← AWS Podcast
AWS Podcast
#751: Werner Vogels’ Tech Predictions for 2026 and Beyond...
00:30:54
For the past 6 years, Werner has published his annual tech predictions, where he’s covered everythin
Loading summary
Transcript82 lines
- [00:00]
A
This is episode 751 of the AWS podcast, released on January 12, 2026.
- [00:10]
B
Hello, everyone. Welcome back to the AWS Podcast. Simon Lesh here with you. Great to have you back, joined by the most special of guests, Verna Vogels, VP and CTO@Amazon.com G', Day, Werner. How you going, mate?
- [00:21]
A
Thank you, Simon. Always pleasure to be on here.
- [00:25]
B
Always good to have you here. And we're talking about something challenging and good, and this is all about predictions. And so before we even talk about the predictions, why, why would you, a man who travels the world seeing the world, say, I'm going to pause and make some predictions. What drives that behavior?
- [00:46]
A
It's hard to say on some level, but I see so many customers over the years that I see patterns arising that I think, oh, others should know about that. For example, in healthcare or whether that is in. In how young businesses help smallholder farmers become successful or. Yeah, all of these kind of things, I think, are sort of, sort of signals towards the future. And as such, you know, I like to write them up. And it became as just something I would write up on my blog. These are the five things that I've seen in the past year that I think are really important for the coming years. And yeah, at some moment it looks like that there's a lot of people interested in actually sort of what might be happening in the coming years, at.
- [01:45]
B
Least thinking about it and taking that almost that ritual to sort of look back and look forward to try and orient yourself for what's coming.
- [01:53]
A
Well, it's a very large part of my job actually. You know, it's. It's listening to multiple CIOs and then trying to figure out what is actually the issue that they're all talking about now today. Of course, that is easy. Every CEO says, what should we be doing with Gen AI? Question number one. The one or Gen AI predictions, necessarily. Byron Cook, our most famous automatic reasoning researcher. Brilliant, brilliant. And I think on one of these postcards he said that this is exactly what Kostim is asking, what should we be doing with? And he goes, well, sorry to answer your question with another question, but why are you asking me this? And that clearly shows that at this moment our customers are confused. They're looking for what they need to do and they have a massive fear of missing out and anxiety and things like that. And so our task as AWS then is to bring back the trust in the path that we are going.
- [03:13]
B
That makes sense. Now, it's interesting to me that your first prediction is. If you had given me 10 guesses, Verna, I would not have got it right. It's about loneliness.
- [03:23]
A
Yeah.
- [03:24]
B
So tell me about. This is a really interesting one. I like the angle you've taken. Tell us more about why this is a prediction than the nature of that prediction.
- [03:31]
A
Well, I always think that now we have lots of roles to play as technologists, and of course, we can write better spam filters or we can build lots of new technology. But there are so many problems in the world, really big, hairy problems that technology can help solve as well. And a number of years ago, almost in Japan, and traditionally in Japan, children would take care of their parents as they grow older. But these days, young people in Japan want to make career. So there's enough young businesses in Japan that are looking at this particular problem of elderly people growing older and older and not having anybody to care for. What kind of technologies can you build to help those kind of people? And I haven't mentioned loneliness yet, but that seems to be almost an epidemic. Elderly people alone have nobody to talk to, have no interaction, have no social engagement. And also, a number of years ago, I saw a presentation by Kate Darling from mit, who is all about sort of these kind of companions. And one of the things that. One of the most amazing things that really got me, got me thinking there, is that people treat these companions as pets. They don't treat them as a piece of metallic, a machine, really. There is an interaction. They give them names. You know that 80% of the people that have a Roomba.
- [05:32]
B
The robotic cleaner.
- [05:33]
A
Floor cleaning thing, iRobot have given their Roomba a name.
- [05:40]
B
80%?
- [05:40]
A
Yeah, 80%. And it's even to the point that if they send in the roomba for repair, iRobot would say, you know what, why don't we send you a new one? And the customer said, no, no, no, no, I want mail. Sweet back.
- [05:57]
B
I want my room back.
- [05:59]
A
Yeah, no. So there are many cases where just robotic companions, digital companions, can help people. And we've already seen that happening in how to learn languages and things like that. But these kind of devices, these kind of companions, become extremely important. A great example is that younger people, There's a device called the Hugger. It's green, it has a funny snout. It's whatever. It's really targeted to its kids, and it's also targeted to its kids. One of the biggest problems that kids in hospitals have is talking to the doctor, have no problem if they have one of these huggers who actually also talks to them. And There's a back and forth and back and forth, and they make them less scared of the doctor. And there's so many different roles to play for these kind of devices to help people not feel that lonely not being on their own. A kid alone in hospital is desperately lonely, needs a companion and scared.
- [07:15]
B
Fear and loneliness are two terrible combinations.
- [07:18]
A
Yeah. And we already know enough about sort of the stories about bringing dogs and cats and things like that into a hospital and what kind of positive effect that has. It turns out that these digital devices or these digital personalities also have that same effect.
- [07:37]
B
Interesting.
- [07:38]
A
Interesting, yeah. And of course, at Amazon, we built.
- [07:42]
B
One.
- [07:44]
A
And that one is actually very interesting. And if you haven't taken your medicine, it will go search for you in the house. Ellie will ask you, did you take your medicine today? So there is tons of small things, and these things don't necessarily need to look like humans, but there need to be some form of interaction. I remember going back to the beginning, Kate Darling's story was that at one moment she did a test. So she broke her students up in two groups, both called a dinosaur, a puppet dinosaur, and got a name, and they did all sorts of things with her. At some moment, she asked them, he gives him a hammer and says, no, break the dinosaur. They can't do it. No, no, they can't do it. They're mostly engaged. Gives the hammer to the other group and says, kill the dinosaur over there. No problem whatsoever. Yeah. So there is a social relationship between these technology, devices and the human. And as such, I think as technologies, we played a role in making people more lonely. Everybody stands at a bus stop and the bus takes five minutes, pull out their phone. They're not talking to the person next to them or whatever, or complaining or whatever. No, they're in their own world. So quite a few of our technologists have sort of accelerated this loneliness. And as such, we have this principle at Amazon that with scale and success comes broad responsibility. This is also on our responsibility, not only to create a problem or contribute to the problem, but also contribute to the solution.
- [09:46]
B
Solution, yeah, absolutely, absolutely. I think it's a fascinating dimension that has real health outcomes.
- [09:52]
A
I think there's important things happening in our society where technology can play a role in. In trying to solve them. I prefer to start with a problem like that that actually not only technologists but everybody else actually knows about. And there's so many side effects of bringing technology into this particular world is that people can stay at home longer. People. People are very lonely, degrade much faster and Our society is getting grayer and grayer and people are getting more lonely and people getting older and such. These kind of devices are becoming more and more important in keeping people to be. The Dutch word is so don't really know what the English word is. On their own. Yeah, they can't take care of themselves.
- [10:55]
B
Independent. Independent and self reliant.
- [10:57]
A
Yeah, self reliance. And so that is a significant cost to society as well. Because if these people need to end up in a care home or whatever because there is nobody else, then that's a significant cost to society. So these things are things that we just need to solve for people in the first place. But there's economic reasons as well.
- [11:22]
B
Exactly. Many good reasons. Now let's change tack a little bit because you mentioned how you get to speak with a lot of CROs globally. Another cohort I know you get to speak to a lot are developers. And I think one of the number one topics developers certainly talk to me about is what does my job look like now that Genai has come along? You have everything from there'll be no developers to developers will save us because of all the Genai slot being created to everything in between. Tell us about your prediction and how you see this from your perspective.
- [11:57]
A
Well, first of all, I think we need to go a little bit through history to put things in context. Today when I went to school, I learned three languages. 68,000 compiler, 68,000 December, Cobol and Pascal. Nobody writes in those languages anymore. And then we got I still love cobol Object.
- [12:22]
B
Don't make me feel bad.
- [12:24]
A
I like COBOL as well, by the way. Then we got into object oriented programming and then we got into functional programming and then we got into from C to C to whatever. We. Almost every few years there's been a change.
- [12:40]
B
Yeah.
- [12:40]
A
There's been new tools.
- [12:41]
B
Yeah, true.
- [12:42]
A
If I think about my IDE was vi.
- [12:46]
B
Yes.
- [12:47]
A
Not Emacs.
- [12:50]
B
That's a whole other podcast.
- [12:52]
A
Yeah. Now what came after that? Borland. Yep.
- [12:59]
B
Yep.
- [13:00]
A
With their look at all the things we got after that. We got. We got Intellij, we got VS code almost everybody uses today. And everybody loved VS code. Nobody talked about VS code anymore. It's all about cursor, it's all about Cairo, it's all about the new tools. So here we have 20, 30 generations of tools that every time made us better, every time took away hard work. We have now quite spectacular change happening. But it's still tools. Yes, it's still a tool. And remember, you do the work, not the tools.
- [13:50]
B
And you're responsible for the work also.
- [13:53]
A
Absolutely. But it's still you doing the work, it's not your tools. Yeah, they're going faster and things like that. But you know, I was happy when I saw the first compiler that compiled down to 68,000 so I didn't have to write the 68,000 stuff. Yes, yes, yes. And, and now of course, yes, of course we're going much faster. We can be more efficient. We can be. But there are tons of things that I think developers need to sort of start to focus on. And I call this the Renaissance developer. Mostly because I love the time of the Renaissance where there was a real change in society where technology and art and everything became more important. Yes.
- [14:42]
B
And valued.
- [14:43]
A
Look at Leonardo da Vinci. There was so much creativity and so many things going on then. I think there's some things that our current developers can almost learn from that time. But in a modern way. Yeah, but in a modern way.
- [15:01]
B
It's the spirit that's really changing too because. Interesting. Like we, you know, we talk about developers. I often met folks that just that have called themselves coders. You know, I do coding, I'm a coder. It's like, no, you know, the machine is the coder, you're the developer, you're. You understand the problem domain, you understand the customer, you're thinking about maybe novel ways of solving things. You're challenging the machine to do things like that's the, that's the job, that's the career, that's the art, the craft.
- [15:27]
A
Yeah. In all of that. And I think one of my things where I think developers need to spend way more time on is on communication. Yes. Communication is becoming more and more important. Last year or two years ago in the Frugal Architect, I gave this example of we can cut up the Amazon webpage into three or four different tiers and each of them are increasingly important. Without tier one, the whole thing doesn't work. Tier two is, is pretty important. Tier three is nice to have. But we as developers don't make those decisions. We need to talk to the business saying like what, what kind of trade offs do you want here? Do you want four? Ninth for tier one? Three. Ninth for tier two and who cares about tier three? Yeah. And so, but to be able to do that, an engineer needs to be able to communicate to the business. Yeah. And communication becomes more important. What does this do? And also understanding at the same time, what is it that my customers, who often are internal and not external, this is that they really want.
- [16:41]
B
Yes.
- [16:42]
A
Nor is what they Say no, no, no, no. And as such I think we've done something really good at. AMSO is putting the people that work on shoes, technology for shoes, therefore the whole thing sit with the shoe team. Yeah, so.
- [17:01]
B
So they absorb.
- [17:01]
A
I think, I think one of the, one of the bigger problems in the earlier days of, of tech, tech technology was that we got an assignment, you went off a came back and I wasn't good enough. Why? Often because requirements changed.
- [17:19]
B
Yes.
- [17:20]
A
And it's always the case. Requirements always change.
- [17:22]
B
It's guaranteed.
- [17:23]
A
This is what customers do. And so being in the same room with your customer makes that you build way more efficient and way more important software because you build exactly what your customers want. And the such communication there is crucial. Strangely enough, the school where I went to my undergrad had an interview class and I felt like interview class?
- [17:57]
B
What am I doing this for?
- [17:59]
A
Why do we need to talk to journalists or so no, it is how do you talk to your customers? Because quite often your customers come to you already with an idea in mind how it's going to be solved and they already have some tech stuff. So for you as a developer or an architect or however you want to call yourself, it is important to be able to listen and to drill into what is it that my customer actually really wants, not what kind of technology does he want to use. What is the end product? What am I actually supposed to build for him? Yes. And maybe he's listened to, he read some magazines and listened to the media and all things like this. This needs to be solved with an LLM. Well, maybe not. Why don't we go back first to what is it actually that you want to solve? And this is a skill. And I find that as a technologist myself, it's not necessarily my boss note, necessarily my strongest strength, but it made me aware that customers, they're thinking ahead much further than where they should be. They should be focusing on the problem that they have instead of how they think they should be solving it.
- [19:24]
B
And they're coming from perspective. So communication and problem solving. And I think you're right that cohabitation becomes really important too. And it's interesting because sort of when I reflect on my career now, one of the appeals of getting to IT and computing was particularly because I came from the pre Internet days was I was on my own on my computer, I didn't have to talk to anyone, I just did stuff. It was great. Look at me now, talking to people around the world all the time, you know, you can't get away from it if you want to build something meaningful, you have to go speak to someone.
- [19:54]
A
Well, it's crucial also and I think, you know, the small team approach sort of promotes this because if I have a small team, let's say less than 10 people, 10 or 12 people, I know what everybody else in the room is doing because we talk all the time. Get it more. If you do scrum and you have to stand in the corridor in the morning and you have a team of 40 people, that's a pretty difficult scrim.
- [20:25]
B
Agree. So let's take another shift. And this particular prediction is a domain that I, I think about in the context of driving my car and I look in my car rear vision mirror and it has words on the side there that say objects in the mirror may be closer than they appear. So obviously we're talking about quantum because it feels like we've been talking about quantum for a long time, but some stuff's been happening that makes it different. Help us unpack that a little bit.
- [20:53]
A
Yeah, I think all. So there's a range of private companies that are building quantum computers. All big companies, whether it's Google or Microsoft or Amazon are all trying to build as well. We have this very strong collaboration with Caltech. We've made a whole bunch of advancements in terms of error correction and networks. Others make advances in other areas and where a number of times ago I thought this is 10, 20 years away before these guys can ever build anything. But I see more and more advances that make me believe that this is four to five years away. And that brings a whole different problem set. Not what this machine can solve, although that's a, that's a whole conversation by itself. But one of the things that it will definitely be able to solve is decryption, correct? Yeah. And so it is crucial for us right now, not five years from now to start anticipating this to everything that you have encrypted now will be able to be decrypted five years from now. And we have enough tools now that you can start to integrate into your business processes already, which is crucial because there is something else happening and that's called data harvesting. That means that bad actors are already harvesting your data. Encrypted data.
- [22:37]
B
Encrypted data.
- [22:38]
A
They can't see it. They can't see it. Four years from now they will. So you know, it's it, it is more urgent than maybe I fought three, four years ago.
- [22:54]
B
And I guess because it's one of those one way doors as we talk about at Amazon where once Once it's available it's, it's. Everything's changed overnight. Basically everything changed and that's why I think the work that a lot of large organizations and NIST as well have been doing on, on establishing the standards for post quantum cryptography, the work our teams have been doing to implement it across things like kms et cetera so that customers can just get it. But, but you also think about in your prediction the, the challenges that many customers who have maybe lots of remote devices, older devices, et cetera, there's a, a big job of work to be done to do it properly across the board. Yeah.
- [23:30]
A
But I don't know but home automation was it problem before post quantum encryption? You know I mean the number of home devices that run a Linux that hasn't been updated in the past four or five years. Yeah. Zero day exploits. Can you open someone else garage door with a little bit of effort? Absolutely, yeah. And in a search note everywhere where we have used keys and where we have used encryption and thought we were safe. We will no longer be safe if we don't do anything make the change. Yeah, yeah. And that is. It's a bit scary to be honest because think about your hotel card. Can anybody now enter my room? Because I can just decrypt whatever is on the card or there's so many places where these digital cards are being used or you know what we have.
- [24:33]
B
Yeah. Pass cards, all that sort everything, everything relies on this encryption that is no longer valid.
- [24:38]
A
And yes and as such in the coming three, four years there is a lot of work to be done now. A lot of work. Is it what the hyperscalers do? And I haven't given credit to Microsoft and Google because they working on this just as hard as that we do. But you can already start using post quantum encryption on the wire because S3 has this cloud font. You can talk post quantum encrypted messages to quite a few of our services and the TLS implementation that we did quite a few years ago and it's open source handles this as well.
- [25:27]
B
It does, yes.
- [25:27]
A
So we have open source technology that anybody in the world can use and they shouldn't be waiting and not because it's us. But I still believe that Amazon's most important part is to protect our customers and protect our business because without that we don't have customers and we don't have a business.
- [25:50]
B
Yep, yep. So true. So true. It's a virtuous cycle now we're coming close to the end of the episode and there are other Predictions that we're not going to talk about in detail. They're predictions around defence and education. But what I want to get is, I guess a big picture view from yourself, Werner, of what are you seeing? Is there a theme, is there a connective tissue that really resonates with you about how you're thinking about things?
- [26:12]
A
Well, the thing is, if it would be up to me, I would be writing four or five predictions that all have to do with, let's say, the biggest problems in the world. Hunger, economic security, healthcare for everyone. You know, I can tell tons and tons and tons of stories about sort of the amazing work that companies are doing. So let me give you one. Can I have time for one?
- [26:39]
B
Okay, absolutely.
- [26:41]
A
So I was in Kenya not that long ago, and of course when we stay in Kenya, we stay in nice hotels and nice neighborhood and things like that. But driving 10 minutes out, you come into the real Kenya and how people live. And understanding how people live there. They go to the bank, borrow a few dollars in the morning, buy some stuff, if that, try to sell it during the day, go back to banking, pay back. Hopefully you, you have 40 or 50 cents made and you can buy food from that. Yeah. But you can't cook it because you don't have money because a, a big canister of, of liquid gas cost $10 and you don't have to, you don't have that money.
- [27:27]
B
Yeah.
- [27:28]
A
Yeah. So there's a, there's a young business called Coco Networks that try to solve this. Yeah. And you have a sort of like ATM machine we can go to with a little canister that can hold the gas. You can plug it in and you can say, I want 20 cents of gas today. Wow.
- [27:49]
B
Like a really micro amount.
- [27:52]
A
Yeah. And so boom. Go home, cook your food, eat. And this is where a young business sees a problem, understand what the bigger space is, where this problem operates in. They understand the people and then build technology solutions for it. Because without technology, this wouldn't help. And this is an easy example. But I can give you dozens and dozens of examples of young businesses that are not looking to become the next unicorn. The thing what they want is solve hard human problems. And the United nations has the sustainability goals almost in every one of those, there is a startup that actually is thinking about how do we solve this big problem of what can we do? Or take the ocean cleanup that we talked about? We talked about that. Yeah. Amazing, amazing work doing. But still young people not thinking about becoming a billionaire, but solving hard human problems at this moment with technology, it's inspiring.
- [29:08]
B
It's inspiring. So you've been doing these predictions for a while now, Werner. Is there one that's really sort of stuck with you that sort of really resonates?
- [29:19]
A
I know maybe this year's, I mean, one that this year definitely resonates with me is that I've always been an amazing fan of Ken Robinson. And, and Ken Robinson always hammered on that the way that we're doing education is wrong. Yeah. We are training people for confirmity, not for diversity. We want them all to be the same. They get the same lessons, they get the same education. And, and he. It tells great stories, by the way. And if you guys want to go on YouTube and look up Ken Robertson, you'll see amazing. It tells various stories about how people are actually more creative if you just let them be. One thing that we do with children is that we kill curiosity at a.
- [30:20]
B
Time that they're most curious in their lives.
- [30:21]
A
Kids are naturally curious. They're always asked these kind of questions and things like that.
- [30:28]
B
I like it. I like it. So, Werner, all your predictions will be out on the blog for folks to look at and to read and to comment on. And I'm sure you enjoy that interaction with folks as well as they assess your predictions and that sort of stuff. So it's going to be great to see the feedback we get. We, of course, love to get your feedback. For listening, AWSpodcast@Amazon.com is the place to do it. And until next time, keep on building.