When Bubbles Become Clouds

Richard Kramer:

Welcome back to Bubble Trouble conversations between, well, we're not deep fakes masquerading as humans, well, we are that real life double act of the independent analyst, Richard Kramer, that's me and the Economist and author Will Page, that's him. We're closing in fast on our hundredth episode of exposing the sycophants and stenographers who create the bubbles and get our portfolios into trouble. And as we always like to say, if there's a bubble that burst, we usually are the ones that pricked it first.

We're going to tap into this topic that we haven't really spent a lot of time with to date. The excitement around the big cloud service providers like Microsoft, Google, and of course the market leader, Amazon's AWS. Are we getting ahead of our skis about what's another form of storage or a place to compute stuff or is this truly a game changing transformation for organizations that changes everything that they do? We're joined by a true expert, Liam Maxwell, who's understood IT from inside government, which is a very hard thing to understand to begin with, to how governments function and from outside as a senior person at a AWS. We're going to travel from the clouds above Amazon headquarters in Seattle to rather different clouds above the Ukraine where Liam played a pivotal role in securing their government's data and future, more in a moment.

Will Page:

Well, we're blessed by tech giants today in that we are here at the World-Class studios of Platoon, an Apple owned company, and we are joined by a key part of the Amazon machine, AWS and Liam, to bring you to the microphone and let you introduce yourself. We usually ask for a tweet like introduction, which puts a limit on the amount of words, but your career is so incredible. We'll come to the Ukraine in a second, but more importantly, why on earth did you go into Whitehall government offices with a suitcase full of fake bank notes?

Liam Maxwell:

Thanks Will, that's a really nice introduction. So I'm a technologist. I worked as a CTO in business and then was a teacher, and then I started to get involved in local politics and got elected and we put together what was the test lab for David Cameron's opposition at the time. And then when they came into government, I was asked to come and join the government, as an official, and became the CTO for the British government. And really working with a brilliant team there, we put together changes to the way that the citizens interacts with the state, a fantastic team led by Tomas Moore and Mike Bracken doing that, but also the underpinnings of that, what that was running on and giving government the ability to be flexible and move off 25 years of building technology in a particular way so that they could be flexible, so that they could get more secure and they could also build brilliant services for much, much less money.

And so that combination of being able to operate at scale, operate at speed, and with security, but also make savings was a brilliant way of helping that move. And it was really good fun doing that within the government. And yeah, we had some ways of explaining that-

Will Page:

Which were controversial.

Liam Maxwell:

But we did actually build a, I think a team which was copied, well, or copied or imitated worldwide. And Amazon came to me in 2018 and said, "Look, we'd really like you to help customers, our government leaders accelerate their modernization and transformation programs." And they've given me the ability to go and do that. And that's what I do now for a living is help operational leaders and governments accelerate their modernization programs.

Richard Kramer:

I want to give a shout-out to Mike Bracken quickly because he had this brilliantly simple idea, which was why don't we make all the government websites built on the same template so they all look the same and when you navigate them, they all feel the same?

Will Page:

Oh yeah.

Richard Kramer:

And it was not only a tremendous cost savings, but the benefit when you go on to multiple websites from the government and they all basically have the same app and the same look and feel. It's just such a blue flame, simple idea, and it's not radical at all. It got rid of dozens of projects of Frankenstein code.

Liam Maxwell:

Well, Richard Pope and Ben Tarrant were with two guys who within that space actually, the conversation that was around was you as a citizen, why do you need to know where family tax credits is and which website you need to go to? There's the government, and the only time actually people want to know which bit of the government is delivering their thing is when they want to make a complaint, when they want to go to government, they just want to go, "I want to go to the government to find out what the thing is and how it works." And it should be simple for people, but it also is a brilliantly simple idea. It did mean there was a lot of cost savings. So for one example, there were 125 million telephone calls coming into the government every year, and each call is two or three, it's 12 or 15 pounds. I mean, it's a lot of money spent, but a quarter of those calls were saying, "Where is this on the website?" So if you can make the website clearer, you save a lot of people's time, a lot of people's energy. And the thing is you can focus the taxpayers' money on building brilliant services and delivering frontline services, which ultimately is why people have a government.

Will Page:

I love it. I love it. It's pioneering. And to your credit, it's been copied by many other countries since. So just to stress the quality of the guests we have, this is something which has gone global, but I can't lay you off the hook. Give the people having dinner parties this weekend a story to tell about that suitcase of fake money.

Liam Maxwell:

So it wasn't actually fake money, but we would go to a meeting with members of people who were trying to build systems and they would explain quite how much it was going to cost, and sometimes the numbers were really eye watering. And so one day I remember just having a bag and just going, "Look, let's work out how much, what does that actually look like? If you're going to spend 60 million pounds to build a particular thing, let's actually see, what's 60 million pounds looks like? And can we use that to help identify that it's smaller?" In fact, in that case, we got it down to 12 just by putting it on the table in front of people. So sometimes stunts work, sometimes they don't. But it was paper the size of bank notes, it was white paper the size of bank notes.

Will Page:

So you'll be canceled by the Forestry Commission?

Liam Maxwell:

Yeah, I'm sorry.

Will Page:

Okay. Right. Now, the key to a great podcast is to assume no prior knowledge for our audience's benefit. Let's get everyone on the same 20 yard line. This conversation is going to go deep, but let's stay shallow for the moment and put a definition of what cloud is. Imagine a company, an institution and it has to store lots of stuff, and that stuff is stored in perhaps local data centers, and then that company says, "Hold on, there's this thing called the cloud. I'm going to shove that stuff up there. What actually happens on the technical backend and what does the frontline office worker actually see as a result?

Liam Maxwell:

And the big thing about cloud is it's really not just all about storage, but so think about it, just in that storage case, it's much simpler, if everyone's doing the same thing and commonly storing data on their local systems, you've got to do the backups, you've got to do the resilience, it's got to fail failover to places. Every single company would be doing exactly the same thing. So what cloud does is it gives you the opportunity to have capabilities as a service so that you can get a storage capability or a compute capability or a database capability, and you don't have to go and buy the server, set up the servers, put them in a nice room, all of that stuff. It's all there for you. And so one of the ways we describe it is we do the sort of undifferentiated heavy lifting of IT so that all of the basics can be done by the cloud provider and then the business can turn up and instead of spending huge amounts of time maintaining the infrastructure and making everything work, they can get on with the brilliant stuff of designing fantastic apps that really work for their users.

And they can spend all of their time and all of their money on that. And the infrastructure part of it, or the databases or the compute or the AI or all of those things are all building blocks on which they've built this.

Will Page:

So your beautifully illustrating for me, the productivity gain. Just at street level, I could spend 80% of my time worrying about backend infrastructure and 20% scaling my business. You want me to spend 100% of my time scaling my business, and AWS takes care of the backend infrastructure.

Liam Maxwell:

That's right.

Will Page:

Perfect.

Liam Maxwell:

That's right.

Will Page:

Second question is we kind of got a visualization here of what the cloud is. How do you pay for it?

Liam Maxwell:

It's pay as you go. So literally you pay for the services that you use and when you use them, and some of our billing is second, sub-second billing, because you're only using things for, sometimes you're only using things for a very short period of time.

Will Page:

There's no breakage. There's none of these onerous IT contracts where you're paying for stuff you don't need.

Liam Maxwell:

Because yeah, no, none of that because you don't need to pay CapEx, you don't need to pay capital expenditure. You go and you rent the services that you need when you need them.

Richard Kramer:

I want to come back to two dreadful letters you put together previously, which was IT. Now when you and I started out here, IT, three of two or three, four decades ago, IT was pure nerd porn, right? It was pocket protectors, thick glasses and body odor, right? It was that the guys in the compute lab at the university who were programming in Cobol or FORTRAN or what have you or basic, and then it became a sexy thing for Tech Bros. But one constant to my mind in this IT world has been the software companies promising that the next piece of software, the next release of software would solve all your problems. How is it that when our listeners think about the frustrations they've had with software generally, with IT generally, how do they guard against the fact that the software industry has forever been creating these bubbles of promiseware, of vaporware, of the next version, we'll take care of all of our problems. And we've all had these regular frustrations. Is the cloud going to take that away or is this just a feature of the tech market that we are, as you might say, if you're a ham supporter forever blowing bubbles?

Liam Maxwell:

Perhaps could I explain? Because we don't think of it that way. And perhaps if I could just explain the way that we sort of think about innovation is that the innovation that we work with is we work with our customers on their innovation to help produce the components that we do. So about 90% of the innovations that come through a AWS are things which our customers have come to us and said, "Look, I've got this problem and it's a common problem. Can you help us build it?" And so that's where we get to be able to build things because we work backwards from the problems with our customers to try and help build things. The other side of that is that we then are always in a position where we make an announcement that it's something we've worked on, we've built, we've put together, it's working, the customer's happy with it, we get to that stage and then we make an announcement about it and say, "Look, this thing, it's a common component, which is a common capability which you can have as a service, this customer, this customer, this customer already using it, you can now have it too."

And that approach means it would answer your particular complaint of that. But that's actually part of the way that we always think is that if you can work backwards from the customer, work backwards from the user, that builds up.

Richard Kramer:

And those customers don't mind sharing their clever ideas with someone else?

Liam Maxwell:

We find because these are common core components that people have, and yes, with customers, but can I also just say one of the big things that we do with governments is that governments don't compete. I mean, I work in the government space and governments don't compete. The driving service in Norway don't have a driving license office in Sweden because they don't need to. And yet they do exactly the same thing. And so quite a lot of the work that I do is about helping governments identify that someone else is already doing something like that and that moves-

Richard Kramer:

And learn from best practice, yeah.

Liam Maxwell:

And going back to Will's story about Tarot and Richard Pope's govUK, govUK the website of the government of New Zealand looks quite a lot like the website of the British government's website. Do you know why? Because it is, because they open sourced it.

Will Page:

Just very quickly-

Liam Maxwell:

Brilliant people.

Will Page:

When you describe the innovation happening with a AWS and off AWS, can you just give a crystal clear case study example of a company or an institution that went to you saying, "I want to store all this stuff," and six months, 12 months later said, "Now I want to do all this stuff." The transition between storing and innovating.

Liam Maxwell:

Yeah, I mean, and storage is one of the first services is to get storing. So there's a restaurant delivery group that I remember advising in the states, and they wanted to come to us with an idea around, they wanted to deliver things, but they kept on trying to deliver things and they had to use a map and they had to pay for the map. And sometimes the map in a particular market was this map and it worked better and sometimes it was in another and they had to continually cut the changes to make that work. So there was an example there where the team that was working with them built a geolocation service, which meant that you could choose which map you were going to use to help yourself move forward. That that's one example.

Will Page:

Better maps, better delivery, better demand, better-

Liam Maxwell:

Yeah. But also if you start thinking around the ability to have a moving, actually from massive storage to being able to get a landing zone or a service, it was called a service workbench, which was just a starter kit that lots of universities have come to us to say, "Look, we store all this stuff with you, but we want to be able to move faster by having all the things we do at the beginning of a project. Let's just speed that up." And so you can get the templatizing, the work that, the routines you have because you've done it all before. So why start from scratch, spin the thing up, just have a kit to get yourself going.

Will Page:

Making academics more efficient?

Liam Maxwell:

Yeah.

Will Page:

You should deserve the Nobel Prize,

Richard Kramer:

Although he did just use the word templatizing, which I think most other guests we would pretty much show them-

Liam Maxwell:

Yeah, I'm sorry about that. Yeah.

Richard Kramer:

But now I'd like to come back to again, this sort of culture of IT, because I was told long ago, three things you cannot tell a man how to do. One is drive a car, two is make love to his or her partner, however you want to say that. Or three is buy tech hardware, right? And you have these IT directors or CIOs for years would love to go and pet their mainframe computers or their servers, close the door, turn out the lights and have this big smile on their face. And we've seen these cycles in the industry of going from distributed computing to centralized computing to distributed computing and centralized. And I guess my question is, when a CIO who sees all their peers going to the cloud, what is going to stop them or what doesn't encourage them to say, "I'm going to be different because I'm going to have my own stuff and I want to have my own stuff and build it my own way." And we see some of the biggest companies do that because they're sophisticated, but how do we know this isn't a pendulum that will swing back the other way when computing gets cheap and personal enough that we don't necessarily need to have the same solution as everybody else?

Will Page:

Richard, just to preempt to Liam's answer here, you're reminding me of the famous James Barksdale quote, which is, "Ladies and gentlemen, the future of business is either bundling or unbundling."

Liam Maxwell:

And you have seen that the pendulum swing backwards and forwards as you went from mainframe to client server to centralized, in and out, and you have seen a lot of that. I think that the reason was, I think, the pendulum is going to swing and stay swung one way for a long period of time, is AI. But can I just go back as to why we think that? So cloud is not just storage, it's the ability to access hundreds of capabilities. We have services, we've got a really broad and deep range of services which we offer, and those services are available on a pay as you go basis. So you pay for what you use, and that means that the flexibility for the way people pay for the services that they consume becomes very much a, you are only paying for what you use. So the leakage isn't there, the wastage isn't there.

It's much more thrifty to do it that way. It also means that you can scale, that you can actually move things very, very quickly. So if you are running your own data center and your own things and you then go, "Actually, we now want to move into the European market and we want to start building more stuff in Germany, we want to get this," say you're doing movies and you want to get close to your customer and you want to deliver those things or some engineering routines, that, you have to go and build a whole new set of things in Germany. Well actually within the cloud model that we have, you can go and deploy your services there. You can expand easily using the same code, the same materials, same structures into those areas. So that helps you scale really well.

Will Page:

You can hit the ground running.

Liam Maxwell:

Yeah, the speed means that you can do that relatively quickly. But the other really big component is the security. If I give you the example, I inherited a government estate when I started as CTO that was, I mean, there's a brilliant guy called James Duncan working with me. And we tried to work out where's all the servers? And we worked over, we had something like 350 data centers around the country, which were running government tech. And I had no way of checking that everything was patched, patched, I mean, let alone whether the security guards' dogs were being fed properly so that they would be out there at night protecting it. So none of those things that around the core infrastructure recorded in place. And so my security posture, I just improved it almost instantly by moving to the cloud where the security is the same level of security as the most demanding customers and moving that helped us increase the security posture whilst, I'll say this again, whilst the cost for doing it reduced markedly. So it was a really big change. So speed scanner security means that's why you swing into the cloud. Why does it stay there as if you are offering the ability? And what we are sort of doing is democratizing access to brilliant services because a small company or a large government or a large company can access these cloud services on exactly the same terms.

You are then able to offer people capabilities that they'd never been able to build before because they can do it as a service. And so people can have access to foundation models and the huge amounts of data that you need in order to put together a really effective and efficient AI on a credit card. And if you do that the old traditional way, that's billions of pounds of investment and billions and lots and lots of years and lots of people putting it together. Whereas cloud actually gives you the ability to access AI on a credit card. And that's why, I mean, I've got a sticker on the back of my phone that's no cloud, no AI, it's you need it in order to do it. And so whilst that continues, that's going to be in place I can see for the foreseeable.

Will Page:

Now you've mentioned the importance of the ability to move fast, to hit the ground running, to scale solutions in an incredibly short space of time, a short space of time that's inconceivable in the age of local data storage, and that's in times of peace. So I think we should now, Richard, turn our attention to war. But before you do, Richard, we should just give a quick hat tip to Feargal Sharkey and we'll link to this in the show notes, whose speech, "And Have I Got News for You" last week pretty much had the entire audience on our standard BBC comedy show in tears referring to the importance of peace. And with that backdrop, given last time we were in this amazing studio, it was with Feargal Sharkey, we should turn our attention to events in Ukraine.

Richard Kramer:

Yeah, I mean Liam, you had told us coming into this about this incredible story when the Ukraine war began, about having to effectively lift and shift, as you would say in the IT world, their entire government and make sure it was backed up somewhere that wasn't likely to be under a bombardment. And while putting all your data in one place does in some cases make it more vulnerable, if that one place is specifically vulnerable, you've got to distribute that data pretty quickly. So do you want to walk us through what you did with a AWS as a government focused person to help the Ukraine government make sure that it wouldn't lose all that valuable data for running itself whilst it was under attack?

Liam Maxwell:

Yeah, so I mean, very early on the Ukraine suffered some strikes on their data centers, but they had some luck because Ukraine's got a brilliant STEM education system. They've got lots of really brilliant engineers, absolutely lots of fantastic engineering and computer science background. And their digital minister is a guy called Mykhailo Fedorov, and he's inspirational and brilliant. And I met him the year before the war, he'd come to London to talk about, and he'd gone to GDS to go and talk to them. And he came in and had a chat with me and we signed a sort of, let's get together MOU. And then we realized between then and the beginning of the war, we started having conversations with them about this, "You need to think about continuity of government. It's going to be really important." The president made sure that a bill was passed, which meant that data could be stored outside the country if it needed to be. And then-

Richard Kramer:

We're talking sensitive data like health records where there would be a concern if they were stored elsewhere?

Liam Maxwell:

Yeah, I mean, people should be concerned where their data is stored, whatever. And then so on the day of the invasion by happenstance, I booked in to go and see a guy called Vadym Prystaiko who was the ambassador in London who actually was a computer scientist by trade. And we sat down and had this strange lunch where the invasion had just happened, just started. And he was serving up this brilliant borscht, and I was sitting there, I go, "So I think we probably need to work out how we can help you." And he said, "Look, the first thing you've got to do is we've got to make sure we've got continuity."

And so I said, "Well, of what? So let's work it out. What's the most important thing?" And he just said, "Population register. First thing, population register. Second thing, land register." Now, those two, because the population register, the first thing the Russians are going to do when they invade is they're going to go, "Oh, why don't we have a referendum? Oh, we'll keep the listing." And it's the classic Stalin quote, it doesn't matter where the votes are, it's who counts them. And you've got to make sure the list is right. You've got to make sure that the land register is there so the land doesn't get stolen and taken away.

Will Page:

I have to say those two boxes you had to tick, they just visually bring the horrors of war to life. It makes you rethink. You see tanks on TV, but no, no, no, no. There's something much, much bigger at stake.

Liam Maxwell:

It's about the people.

Richard Kramer:

Well, and equally there is an entire theater of war, which is in the cyber world, which is constant and ongoing. And in this case, in the Russia Ukraine case, is ongoing every day. We don't hear about it the way you might hear about bombs and invasions and counterstrikes or counteroffensives. But this cyber attack mentality is relentless.

Liam Maxwell:

And we and lots of other tech companies did huge amounts of work helping the Ukrainians protect. But in terms of the data, so we had the population register, land register, then we wrote down tax register, criminal records, healthcare, education, welfare. And in a way, there you go, you've got a state. With those registries, you've got a state, you've got the ability to start to move. So-

Will Page:

You have a country.

Liam Maxwell:

So we then said, "Okay, so we've got to back those bits up. How do we do it?" So I went and got these device, explained to ambassador what they were, which is a device called a Snowball. So a Snowball, imagine a 90 terabyte US USB stick. Okay? It's about the size of a suitcase and there's edge capability on it for communications, but it's a classic storage device. Okay, so you put it, but you can run compute on it.

Richard Kramer:

Giant data suitcase.

Will Page:

I've lost USB sticks in pubs before. I'm not going to lose this one.

Liam Maxwell:

Well, that's it. So they're about the size of a suitcase. And we took three of those and a colleague of mine, they took it, flew them through Amsterdam onto Krakow and handed over the Snowballs to a member of the Ukrainian team who was a truck in a people wagon. And they stuck the suitcases in the back and off they went into Ukraine. And I've got to say, he said to me at the time, he said, :"That's the most nervous I've ever been at a baggage claim," because these things go in the hold of an aircraft. I mean, they're really rugged. And so they took the snowballs, off they went, took the snowballs, and on the Monday, we got these calls coming back in which were, "I've got this thing, what do I do?" And so we had this brilliant set of, and everyone volunteered in the company volunteered to do it, and they set up a Ukrainian center called Project Sunflower, which we did over that weekend.

And the Sunflower team then started helping the Ukrainians boot the devices up, plug them in, make sure the manual worked for them because each of them, I mean the manual is a kindle stuck to them, and then you...

Will Page:

Brand.

Liam Maxwell:

Yeah but simple thing to have. And then translating the instructions from English into Ukrainian, Ukrainian into English, thank you, Google Translate. And then they very, very quickly got that moving using encrypted messaging service. And they then started to back up. And at the same time, there were the banks and there were the universities. So we're going, "What do we do? How do we save our stuff if we get bombed?" And there's a brilliant guy in one of the banks, Privat Bank, which was, it's a bank of about 18 million people in Ukraine banked with Privat banks. So very big. And they backed up and migrated completely. And a migration in normal times takes a few months and it's very, very complex. Got going. This guy led team, this is the customer, 45 days migrated, I think it's two and a half thousand servers, 350 applications into the cloud in 45 days, two terabytes of data.

Richard Kramer:

And if we rewind that quite simply now, the Ukrainian government can run all of its technology backend without any of it being in Ukraine.

Liam Maxwell:

Yes. And some of it, it still runs some services inside Ukraine, but the really important thing there is that as any Govtech nerd will go, "Brilliant, they had the registries. The registries were." But some things are still running. And Misha's been really clear about this in public. He said, "Look, DIA," which is their super app, which is a brilliant device, means that everyone can use it to report. So to do their welfare benefits, their education, they've even built services now where you can tap in, there's a Russian group around the corner from me here, so it's the citizen app. DIA runs in Ukraine in an undisclosed location or several undisclosed locations, but if it fails, if they get taken out, it fails over onto AWS.

Richard Kramer:

Right. Somewhere outside of Ukraine.

Liam Maxwell:

Somewhere outside Ukraine.

Richard Kramer:

In the rest of the world, not to narrow it down.

Liam Maxwell:

Oh yeah, no. And one I can't, no, I'll be really honest. I can't narrow it down. I don't know where it is. And there's no need for me to know where it is. I know it's in the cloud and it's being looked after properly and it's in the right place.

Richard Kramer:

Well, that is a fascinating and inspirational story of how the cloud might work under the direst of circumstances, which we don't really want to think about too deeply, you want to...

Will Page:

I just want to, just reflecting on that incredible story. Big tech often gets bashed, and I often defend big tech by saying, "Don't compete on cost. They compete on convenience." Rather than what the textbook says that you do. They expand output, and in the case of the cloud, you've expanded output and they eliminate costs. In the case of the cloud, it's a lot cheaper than data storage. They also contribute to solving conflict. Do you feel this Ukraine story needs to be wider told? I mean, it does feel like it's been, it's not a page one story yet, but I wouldn't be surprised if I go to the I IMDb website and see Liam Maxwell in production, the Ukraine a AWS story.

Liam Maxwell:

Well it won't be me because it's a brilliant, there was 600 people on the team doing it. I mean brilliant load of volunteers all the way across the company who just donated their time to do it.

Will Page:

That's important to hear. All hands on deck.

Liam Maxwell:

And if I just explain, yeah, the way of doing it was as it happened, we have two ways of starting any program, which is you say, "This is chart so this is the mission we're going to do." And then you set out a set of tenets which are, "This is how we're going to do it." And a brilliant woman, Maggie Carter, who I worked with during COVID when we did a similar thing for a diagnostic development initiative, we sat down that weekend and said, "Right, here are the tenets, this is the mission, this is what we do." We announced it to the teams across the piece. People picked that up. The mechanisms were in place from her brilliant disaster recovery team. And that meant that you could just get, the ministries could say, come to us and say, "Look, I've got these workloads. How can you do it?"

Strip straight through. So long as they met the tenets, they could move fast. And that's the weird bit about working in Amazon, is that the ability to make a decision is so much quicker than anywhere else I've worked because we have a real concept of two-way doors, that if a decision is a two-way door, if you can go through it and come back then if you're 70% sure of where you want to go, take it, take the decision. And that's how we devolve power right down to everybody within the company. And the way you take that decision is you look at the team you're working on, what are the tenants of the team? 70% certain I'm okay on this, bang, off we go. And that means you can go fast, and that also means that you can really get ahead of big, big challenges. And that's not just getting ahead of an invasion, but it's getting ahead of a disease as it was traveling around the world.

Richard Kramer:

Well, we need to wrap up part one with Liam Maxwell. I would only say one of the shocking things I pointed out to Will very early on in our Bubble Trouble podcast was that the 22 billion pounds spent by the UK government on their test and trace app, which didn't work, is the equivalent of the entire annual R&D budget of Google that year. And you just wish that you had brought in some competent people to write that app in a weekend as opposed to allow the UK government to spend months making a hash of it. We'll leave it at that for part one and come back in a moment with more with Liam Maxwell. Welcome back to part two of Bubble Trouble with the estimable and very experienced Liam Maxwell from AWS, formerly of her or his Majesty's government of IT fame. Will, you wanted to discuss the cloud and GEP?

Will Page:

Yeah, I mean you've laid out beautifully in part one just how transformative this tech is, and that's not just for the private sector, but the public sector as well. And I have this bug in my base bin, which is, does the government actually capture what you do? The government captures what agriculture and fisheries do. The government captures what construction does. They look at cranes, the government captures what a car plant does, they produce cars, but I want to explore down a rabbit hole with you. Does it even capture the cloud at all with three...

Liam Maxwell:

It's one of the things where, because cloud delivers a huge amount of productivity because instead of everybody running and building and raking their own servers and putting their things together and spending time spinning up an AR routine, et cetera, et cetera, and learning how to do it, it's lots of this stuff. The services are there for you to use on a pay as you go basis. And there was some brilliant work done by some of my colleagues in the UK, which we looked at, which was how much value does the cloud actually bring to government? The thing was that, the thing we discovered was that we found the economic value of leveraging cloud was several billion of gross value add to the economy. In my head, I remember it coming out as the same value as the Premier League to the economy. That's just the AWS bit of the cloud.

Will Page:

What you do on top of that, the innovations side, well, that produces gross value added, which is going to be captured elsewhere, but traces back to the cloud as its instigator.

Liam Maxwell:

Yeah. And it's one of the things that I genuinely don't know whether the RNS model or capture this, but the thing is the productivity gains that people can see and the ability to again, just do things so much quicker. So go from an idea to a brilliant app where all you need to do is work on the app and not on the infrastructure and the communications, all of that stuff because you can just get that on a pay as you go basis. That's priceless because again, if you look around in the competitive market, everyone else is doing that as well. So you need to go that fast.

Will Page:

Well, to go a little bit deeper in this topic because I think we can uncover something quite crucial here, which is this transformative technology is invisible in government accounts. That's my assertion. I just want to probe it further, which is let's look at price, let's look at CapEx and let's look at location. On price, I published some work this week, which showed in 2001 you gave 9.99 for Rhapsody, a music streaming service, and you got 15,000 tracks. Today you give 9.99 and you get 110 million tracks, you're getting a whole lot more for your money. That's hard for government to measure. The unit cost is the same, but the output is so much more different. Without getting too technical here, surely that same dilemma has to exist with cloud. If I spent 100,000 on cloud services, let's say 10 years ago, I would've got X. If I pay you 100,000 today, I get Y.

Liam Maxwell:

And also you get a much broader set of services that you were using because instead of, this idea around common infrastructure, which started originally with storage, compute databases, but then built services, services, services. So as you went up the tech stack, more and more of the components of the tech stack were becoming available as a common service because the cloud enables you to do that. I mean, even if you think about the capabilities that are open to us to leverage things like Lex and the voice to text, translator, transcribe, those sort of AI services which are now AI services that are, just use it as a service and you can get that from the cloud on a credit card. And so it's much more efficient and effective. Coming back to you though, in the olden days, people would've bent lots of GDP and huge amounts of money building those routines and putting those things together themselves. And somebody three streets down would be doing exactly the same thing and three streets down, they'd be doing exactly the same thing. And now it's just a common service. And so people who want speed scan scanning, security go, "Right, well, I'll just go and use a credit card and get it as a service."

Will Page:

And it's not just a contribution to the economy that's getting lost. It could also be inflation. I mean, let's just park up for a second and think about the smartphone. I could have a printer in the CPI inflation basket. I could have a scanner in the CPI inflation basket and I could have a telephone. Now I've got a smartphone. I don't need those other two items, but it could still be there. And it's the same thing with the cloud in terms of what you're able to achieve with that expenditure is so much more and it's displacing activity that would've happened elsewhere.

Liam Maxwell:

Yeah, I mean it means that people are able to invent. In a way, the best thing about this is that it enables people to invent more quickly, more effectively, and also their chances of success are better because they're not investing all that mind share and bandwidth in keeping the infrastructure going.

Will Page:

Quick, cash prize Scottish question for you. This is for a Scottish 20 pound note. How much GVA does a smartphone add to the US economy?

Liam Maxwell:

No idea.

Will Page:

A big fat zero because international accounting means you have to attribute it to Taiwan and South Korea.

Liam Maxwell:

Really?

Will Page:

Interesting.

Liam Maxwell:

Right. Okay.

Will Page:

Question two is to take economic theory just real quick here, the argument goes, you expand capital expenditure and you boost productivity. In part one of the show, you said you're not doing CapEx anymore, you're not having local data servers, you're leasing, that's current expenditure. How does the economic theorist figure this out? Because now I'm reducing capital expenditure, but I'm absolutely smashing productivity through the ceiling.

Liam Maxwell:

I'm sort of stuck with the question. What-

Will Page:

I think the ONS are stuck in this question too because you overturned the rule book.

Liam Maxwell:

Yeah, I think it is a new path. Using and leveraging common capabilities is something that when we were making things and in the manufacturing industry, it's very difficult to do, but in the technology industry it's where the speed and the scale is. I suppose the other change is, just to give you a view of how different it is. When I did a startup in 2000 and we raised, had to go and borrow 4 million quid and we went and bought a load of servers and we built a development network, a test network, and a production network. And this was running an online recruitment company and it was in Hoban and they were all there and you had to back up into different places, et cetera. Now all of those things, gone, all of that is just, I'll just put it on a credit card.

Richard Kramer:

But look, I've got to derail Will's nerd porn about statistics and ONS right now. But what explains the productivity paradox, the fact that if you look at productivity in many western societies, it doesn't seem to have moved greatly even though we've got all these tools available to us. What is holding it back? Because you can see the capabilities which are now at our fingertips. When I came to London 30 years ago, of course everybody had an A to Z. Today, it's like a historical artifact. We all have maps on our smartphones. Where is the productivity paradox disappearing into that we don't get to see it in the economic statistics, in GDP growth in our lives that we're not showing that productivity?

Will Page:

I want to answer this for you before I hand to my guest with a simple sentence, which is what matters most to our lives is being measured least by our statisticians. And this is what I'm thrusting at. I just don't think what you're doing to society is being captured by our governments.

Liam Maxwell:

Or any government because where is there a government around the world that does capture that and this enormous change? And I suppose part of it is to say that, I mean, I'll be honest and say that my focus isn't on that. My focus is on actually building and delivering brilliant services to government leaders so that they can go faster to build better public services. So less taxes used in administration and more taxes used in frontline service. That's what we, we use the cloud to do that. And those are the numbers that really affect me.

Will Page:

Just to wrap it up and put a pin in it, I think I do have to raise a Victorian concept with you and see whether it resonates in terms of where this whole thing is like a square peg trying to fit inside a round hole called Jevons paradox, which is during the coal era of the Victorian times Walter Jevons said, "Why is it that when we become more coal efficient, we end up using more coal?" My favorite one, it's not that sexy, it's to look at refrigerators. When refrigerators become more energy efficient, what do we do? We buy even bigger refrigerators or if cars become more fuel efficient, we live further away from our offices and we spend more time in our cars. Is there a Jevons paradox thing here which happens with the cloud, which is as I become more cloud efficient, I use more of the cloud, and that just doesn't sit.

Liam Maxwell:

And there's also another incentive for doing that because you see, the more that you start using cloud, I mean if you start, and I can only speak for the numbers of the company I work for, but if you start using cloud, our research is that you use 80% less carbon, that your carbon footprint goes down by 80% the more you move into the cloud. So there's actually a really good benefit because you're using much, much less energy. You're pooling it. You're not needing to keep all those data centers cool, all the rest of it because someone else is doing all that for you in a strong and resilient way. But it also means that you get much, people use and generate much less electricity. And that means there's much less impact on the environment from IT. I mean the impact of large numbers of low efficiency data centers, the old model is quite enormous in terms of just the carbon that they're using up. And so the cloud actually gives you a route to much lower carbon.

Richard Kramer:

And it is astonishing, this relentless march of IT expenditures broadly defined as a percentage of GDP, which keeps going up and up and up even when the individual costs are per unit of compute capacity or storage or any of the elements of it keeps going down, down, down. And that is obviously reflecting the tech takes a bigger role in our lives all the time.

Liam Maxwell:

And we have many more tech jobs. So one of the big things that we, one of the numbers when I was in government, I was to look at the digital economy. I actually was responsible for growing the digital economy. It was part of my job spec that my brilliant boss-

Will Page:

Doing something that can't be measured?

Liam Maxwell:

You can't measure it.

Richard Kramer:

Isn't it best to stay out of the way if you want to grow the digital economy?

Liam Maxwell:

Sorry.

Richard Kramer:

Isn't it best to just stay out of the way?

Liam Maxwell:

Well, in many ways, yes, that was part of it and you can measure it, but one of the things we also measured and came back was the wages of people who worked in the digital economy were about 40% higher than the wages in the rest of the economy. And so there was a definite incentive to start to help people move there because you can get a better paying job. And that meant that that's much better for everyone concerned.

Will Page:

And just a quick, to put a Christmas bow and tie on this topic of my geekness about statistics. I want to remind Liam Maxwell and my audience, my favorite example of this, which is when you'd raise the issue of the environment, carbon emissions, that if you think about Wikipedia, it is the sum of all the well spoilage. It has no environmental damage and it adds nothing to gross domestic product, to which Liam Maxwell can think of lots of things which are not the sum of the well spoilage do terrible environmental damage and add lots of gross domestic product. What matters most is being measured least.

Richard Kramer:

Yeah, I mean undoubtedly there is a measurement issue with this, the broad march of IT. And I would argue because we had Corey Doctorow sitting here who's probably on the other side of you of many things, but a very brilliant critic in many respects. He would argue that a lot of those non-tech jobs have been relentlessly devalued to where we aren't paying enough to people who are working in social care, are cooking our meals, are doing the kind of things that we should be valuing higher. And that that Gulf is really unfortunate and highly damaging and we don't pay very much at all to the people who stand in front of our children and teach them however many cloud enabled tools we give them. We're just underpaying these people relative to their value to society. Whereas we hugely privilege these tech jobs because the tech industry is brilliant at promoting itself.

And in some cases I would regard the statistic you just gave us as a symptom of a problem, not necessarily something we want to aspire to, because there's a huge range of jobs, and I feel this very, very personally because I have an elderly mother that, finding people to care for her when she can't care for herself is very difficult. And those jobs can't be outsourced to tech. They aren't cloud jobs and we're not paying enough for the people who do those jobs because we've run towards the tech enabled future. And I don't know how we're going to deal with that in the future as more and more of those jobs aren't finding the simple volume of people willing to do them at the clearing price which we offer to pay.

Liam Maxwell:

Okay. So I mean I suppose my focus on that, and I can't change the way the industry goes, et cetera, but I think one of the things I'd say is that it is really important that we can, if I go back to it, help governments and government service delivery focus on delivering the frontline service.

Richard Kramer:

100%.

Liam Maxwell:

Not the administration. And being able to be flexible to do that is that, so one of the guys I work with a lot is a guy called Mark Thompson who was a professor at Cambridge and Exeter, and he did a brilliant piece in the Guardian, if you look it up, but it's about a Dutch system called Buurtzorg, and what it was was that this great straightforward selection of how you do the, get social care selected and who you get and what somebody turned up and just reformed it. So it was the most efficient, most simple, most effective way of doing it. And this monolithic system was put in place to make it work and actually made everybody's care much, much worse. And what they were able to do was to take a step back and say, "Let's go and design a service and system that's built around the pensioners, the people that were getting the care who were being given care and build that and make that work." And the flexibility and the ability to make that requires a lot of different thinking, but these old monolithic IT things were getting in the way and Buurtzorg was a way of actually being flexible and bringing that together and delivering more service and frontline service and less administration. And that really is where if I can do anything in the gig I've got is that's what we're really focusing on.

Will Page:

Maybe a philosophical question before we get you to smoke Signals, which is, I'm flying to Denmark next week to work on this Women in Music award and I always admire Denmark as a country, 5 million people, the way they open their shutters in the morning, the way they close them at night. It's a pretty impressive way of running the show. Is there a country government market backend that you admire and hold up as that's what good looks like, that's what we should aspire to do?

Liam Maxwell:

What good looks like in doing a citizen app, as I explained, was actually Ukraine. Okay, that's really interesting. Also Singapore and [inaudible 00:47:17]-

Will Page:

Singapore's incredible.

Liam Maxwell:

And I'm really lucky because when I joined AWS, I said, "Go and find the best people in the world to come and work with you." And I worked with a woman called Choy Peng Wu, she's the founding CIO of Singapore. And that's her vision and it's brilliant that you've seen that come into play. So that's something that I found really interesting. Service, New South Wales, just focus on the service and get it across. That's a really good example. And then of course the one place which I've always found absolutely spell binding and how they did it, there's a guy called, there's two guys, Pramod Varna and Nandan Nelekani who put together the Aadhaar program in India, which is beautiful in its simplicity about how it condensed the identity program into just a question of is Pramod Pramod. Is the person in front of me, the person in front of me. Eight pieces of data, one 12 digit thing, 1.4 billion people signed up to it. Absolutely unbelievable. But-

Will Page:

Economic and societal progress that unlocks.

Liam Maxwell:

But the economic progress there because then you build the banking stack on top of that, all of those components and the effect on female economic empowerment in India is now starting to be felt and starting to see it. And you look, if you think about where competition is, absolutely amazing place to understand.

Will Page:

How often does the word leapfrogging into your mind?

Liam Maxwell:

Oh yeah, yeah. That's like M-PESA in Kenya.

Richard Kramer:

Yeah. And I'll tell you that it's sad because you see also too much attention paid to these terrible examples like as you and I both saw the slow motion car crash, which was this giant NHS IT project, brought in every vendor under the sun and spent tons of money and yielded nothing.

Liam Maxwell:

But because very much, and obviously I wrote a pamphlet all about that called Better For Less in 2010. The whole point around that, was that was built around the system. And going back to a guy I mentioned earlier, there's a book called Tom Loosemore who's just brilliant at identifying that you need to build the services around. There's two things. You need to build the services around the user. I used to have on the back of my phone, "What does the user need?" You've got to build around the user. And the other thing is that the component of delivery is the team. And that's really, really important that you have the teams together and you have, small teams can do huge things. One of the other things, brilliant thing about where I work is everything we start is built on the basis of two pizza teams. You start off with a small team that could be fed by two American pizzas, and if you can't make change happen with that team, it's not going to happen. And that means, so you can innovate, spin things up really quickly with small teams to do it, but it's the team that matters.

Will Page:

And those two American pizzas are delivered using that geographic app you built on top of AWS and the case study you gave us in part one.

Liam Maxwell:

Yeah, you could say, I think could be, yeah.

Richard Kramer:

I don't know. I can eat a whole American pizza, but that's me. And by the way, I also, I lived for a year in Singapore in 2011. Going there from the UK with three young children, we had a wonderful year there. It could not have been easier because everything worked. All the utilities were paid by barcode at that point, which was minor miracle. I mean, you would still be able to go to the post office and pay your utility bill in cash in the UK. You might've had to. The medical records, all the citizenship records, all the things that I needed to live my life were immediately digital and available to me in 2011. And you had a chip pin in your card in your car that you stuck in your car, you went into any parking garage and it just charged you and it told you ahead of time whether or not there was a space free and you drove in and drove out. You never saw an attendant. It just charged you what the cost was and smart metering and all that, now-

Will Page:

Just to add to that as well, the only time back to where we are in Britain in 2023 going into the next year's election cycle, what's the only time you ever hold a pencil? It's when you have to vote. It's just weird, isn't it? I can't think of any other time in my life where I hold a pencil than when I go into some primary school to commit an act of democracy and vote.

Richard Kramer:

So Liam, we've got to get you to our final segment of the show, which is our favorite. We ask all of our guests to give us a few of what we call smoke signals, the kind of things that you sniff in the air and you say, "Well, that spells trouble," and let's just confine it to your area of expertise. When you're sitting down with government ministers and they say to you, "X," or you're sitting down with your technology partners, because I know AWS works with literally everybody under the sun in the tech world and they promise this, that or the other. What are the couple of things that, phrases you hear or things that customers ask for where you just go, "Huh, this is not going to end well,"?

Liam Maxwell:

So I think part of the approach we take is very much about listening to what people want and what people need. And that's really important. And sometimes the way that people express what they want doesn't describe what they actually need. And there is a big difference between the two things there, and that's part of where we can help people with that. I think in industry, why, I think, you can still find these neologisms, and certainly in government you find these neologisms that kick in. If I give you an example, when I started in government, the Institute for Government released a report which said, it used the word agile a lot. Everybody passed-

Richard Kramer:

With respect to the British government?

Liam Maxwell:

And so everyone then decided they were going to go agile and...

Will Page:

Smoking pretty densely.

Liam Maxwell:

And Mike Bracken and I went to a particular department and we tossed a coin and he went on the shop floor and had a look at all the administration. I went in the corporate boardroom and they were giving me the presentation of how they were going to change their particular department, agency, whatever. And on the board, I went through this huge PowerPoint that they gave me, and then on the board appeared this word and Mike just appeared at the door, open the door to come in and on the board was the word "agilution" because they wanted to go agile, but they wanted to do it at an evolutionary pace. And I just remembered this door opening, him looking at it, and then the door shutting and he just went.

And so there's the ability to, the thing that I noticed is the ability to grab hold of those neologisms, but being vocally self-critical. We then went and created our own one, which was called Digital by Default, which then lots of other people came back to and said, "Oh, this is digital by default." And because we hadn't been clear enough in describing it, it didn't, other people went off and took it. So the bit where I find challenging in those neologisms is where you can't follow it up with the substance. And that, I suppose, without wanting to blow smoker where I'm now, and I really enjoy working where I work now because that's so anathema to what we do.

Richard Kramer:

Analog by actuality as opposed to digital by default.

Liam Maxwell:

Possibly. So it's those, yeah. So it's possibly sort of phrases like that when you get people talking about blockchain a lot. That's one of those bits. And-

Will Page:

Our producer's going to edit that word out this podcast, I can tell you now.

Liam Maxwell:

Yeah. It's one of those bits where the hype, we all know there's a hype cycle. And I suppose the thing that we...

Will Page:

When Mckinsey announced that the metaverse would be worth 5 trillion, which I worked out was the GDP of Japan.

Liam Maxwell:

Right. Afraid I'd never saw that. But the ability to, I think you've got to get back to that ability to actually have a thing that you talk about as opposed to a bit of vaporware is still the thing which people trust. And earning trust with people is the most important part of my job. And you just don't get anywhere if you suddenly turn up and go, "And in three months we're going to do this," it's like doesn't happen. It's much easier that you can make all these things work. Go back to the Ukraine example. One of the reasons why when the Ukraine, we talked about the Ukraine thing inside the company for the first time, and we didn't do it until it was well underway. We did it in the October and all the change had happened in the year up until June really, we didn't want to talk about it because it wasn't done. And the other component with it was that all of the components that we used were things which had existed for a few years. I mean, Snowballs had been around for a while and we just used existing techniques in a simple way to help someone meet a simple goal. And that was the way of avoiding the bubble.

Richard Kramer:

Yeah. I think we're going to wrap up by saying this podcast needs to be more agile and we'll have to evolve

To doing so in our next 200 episodes, but after you listen to this, be sure to tune in next time. For our very special guest for a hundredth episode, the only man to have ever gone from having CFO of the Year to a prison card within one year, Andy Fastow, former CFO of Enron. But I'd like to have a huge thanks to Liam Maxwell, who's had an incredible experience at the heart of government and somehow survived with his sanity intact to join the private sector and help other governments at AWS. Thanks very much, Liam for joining us.

Liam Maxwell:

Thank you.

Richard Kramer:

A very special thanks to Oliver Blois and Stella Massonet at Platoon Studios in London. If you're new to Bubble Trouble, we hope you'll follow the show wherever you listen to your podcasts. Bubble Trouble is produced by Eric Nussbaum, Jesse Baker, and Julian Natt at Magnificent Noise. You can learn more at bubbletroublepodcast.com. Until next time, for my co-host will page, I'm Richard Kramer.