 1 or 2 people billion dollar company and so on. And to the point of I’m tapping out of my potential, which is kind of an interesting challenge that I faced about myself now after 10 weeks of building so much with AI and I’ve I’ve got probably like 5 different projects going at any time right now, is I realized that there’s almost like a bit of a ratio of yourself and the Claude conversations you can have, or yourself and how many agents you can manage. And that ratio is really real. Like there’s just that much context that you can hold that many different conversations. And again, just to illustrate this a little bit, like on my laptop, if you looked at it right now, I have 10 desktops open and every single one of them is a Claude Cardin conversation, right? Working on 10 different problems at any given time. That is kind of like my current status quo of how I work. And then I’ve got probably like another like 10 or 20 agents working for me, throwing information back at me. But that’s kind of almost like my time limit and my context limit. So it’s like I can empower that, but at some stage, and this was literally a conversation from a day or two ago, I’m now looking into getting people in to almost like co-found those businesses with me. Because I almost need people to scale me further. And like, so there’s definitely like a limit there.

James MacDonald [44:34]: Yeah, no, I like it. I, I completely agree. Everyone— and you can write, you can have the best agents working for you to go across at 10 projects, what are my biggest priorities? There is only so many hours in that. It’s only so much context you can have.

Michael Batko [44:46]: I mean, I also like, just to take it to extreme, I’ll give you another example because I actually did try to take it to the extreme where the Batgod.ai company that I mentioned, which is helping founders solve problems with AI, I literally went on a tool called Paper Clip. And what it does is it spins up a whole organization for you. So I literally gave PatGuru.ai a strategy, which is very comprehensive, and I hired the CEO. I gave the CEO the strategy and the CEO then the first job is hire a team underneath you, CTO, et cetera, et cetera. Then they hire a team underneath them and then you just unleash that CEO to just execute in your strategy. And that is one way to work with agents and you can hire yourself a CEO of that role in a way. But even then, this is like, yeah, how many of those types of projects, how many of those types of CEOs can you do? And again, like, I’m just like tapping out because there’s almost like too much information coming back my way to make decisions. To be fair, like, another skill set that’s going to be really important working with AI is actually decisions are like off the charts. Like, the number of decisions that you have to make, you get into decision fatigue as well. But to be fair, that’s also the fun part of the job.

James MacDonald [45:53]: Humans from dawn of time have been operating in tribes and groups. There’s a lot of We went COVID, everyone was isolated, and I feel like the mental health challenges on the back of that and a lot of loneliness come on the back of that. AI is, you know, providing opportunity for these one-person companies. Do you think that there will be a switch back though to humans wanting to be around other humans, humans operating better in little tribes and teams? Because I think if you have a look at most businesses, even the way businesses have been designed in smaller pods and a pod organises something, different pods working together, and then they all accumulate together to come to form a company. Company and, you know, push a company forward. So I’d be interested to get your take on just that human nature side of it, which overflows to your one person’s own got so much context.

Michael Batko [46:39]: Yeah, I, um, I probably have a slide, as in, like, I probably have a, um, a different thesis there to what is actually probably out there, which is previously, like you said, like, teams are created almost just like in order to make more work happen, right? Like, you almost needed a bigger team to just like execute on more things, which which now with AI isn’t the case anymore. With agents, you can actually scale yourself to actually do quite large amounts of work itself. I do think where teams and other people really come in play is, well, number one, the loneliness. It is just like, sure, you can talk to your agents. Me and my wife joke all the time that all I do is talk to Claude, my friend Claude. So that’s one part of loneliness, actually having somebody around. But the other one is actually why you want other people around you, why you would hire a team, is actually inspiration. With AI because you and I talk to the same AI window. You and I are going to produce completely different outcomes just based on our previous experiences and what we know AI can do. But as soon as we read a blog post, see, have a conversation, read a newsletter from someone else, suddenly you’ve got this inspiration. You’re like, “Oh my God, of course that’s possible.” And next minute you just throw it in there and you’ve got a way better output and outcome. So back to teams though, the reason why you want other people around you is almost just a vibe on it, to see how they work, to get inspired by the questions they ask, by seeing them, overhearing them say something, and like, hey, that’s really interesting. And that’s actually what really unlocks you as a team to get better and better over time.

James MacDonald [48:07]: What you said there feeds back to something you mentioned before about companies getting rid of, changing over 10-15% of their staff each year, which I think is great because what it allows is then fresh eyes to come in. I had this conversation earlier this week about big successful Australian-based organization, 250 people in the technology team alone, own, just brought in some fresh, fresh eyes, people from a different vertical altogether, but applying that vertical and what they’ve learned over here in this new industry has brought along new ideas, fresher ways of thinking, as opposed to being able to sit within your own, you know, your own context and what you’ve only ever known. It’s the same thing to your point, what you’ve just mentioned there. From an individual perspective, reading a blog post, listening to a podcast changes perspective. In a bigger organization bringing in that external talent. And I think that potentially feeds into what you’re actually doing day to day at the moment, where companies are employing you to come in and fresh set of eyes, different ways of looking at things. You’ve seen a bunch of startup works, but you’ve also worked in, you know, corporate and enterprise. So taking a different perspective and applying that to a new vertical, new size company provides different ways of thinking.

Michael Batko [49:17]: That’s the most fun part right now, honestly. It’s like that first principles thinking is just absolutely game-changing. And if you are the type of person it’s like a kid in a candy store because you can literally rethink anything. And the question which you need to ask yourself is just like, what is the best way to do it? And previously we were all limited by engineering resources, by design, whatever it is, by getting it live and so on. Whereas right now, what is the best way to do it? And you can actually make that happen. Just to give you an example, yeah, what we do is we come in for 30 minutes, we interview your team, and then we present back of like, here are lessons. The way to do it is like traditionally would have been like, here’s a PDF presentation, here’s an Excel sheet just like, and I come in for an hour and present it back on a big screen kind of thing. Whereas I literally asked myself the question of like, well, what is the AI native? What is the most incredible experience you can possibly have? And within a couple of hours I spun up the entire new experience, which is as soon as you come to me, I literally send you the link to your own website. And on the website you fill out the whole survey, give me all of the information, tell me who you want to interview, that’s automatically triggered to actually get the 30-minute setup of your to the team. As soon as they give me the— we finish the conversation, all of those notes are sucked into your website, automatically created already, all the AI tools that we can create for you on the roadmap of quick wins to bigger builds and stuff instantly with budget and ROI for you. And rather than me creating any PDFs, any Excel sheets and stuff, it’s literally just there on the page where you can log in anytime and see the progress of any tools that we’re building. I’m just like, that is just like, that is a slick experience for any client being like, I know where to go. I’ve got one place and all of the information in there, every single word is transcribed. You’ve got full context, every email that you’ve sent me, every conversation that you had. And just like, that’s a small example of like, even just rethinking like, Jesus, like PowerPoint presentations take me forever. And most people as well. And it’s just like so annoying when I relocate like a picture to the right or stuff.

James MacDonald [51:10]: Not with Grammar anymore though, like that’s changed the game as well, right?

Michael Batko [51:13]: True, absolutely. But like the beauty is kind of like you get to ask the question, which is like, what would an incredible experience look like? Like and actually just not be limited by how things have been done before.

James MacDonald [51:24]: I’ve had that conversation with Claude a couple of times. Like, what would this look like if we wanted a 10x, a 5x? What else would that take? What agents do we need to be building? What do I need to be doing?

Michael Batko [51:34]: My favourite prompt is, which is, I make this the most incredible experience possible for people to remember forever. Let’s see what happens.

James MacDonald [51:40]: Pretty high standards. Yeah, yeah. Mate, you’re working with different size companies, you’re working with different individuals at the moment, individuals coming to you and asking, what’s my job look like in the future? How do you approach that? You’re obviously working with companies and you’re offering them AI solutions that can change the way they work and the way their company operates. But on the flip side of that, fear. Fear is a very real thing for a lot of people in all sorts of jobs. There’s a changing nature of— there’s a different article every week about legal profession won’t be there in the future, the accounting profession, the recruitment, any white-collar roles.

Michael Batko [52:15]: Yeah. And honestly, I actually believe in all of that as well. Like, it is scary and it’s like if all those types of jobs will be fundamentally shifted and changed, if not nonexistent in the future. The only positive I’ve got for somebody who is listening to this right now is like, at least you’re listening to this right now and you are literally thinking about that already. And either it’s you who wants to make a change or it’s your employer who’s pushing you to use AI. And if you are in that situation, you’re already in like the top 1% of actually just like getting ahead and forging the future and making it happen. Especially if your employer is pushing you to do AI, they’re literally trying to get you in the best possible situation. Sure, you might get fired, you might leave the job and so on, but the skill set you’re going to learn, you can then employ in any other organization. So you’re already ahead of like 99% of other people. So if anything, I would literally be like, lean into it 100%, automate yourself out of the job. Job doesn’t even matter, but that skill set you can take that anywhere and you’re going to be so crazy employable anyway. So it’s like the positive is almost just like don’t think about your current job because you’re already creating your future job within.

James MacDonald [53:18]: Yeah, I agree. It’s building that skill set, it’s transferable, right?

Michael Batko [53:21]: Oh, like I mean, we’re literally talking every single person we need to hire is somebody who’s already tried AI and done it. Like that’s exactly it, especially if your employer is pushing you to do it, like lean into that. And I think the other funny part that I keep hearing people say is one of the best employee perks you can get is just essentially like a Claude called max subscription or like a billion tokens to essentially play with, right? Because like that perk is better than any snacks in the office.

James MacDonald [53:46]: Oh yeah, ping pong table doesn’t, uh, yeah, because there is nothing more frustrating. I think my early days of Claude, uh, you know, you’ve hit your limit, uh, buy more, buy more. I think the fourth buy more, then I upgraded to the max plan. I was like, I don’t know how much this is going to cost me.

Michael Batko [54:00]: Yeah, 100% the same here.

James MacDonald [54:02]: But there’s nothing more frustrating than hitting your limit and you’re having to wait 24 hours That’s literally me last night.

Michael Batko [54:07]: I ran out of credits, not 24 hours, I’ve got nothing. And I literally had to look for my to-do list, been like, what can I even get done?

James MacDonald [54:13]: Anything else really interesting that you’re noticing, that you’re seeing in the market, interesting takes that you’ve heard that you’d like to share, which is, you know, top of mind for you?

Michael Batko [54:21]: Yeah, well, one question that I asked myself, and I don’t have the answer to that one, but I actually find that a really interesting question to ask, and is the one of, um, do we value work which has been done with AI more or less. Just to illustrate it, like you hire a consultant and the consultant back in the day charges you whatever it is, $20 grand for a project that they deliver. The same consultant does better work, delivers it within one day rather than two weeks, same $20 grand, but it’s all AI kind of like enhanced generated, but the work is actually way better. I just find it a fascinating kind of question of like, should you be paying the same amount of money? Should you be paying more money because it’s actually better? Or should you be paying less money because they’re actually like like, you know, it’s AI enhanced. So that’s kind of like a really interesting kind of like, I feel like the human perception at least is the one of like, oh, it was just AI generated, I could have done that too. Which actually felt like the really counterexample there is, well, the way, the reason why I was able to do that better is because of all the experience I’ve had over 10 years and I’ve fed my like Batko brain, for example, into it. And that’s why it’s so incredible.

James MacDonald [55:24]: Completely agree. And I think what’s quite interesting, if I look at the legal profession and I don’t know a great deal about it, But from what I understand, the top people know things inside out, back to front, didn’t they? They become, if they’re already unbelievable, they become even better using AI. At some point they die out. So you’ve lost that experience plus AI and you’re being replaced by no experience because you use AI from the ground up. I have a look at from a recruitment perspective, I’ve worked in a day-to-day human-to-human business. But I can AI a lot of what I do. There’s still a human element which makes me being able to prompt, maybe able to set it up better, will provide a better experience. But that’s based on my experience. Without that experience, the AI experience wouldn’t be as good, the, the output wouldn’t be as good. So what happens when the people that have used your, your 20 years of experience— you’re better for that 20 years of experience. Without that, does the next generation trying to provide the same services you, are they able to do the same things? Or if they fed your entire brain into it, can they, can they take that and go from there?

Michael Batko [56:31]: So it’s really— yeah, so that’s how it— like, just, I love that kind of question, right? Because, um, to illustrate it in a slightly different example here, again, is like, you want a bit of legal advice, you hire somebody with 20 years of experience, has been in the industry forever, versus a university student who just throws it into AI, scrapes the entire internet. The advice yourself that you get is probably like equivalent or on par. Of course, the person with the experience can then actually sign off on it and actually give you that like, I take the responsibility for that, which a uni student can’t. But the quality of advice actually potentially is actually very, very close to each other, which is kind of like the interesting point there of just like, well, yeah, if those people die out and anybody’s ever done legal advice, for example, with AI only, that actually becomes like scary or the very different world. Which kind of led me to another kind of thought prompt, which is like, how do you even get ahead in the world of AI? My prompt is just as good as your prompt. It’s like, it’s the closest comparable I have is, and this is a bit of a geek one, is I don’t know if you’ve ever played a computer game such as Warcraft or League of Legends, and everybody starts at the same, everybody starts at zero. They have no gold, they have got no skills, they’ve got zero experience points in a game. But actually proper gamers who just like actually do it for a living, they just like will win 10 out of 10 times. Even though you start at the same kind of like starting point, which is exactly the same way with AI. You and I do the same task with AI and in theory we have exactly the same tools ever from the get-go. But because of your experience and because of the knowledge and the decisions you’ve made, the uniqueness, all the different legal deals that you’ve seen over the years, all the negotiations you’ve gone through and stuff, all of that kind of experience, experience is the kind of stuff which allows you to essentially throw it into your AI brain, but also allows you to prompt way better than anyone else. Which kind of comes back to a little bit of like prompt engineering, which isn’t really prompt engineering, but actually knowing what questions to ask.

James MacDonald [58:23]: Even if it’s not prompt engineering, it’s having that conversation with context because you’ve got that proper context to be able to provide it, right? Because you’ve seen more, done more, had experienced more in the past.

Michael Batko [58:33]: Totally. Yeah. Which then comes like It’s almost like your experience in life becomes your moat in a way. Like the decisions you’ve made, the failures you’ve had, the things you’ve seen which are not documented, not public, the uniqueness of your life which is private, and only the things that you hold in your head or have documented for yourself, that is almost like your moat itself.

James MacDonald [58:52]: Which comes back to the more human, human part because it’s actually who you are as a human being, not just, you know, your prompt or your Claude Brain. Which I think comes back to this point. I sort of deal with this the two sides. I’m like, one side’s a very, very AI, very, very— it’s going to provide me a better outcome because it’s more accurate. And the other side’s this human nature part to understand human relationships and empathy and what experiences I’ve gone through which has made me me, which might be a reason you like me or you don’t like me. And so it’s that quite interesting part where you, you’re getting pulled from both, both complete different ends.

Michael Batko [59:26]: Yeah, which is, I guess, like the other thing that you, you might have been reading a lot about, which is like in the world of AI and the world of abundance, in the world of just like having news and information thrown at us in all different directions. Taste becomes the thing which is actually the unique differentiator. Personality becomes the unique differentiator. Trusting somebody because they curate things in the right way becomes the differentiator. And which I actually find quite interesting, that is almost like the bright side of all of this. It’s just like authenticity of yourself, of the way you look at the world, and that actually attracting people to you and eyeballs or clients or whatnot, that actually is almost just like almost the exciting thing to me.

James MacDonald [60:06]: I agree, because again, authenticity comes back to the actual human behind it, and they’re not— I’m trying to AI wash myself or, you know, throw this filter over top because I have access to all this knowledge. Totally. I like it. They were talking earlier about the value, and the value of somebody actually providing their context done with AI is more valuable because they can turn around quicker. I think the real interesting part at the moment is for creatives and for artists. You have a look at it from a music perspective, you know, a lot of music, they understand the, the formula that works in pop, so then it’s just being generated, right? And then there’s this flip back to somebody that can actually play an instrument, can actually write their own music, which people are really leaning into that. AI-written books, for example, there’s hundreds of thousands of them being produced. What’s the quality then versus an actual human being that’s put their thoughts on paper? So be keen to get your take just from a human creativity perspective, that authenticity, is it still— does it become more valuable as we go forward where people are like, there is so much AI slop out there, I want to hear the real human behind this, the real creativity?

Michael Batko [61:13]: Yeah.

James MacDonald [61:14]: Or is it just, hey, this is produced at a higher quality so it’s better?

Michael Batko [61:19]: Honestly, I mean, I don’t, I don’t know the answer to that. I like thinking about it on a scale of what’s the progress of technology and what is inevitable. Like, what is this kind of like progress that cannot stop and it will never stop whatever you do. There’s actually a great book, I think it’s like The Seven Powers of Inevitability or something like that. And that concept is quite interesting because around most topics you can argue like a pro and a con. But what is actually the rate of change where nothing you could do will stop it? For example, the advancement of AI. Governments can shut it down because it’s scary and so on, but then another government lets it run loose and then AI AI will inevitably keep getting better. Like there is almost like, what is the only limit is potentially like power, like compute power, et cetera. But again, like that’s inevitably gonna get better and better. So like to back to your point of like creativity and like AI producing stuff, again, inevitable, like it will keep producing creative things. It will keep producing better books, better videos and stuff. So like that is almost just like, like without doubt, whether it’s going to take a year, 10 years and stuff, you’re going to have insane podcasts which are generated. I think there’s like whole like artists’ songs which are already ranking in the top kind of like whatever the rankings they are. So back to the human element, where does the human element then kind of come from? I do think like humans, and this is maybe more on the psychology side of things, do still romanticize that kind of like human done versus stuff. Like computer done, right? So it’s not even necessarily around the quality, ’cause the AI stuff, honestly, it probably will be better than the human stuff. But it’s actually still the connection to the human that we value. That is actually the reason why we choose a certain architect for your home or choose a certain artist to like get the paintings for your home and stuff, because you actually have the connection, a story, they live in your town or whatever it is. And it’s not the best painting you’ve ever seen, but actually it’s the stories and the human element, which is still like, back to taste and authenticity, like stories become super important. Like you actually just want to find connection to other people.

James MacDonald [63:21]: I love it. I think that plays out in the music world, right? You engage with the people that you feel like you can relate to or tell the stories that, you know, to pull on your heartstrings or your emotions, right?

Michael Batko [63:31]: Totally.

James MacDonald [63:32]: I love it. Any other one piece of advice you give to bigger companies, small companies, people that are considering the changing nature of their technology teams at the moment? Are not currently using AI, using AI a tiny bit, want to change the way they’re operating, want to lean into this a bit harder, what’s the first couple of steps? What’s the first couple of hires?

Michael Batko [63:54]: I mean, number one, it always comes from the top. So like, I think any kind of culture change and any bigger change like that always is driven by leadership, just from even like a cultural perspective and a wanting to change perspective. The other one is just to literally encourage the kind of like playing around that I mentioned before. It’s not around getting AI into the core of the product. It is just around giving people the permission to just have fun with it. It literally is around creating funny pictures on Gemini or whatever it is. And that in itself, once enough people have used it, they start chatting about it because it is exciting. And before you know it, it actually keeps building an organization to just push the boundaries more and more and more and more. And once those boundaries are pushed, you’re going to get to the natural stage of like, all right, cool, what is our AI policy? Policy and at that stage you actually then really unleash the whole team because they know what they can and can’t do. So that would be probably my biggest advice is like get leadership top down to actually say what people should do and then allow people bottoms up to essentially just experiment and then you’ll meet them somewhere in the middle to actually now set an AI policy as well as this kind of brain for your organization itself.

James MacDonald [65:03]: Yeah, and I think on the back of that you just gain some momentum. It’s getting the early wins and the wins will multiply and it snowballs.

Michael Batko [65:10]: Goals. Yeah, absolutely. That’s it. That’s the funnest part because like you’ll see like people literally around the water cooler just starting to vibe around AI. Like you and I like started building a couple weeks ago and now it’s literally like all you talk about because it’s just exciting.

James MacDonald [65:24]: I spent my entire weekend, my wife’s questioning what’s going on.

Michael Batko [65:28]: But this is also because like the only limitation with AI is inspiration and it really is just the ideas that you get from other people and it’s just like magical moment where like somebody says something and you don’t have any, um, you don’t have any learning curve or onboarding one. Like you hear an idea, you throw it into Claude and it just like makes it happen. And you suddenly, you’re just like, did the whole pipeline of idea into execution has shortened so much. Then it’s just so fun, just try it out.

James MacDonald [65:55]: Yeah, I think it is. It’s a fun time to be alive.

Michael Batko [65:57]: Oh, 100%.

James MacDonald [65:58]: And for those that are interested in, you know, either your journey, following your journey and what you’re building at the moment, or potentially engaging you to come and have a conversation with their business, how people get in touch with you?

Michael Batko [66:08]: Yeah, um, best to jump on batcode.ai or, um, just hit me up on LinkedIn directly.

James MacDonald [66:14]: Beautiful, we’ll tag it up in the show notes. Appreciate your time today.

Michael Batko [66:17]: Well, super fun.

James MacDonald [66:21]: Thanks for listening to this episode of Building Tech Teams. If you’re currently building your technology team, having some challenges, or interested in what’s happening in the market, feel free to hit me up on LinkedIn, forward slash James McDonald AU. Or look at my company website ntp-talent.com.au, where we’re helping companies up the east coast of Australia find and help recruit the best technology talent into their teams. This episode was produced by Day One. Look forward to seeing you on the next episode.