EPISODE TRANSCRIPT:  

Welcome back to Things Reasons, the Independent podcast for Australian IT leaders. My name is Jono Staff. I’m joined by my fearless co-host, Naran McClung. Naran, what have you been up to lately?  

Well, uh, lots of different things, uh, in my personal life. Jono, of course, you know that I’m a very keen cyclist. 

Um, I would say I had the most, uh, impactful sort of, I would say, high stress race of my life. Last Sunday, it was the World Championship race for Grand Fondos, which is basically a polite way of saying racing for men that are too old to be elite. Pro-level cyclists, of which I wish I was one. Sadly I’m not. 

I’m in the 45 to 50-year-old age bracket, and I got to race against 176 competitors from all around the world on a course in beautiful Lorne Victoria that finished with 50 kilometers of racing along the great ocean Road that I have to assume just over my right shoulder was beautiful scenery, none of which I saw. 

Because I was racing my heart out, and so I did that and it was a momentous situation for me, and I, I had a good time doing it. It was a lot of hard work, but it was wild and I thoroughly enjoyed it.  

That’s huge, mate. I know you’ve been training for that, uh, like a monk. Mm. For about a year. Yeah. How did, how did you fare? 

Uh, you know, what, what are the results?  

Look, uh, I ended up coming 17th, um, out of 176. That was very much the best I could do, you know, every time we do one of these races, and I’ve never done one to this level of, um. Magnitude, um, or scale. Um, you always want things to go well, right? And so in a cycling race, you can crash, you can have a flat tire, you can have a mechanical, any number of things happened. I was firstly super grateful that I had an event less, is that the right way to describe it? Uneventful and uneventful that’s, and I had an uneventful race where I know for for sure that I gave it 120%. And so, yeah, I chuffed with it.  

That’s awesome. So you’re officially ranked 17th in the world. 

For your age category. That’s right.  

You know what I do now? Whenever I see someone who looks roughly my age guy, he’s probably not better than me.  

Yeah, it’s important.  

I think so. Yeah. That’s great. It’s about having goals though. Okay. That’s what it’s all about. It’s about a goal oriented mindset. A performance mindset, which I’m told Jono means it makes us better in our jobs as well. 

Right.  

I’m sure that, I’m sure there’s some benefit, but you know, congratulations. Thank you, mate. Fantastic achievement. Appreciate it. Um, while you were doing that, I was busy gallivanting around Japan with my lovely family for two weeks. Amazing. And can I tell you, first time in Japan and, uh, you know, what stood out to me was that uh, you can have nice things. 

Right. You know, Japan’s just such a wonderful country. So many people, yet nobody’s in a hurry. Hmm. Everybody’s polite. There’s no graffiti, there’s no rubbish. I just, I just couldn’t believe it. Beautiful. Um, people talk about it. But it was great to experience Disneyland, universal Studios, bullet trains. The whole shooting match was fantastic, so that, why aren’t we living there? 

Well, uh, you know, it’s, it’s a good question, but I do still think we live in the best country on earth. Okay.  

So shout out to Australia.  

There you go. Shout out to Australia.   

Fantastic. Well, let’s get into it. Let’s go. We got a new segment today. We do. We, we, we, I think borrowed the title Yes. From one of our favorite shows. 

Okay. Some of the fans of the show might know. So what, what’s the name of this segment? Uh, this is what grinds our gears. Shout out to Peter Griffin from Family Guy. Um, we have things that grind our gears and maybe it grinds the gears of our audience as well.  

Yeah, why not? Let’s, let’s put it out there. 

There’s a few topical things going on lately. Yes, we have feelings about these things. Yes. Uh, we have questions. We have questions. I don’t really think we’ve got the answers. We don’t, but we’re gonna, we’re gonna talk about it.  

We are seeking feedback as well. So as we rant, as we grind these gears, we, we, we will encourage our audience to reach out to us and tell us how you feel. 

Tell us whether you can relate. Did we miss anything? Was it a balanced opinion? Should we offer one, which we’re going to. Um, I think we, we, we need to be sense checked on this.  

I think so. But we have opinions and, uh, and we’re gonna, we’re gonna say them. Mm. Because this is an in independent show. Yep. We can, uh, we can say whatever we want. 

So, um, so get, get us into it mate. What’s, uh, what’s on your mind?  

There’s three things that I think we should talk about. Um, I wanna talk about AWS. Um, now AWS have obviously been in the press lately with, uh, a major global outage. These things do happen, right? So nobody’s pointing fingers per se. Uh, we know that CrowdStrike had a bit of an epic fail not that long ago as well, but it’s not so much the outage, um, that I wanna talk about. 

We’re gonna talk about spend, uh, we’ll talk about fiscal responsibility and things of that nature. We’ll talk about our own government, uh, and expenditure on AWS. That’s one thing. Second thing I think we should talk about is the commonwealth Bank. And Commonwealth Bank in relation to, uh, their recent epic fail and the optics that they endured to do with generative ai. 

And there was a whole saga there that I think is worth  

talking about. It’s the one thing that I think every business and perhaps agency is aware of and conscious of, and that is, is the AI gonna take our jobs? So let’s talk about Commonwealth Bank’s approach to that and how we feel about that, and whether those gears were in fact ground as a result of that. 

And I think the final thing we should cover off, um, is, uh, more generally, uh, AI adoption, return on investment. We have covered this in the past. It’s an evolving story and I think we can link that back to the other topics as well. Three things.  

Great. So AWS I think, uh, you know, and maybe some of our audience are rolling their eyes ’cause you’re the famous Azure guy, but, you know, um, doesn’t mean  

I’m not objective. 

Come on. Yeah, yeah,  

that’s right. AWS and then by extension, like hyperscale and SaaS. Sure. Uh, where, where are we going with this? Well, you know, what’s, what’s of interest to me specifically is, uh, for those of our audience who are familiar with it. Some of them may be the whole of government AWS deal. Yeah. 

Managed by the DTA.  

Surely that means value whenever something’s whole of government. That means that there are advantages that the whole of government get to enjoy. That’s what we’re to believe.  

Well, you know, you, you, you would think so. Mm, you would think so, but this is a, this is a big deal. 

Huge. And I haven’t seen anything in the mainstream, uh, press around this. Mm. Uh, I know we’ve done a bit of research on the numbers we have done, remember, but take, take our audience through the numbers.  

Alright, well, so I think, um, I can’t remember the exact year it started. You might be able to jog my memory. 

I know the figure. So there was $39 million, I think in the first agreement. You then cast forward three years and it went to about 174 million. It jumped shortly after that to about 390 million with about 140 agencies benefiting. From this whole of government buying power. Um, today, if we cast forward, so I think it was 364 million in 2022, and, uh,  

should, uh, discerning buyers send away an asset to figure out the exact figure, pundits will tell you it’s probably north of $500 million of AWS spend in Australia. For government specifically. That’s just government sector.  

That is a, it’s a huge, a huge contract.  

Outrageous.  

And I think what’s, uh, what’s of of interest to me is this is, uh, so this contract was up for the second, uh, renewal? 

Yes. I’m assuming  they went for another three year term. Yes. It’s been, uh, two, three year. Yeah. So this will be another three year term that is, uh. Beyond exponential growth.  

It’s insane.  

And, uh, I think it’s fair to say it has grown in terms of value far beyond the original expectations, which is not uncommon for, uh, I think not just, you know, AWS but hyperscale cloud in general. 

It’s why the finops industry exists, uh, to, to make sure that you are in control of your spend. But it seems to me like, uh, with that kind of growth curve spend. May not be as in control as the DTA would like to have the Australian taxpayer believe.  

So this is the Digital Transformation Agency. Um, they’re on record saying they’re proud of their finops. 

I, I think so. Yeah. There’s, there’s very various sources there.  

So here’s what we can deduce certainly within that 500 million we could say that. A significant proportion of the workloads in that spend would be IAS centric workloads, virtual machines, databases, things of that nature, workloads that could conceivably live somewhere else. 

Well, I think before the hyperscalers got to Australia, they did.  

Right. So there was some, there was a, there was an ecosystem of IT services and hosting and, and vendors, perhaps vendors that pay tax in Australia. I don’t know, just putting that out there. There were other ways to host these workloads before AWS rocked up. 

Yeah, that’s right. Right. And uh, so we’re asking the question. Mm. We’re asking the question, what, what, what’s actually going on here? Is this, uh, is this a good thing?  

Mm. Or is this just the easy thing? Right.  

And so what’s the difference between the good thing and the easy thing? Because sometimes, uh, doing the good thing is hard. 

And, and why is, why is this happening?  

Well, look, if I was to get my broad brush out and I like to occasionally, sure. 

I would’ve thought that any discerning buyer today cares about cost and risk. Any buyer. Right. So any buyer should want their money to go as far as possible, and they should care about risk in this world of cyber threats and all the rest of it. It wasn’t that long ago that cyber was the number one interest in Australia before generative AI came along and piqued people’s interests. 

And so cost and risk is supposed to be a thing. So I’m asking an open question, particularly with what feels like unbridled runaway AWS spend possibly for convenience. Don’t know posing that as a question. Um. Do our government agencies care about spend?  

I hope so, right?  

I hope so. I’m a taxpayer. Jono. I hope so. 

I think as a, as uh, professionals in this space, I think people are trying to do their best. Mm. It is a difficult problem to keep on top of. Yep. Hyperscale spend is a difficult, difficult. Problem. Right. Uh, it requires expertise and effort and doing what is sometimes the hard thing. Yes. Not the easy thing. 

And, uh, I, I, I would hope so.  

Alright. So. Clearly, I think there’s some questions that could be asked and answered in this space. Um, so I would, I would encourage our audience to reach out and just give us an opinion on this, whether you’re an AWS fanboy or not. All good, just feels like a lot of money, a lot of spend. 

And I wonder, you know, whether, whether, whether we’re seeing discerning spending going on there or whether, to your point, maybe it’s just easy and therefore that makes it right. I don’t know. That was another, the last time we interacted with the government agency. I mean, has service improved?  

Good question. 

Good question. All right. Okay.  

Now that links us into our next conversation around the Commonwealth Bank. So the Commonwealth Bank has been in the press recently. Jono, what went on there?  

Well, uh, from what I can tell, what went on there is, uh, Commonwealth Bank made a huge mistake. Yep. And I would preface this conversation with great power comes great responsibility. 

Yes. AI is powerful technology and you’ve got a responsibility, I think, to deploy it into your business and change the way that people work in a ethical and uh, responsible way. Yep. Uh, and I think CommBank found, found that out the hard way. Right. Some, some mistakes were made and, uh, fortunately they were called on it. 

Um, you know, some of the news reports I read, were along the lines of employees unwittingly training their own AI agentic AI replacement,  

and this is this fearless lady. And shout out to her as well. Like she’s obviously been very brave to come forward the way she did. Well, she was a 27 year veteran, spent six months training agents for the bank to then get the sack. 

Yeah, right. I I, I think that’s right. And look, um. You know, uh, from what I can tell the employees have been reinstated. Uh, they were all made redundant, and it was, uh, you know, it was clear that they weren’t redundant. What was worse was, you know, in the, the pilot of these programs, uh, the business case, the ROI was projected that it would reduce the volume of calls. 

Mm-hmm. And, uh, and have a net benefit in terms of productivity and you know, therefore these roles were redundant. And then once the technology was rolled out Yeah. Uh, it was proved to be the complete opposite. Yeah. Uh, call volumes actually increased. Yeah. And they were less efficient. Yeah. And so, you know, that’s it. It’s very, very interesting. A lot of mistakes made there through the whole cycle. Yeah. You know, and a lot of assumptions, and it links to some of the other stats around agentic AI adoption in, you know, that are kicking around. But a well-known one right now is that it’s estimated that 95% of agentic AI pilots fail. 

Yeah. This one was clearly, uh, an epic public fail. Right. But I’m sure it’s not the first. It won’t be the last.  

Well, so MIT described that as a hype experiment. Right. Linked with perhaps an ethical misjudge, which may be without the optics, without the public getting a chance to opine on it, the pub test. 

Um, I wonder if the Commonwealth Bank would’ve deviated at all.  

Well, uh, from what I can tell, it wasn’t until they, uh, faced some pressure from, uh, the relevant union. That, uh, that this decision was reversed. Yeah. So I think it, I think it’s a, a bit of a very public lesson mm-hmm. In, uh, what can go wrong. And I think, yeah, with great power comes great responsibility. So, um, that, that, that was a very, you know, just an interesting situation. We do talk a lot about AI, uh, and the the benefits. Yep. And, um, and how it can change everybody’s life for the better. Yep. I think that’s appropriate. Yep. But there is, there is another side to this, and I, well, there is, I think it’s down to just making good decisions  

a hundred percent. And it, it, it triggers an emotional response and it also, um, can incite fear. Now I, I’ve done some guest lecturing with a prominent university recently, and, uh, I sat in front of graduates from both electrical and software engineering. 

I talked about our own experience with AI and AI adoption, um, and the best use cases that we’ve had, and I’ve talked about this on the podcast before, around back office automation and where we were able to take some 280 hours of SOC analyst time out. Now those SOC analysts are still working for us. 

They’re gainfully employed and they’re doing higher value tasks that interest them more, but also benefit our customers as well. Right? And so an amazing outcome for us where everybody wins, our customers, our staff, you know, and we didn’t do the shameful thing, you know, like the Commonwealth Bank did whatsoever. 

It never ended our minds to do that either. Now, why do I bring that up? Software engineers put their hands up at the end of their end of my, my little talk, and they said, are we gonna have jobs in the future? And it was a sincere question.  

That’s a legit question coming from people investing their time, right. Studying and trying to get into the industry.  

Yep. And doing the right things right. Putting in the hard yards and at their end of their career, end of their tenure at the university, wanting to know what the future looks like for them. My answer was, yes, you’ll have a job, but your job may be different influenced by different tools, and maybe the expectation on output from you is gonna change. 

But you’re gonna be using some fantastic tech and capability. Your skills are absolutely necessary. You know, your knowledge of the way these things are fit together or put together, I should say, is just as important now as it was any other year. And so, um, there’s a way to, to embrace this stuff and to be optimistic and to lean into the tech and all that great learning that the universities bring to bear is still absolutely relevant. 

Yeah. A absolutely relevant. Coming back to that, I think it, it comes down to the, the premise mm-hmm. Behind implementing AI into a business. You know, the why. Yeah. Why are you really doing it? Yeah. Are you doing it just to cut costs and dressing it up as something else? Yeah. Are you doing it to actually improve your customer experience and improve your business? Uh, these, these questions need to be asked before you head down that road, and I think, uh. Combat missed a couple.  

Yeah. Well they did, right? And so, um, to any other business, thinking about doing the same thing. Yeah. Look out for the optics on that, right. And make a decision, you know, is it worth it? Is brand and reputation worth sacrificing way? They say the fish rots from the head.  

Yeah, that’s right.  

That’s what  kinda organisation do you want to be and yeah, computer society. Right? So let’s talk about skills and demand. Um, apparently Australia’s gonna be about 1.3 million tech professionals short by 2030. So if we go back to the MIT study, it was the AI survivability matrix. 

This is a fantastic little writeup by MIT, which I really liked. And it said that, um, vendor led projects. And we’ve talked about this before, are two times more likely to succeed than internal? I wonder how internal projects are even gonna get off the ground if there’s such a gap, uh, and lack of resource going forward with skills. 

I think this is only gonna help to serve, to compound what might be an inflated market expectation. And a good friend of mine sent me a stat the other day. Apparently 80% of all gains in the US stocks in 2025 were to do with AI 80% of stock gains.  

I think some people are calling it a bubble.  

Yes.  

Uh, what does that mean? 

So there’s been a few articles that I’ve noticed recently kicking around, just talking about that space in general, having Ponzi vibes. Yeah. You know, and, uh, look, it’s, it’s not, it’s not a, a Ponzi scheme by, by technical definition, but I think there is, um, I think there is some, some high risk. Yeah. You know, flags. 

Uh, in terms of the speculation around the value of, of this space? Yeah. Um, I think it is ahead of real demand. Mm-hmm. And real revenue coming from paying customers. Mm. Uh, that, that’s for sure. So it’s propped up, you know, like a lot of markets by, um, you know, other capital.  

Yes.  

 So that’s, uh, that’s interesting. 

What does that mean for consumers? I think it just means be be careful about how much you pay for this stuff.  

Yes.  

And, and where you deploy it.  

Yes.  

Um. Have, have a good premise. Understand your why, what you’re trying to achieve for your business. Mm-hmm. Keep an eye on your costs and, uh, be discerning about what tech you choose. 

Yeah. And what partner you trust. Mm-hmm. I think that’s really important.  

Think so too,  

uh, for people thinking about investing. I, I can’t give inve investor advice. No, we can’t do that. Not qualified for that, but it is, it is, uh, interesting. Well, I think for your research,  

I think that the market is expecting the next industrial age. 

Now we, you and I can sit here and say that. There are gains, there is return on investment. We’ve both got examples within our own businesses where we’ve seen return on investment. Absolutely. Investment. Absolutely. Absolutely. So there’s no question about that. Um, I know that our costs were considerably smaller than what the market would need in order to prop up the investment. 

You know, and I’ve talked to you about this before, within the, our own security business, we’re not spending much on the technology at all to get. Absolutely astonishing returns.  

Yeah.  

If every example was like mine, well then I think there’s many investors out there that have spent too much money, but they’re spending it primarily on this hope of artificial general intelligence. 

And are we actually getting closer to it now? I wonder what’s gonna happen first. Well, firstly, actually the question is a binary question. Will we land there at all? Some may argue we’ve got domain specific AGI right now. Right. Do we right or do we have statistically relevant answers that would make us think that we do? 

Like what would be anyone’s definition of AGI, you know, we joked about that with Steven Worrall. Somebody’s got a definition of it. Um, I don’t know, but I wonder whether it’s diminishing returns now. Um, there was expectation that ChatGPT 5 would be more than it is. Um, how much more money do you have to spend to add IQ points? 

And is this a. Is this a target that’s realistic? Are we actually gonna arrive at a GI, yes or no? And if the answer’s no, there’s too much money in this business.  

Yeah, I, I, I think, uh, it remains to be seen. Mm-hmm. You know, there’s so many differing opinions. I think the domain expertise angle is now a new angle that’s coming out. 

You know, we we’re not at AGI, but for this particular, uh, use case. You know, you, you could argue that we are. So that seems to be a new thing. Um, it feels weak to me. Yeah. It feel, well, it feels like the bar is continually being raised or moved or, yes. You know, goalposts are, are moving around. So, uh, I’m like you Naran. 

I’m not sure if we’ll ever get there or not. Mm. But a, a certainly a very interesting space and a lot more mainstream press outside of the industry talking about, uh, investments in AI. Yes. Returns. In the context of, uh, bubbles. Yes. Ponzi. So watch this space. Mm-hmm. Very interesting to be a part of and to commentate on. 

Yep. Fascinating. Um, and choose a partner you can trust for sure. Choose a partner  

you can trust. Trust, absolutely. Um, with some runs on the board, there are out there, now we have them. You have them, right? So choose a partner who’s done it, who can demonstrate it and show you that, hey, listen, here’s what it costs, here’s the return, and here’s where we think particularly if it’s a back office. 

Uh, automation style project. I mean, you can quantify the gains there, right? It’s not too hard to do it. 280 hours of SOC analyst time. I know what that’s worth.  

That’s right. Right. Well, there you go. We’ve got the, a few big topics there. Um, quite interesting to get into, you know, uh, hyperscale spend, uh, taxes being paid in Australia, the local industry contributing, you know, that’s all going on. 

Uh, so real, I’m really gonna keep a keen eye on, on some of these big deals. Yes. That are happening there. Um, and if anybody’s got any information, uh, feel free to ping us.  

Yep. And we’re after a balance view here too, by the way. Absolutely. It’s all right for you and I to have opinions. Yeah, that’s, that’s right. 

We’re nothing if we’re not open-minded. If somebody vehemently disagrees and they wanna come on the podcast and pitch to us a different angle that we are not aware of, perhaps we are ignorant. Who’s to say We’re open-minded and we are prepared to learn.  

Definitely, definitely. And definitely, um, you know, anything’s possible. 

So open to hear it all. Um, you know, and CommBank uh, you know, people are gonna make mistakes Yep. And, and get it wrong. But I think it’s just that, that one’s a, just a good, a good lesson in just, you know, just pausing and thinking and, and, and doing the right thing. Mm. This tech is really powerful. Uh, it can change a business for the better. 

Um. But, uh, but be careful. Be careful. Things can be a lasting memory. Yeah. And be good to people, business to is just people doing stuff with people at the end of the day. That’s right.  

We’re all human beings, aren’t we? We’re supposed to be looking after the human race and That’s right. And looking after each other. 

Yeah, absolutely. Uh, and, you know, uh, big tech AI ecosystems, how the money flows. Uh, I’m not an expert, but it’s definitely topical. Mm-hmm. Um. What, watch this space. I know in my own experience, uh, in Australia specifically, I’m not seeing the demand from consumers there to support go going and, and making big investments ahead of that demand. 

In, in GPU no, there’s, uh, there’s a bit of private equity in VC money in the space doing that. Yes. I wish them well. Um, I think they’re betting ahead of the demand. I think it will come. The question is how long can you hang on?  

Yeah, that’s right. Will, will it happen in time? Interesting.  

I think lots of people are asking that question. 

John, a lot of bank is in tech. 

I think our, our gears have been ground.  

They have been sufficiently ground. Sufficiently ground. Excellent.  

Good segment. Uh, lovely, enjoyable.  

Right o. That’s a wrap. This is Things. Reasons. And it’s another episode recorded. Uh, I enjoyed that. Thanks very much. Now, for our audience, you know where to find your podcast. 

It’s wherever you get ’em, right? It’s YouTube. It’s Spotify. It’s Apple Podcast, absolutely everywhere. We are, of course, produced by Pru Loon. Our digital producer is Karina Aguilera, and you can get in touch at Pru. That’s pru@thingsreasons.show. All feedback, all suggestions, questions, anything you like, we will reply because we are reasonable, decent people. 

That’s a wrap. Thank you very much.  

Thanks very much. See you again next time.