Featured Session: 10 Breakthrough Technologies of 2026

00:00:05 Speaker 1 How are you doing? Good. Hey, yeah, so nice to be here. So nice to be here. Um, Jeff from Neil and the tech guys and I think everyone who reviewed and so happy to be back here at Oxford University because I think our plans are breaking technologies again. I feel like back all here in this room, I just feel like we're friends. I'll tell you why, because it seems to me that if you talk about stuff like this and you talk about what's going on, maybe curious what's going on more about how the world is changing and what can we. What's happening next? That's why I do this job. I often know what's happening next. And you know, what's happening next, we're in the right place for that, right? We've got the right stuff like that for us to get a sense of what happens next. If you read these journals and magazines, a thousand different views and projections on how the future is going to change. And I think it's worth sort of reflecting on this over ten years. :. One of the possibilities is that making predictions about emerging technologies. 00:01:05 Speaker 1 So, it's very easy to speak about anything and not to hold you. But what's really, really hard? And, that's not just because technology itself is ever changing the context for us sometimes today, right? But also because technology collides with so many other things in our hearts and our heads. So, for example, we have funding policy problems of scale and sometimes people don't really want. 00:01:36 Speaker 1 The future in the target is always moving. How that makes predicting stuff hard, I think getting it right, isn't just about simple as saying " this is what's coming next." As technology continues to improve, I think the thing is that we are looking to say which technologies are going to matter in the long term. It's not just which of your thoughts on the radar. And a different way is a side of that. So could be scale, could be how many people. 00:02:06 Speaker 1 What we're looking for. Or is it that how completely does it change how we live or work or see ourselves? And the answer that we found in twenty five years of experience at Reboot is, uh, is that really what people do things with their time? Uh, so I'll give you my my job actually. Um, yeah, technology that changes everyone but changes nothing isn't really. 00:02:36 Speaker 1 And the mind-blowing idea that never kicked out throughout is what here. So, A technology that just changes how very few people do their jobs isn't really innovation for us. We don't believe technology exists in a vacuum like that. Brilliant algorithm or brilliant invention is just that, it's just that particular curiosity. What do we have to tell them? And why do you hear it? What's the knowledge coming out? And the area is not just knowing what's coming out, but like to know what people need to need to realize. This is really prepared how to prepare for the future of society and I think it's like. Social media is a great case in point. So we've got social media means to generate social media that means interacting with one another in these online spaces and sharing stuff about our lives, which is really cool, great. Um. 00:03:36 Speaker 1 But we didn't really think about, I think it was a Pat Hockney quote, the other side, the downstream impacts on teenagers with their body image. Just our basic collectibility and bring awareness impacts. So, we're in a similar place with many of the technologies on this list now. That's precisely why we published them ahead. Not just can we build it, but how do we build it? What happens when we do? All right, just give me a moment, tell you a little bit about who we are and what we do. My name is Neil Halligan I'm, the head of the marketing group. Uh, he's going to speak to us in a few minutes about some really interesting concepts. So who are we well open mind might be who you've been in contact with? But we're probably the most independent thought leader I think on how to engage today in what we promote. So we have our main new market access technology campus in Sydney. 00:04:37 Speaker 1 New York, I love it. Um, could be pretty about those you put on a band. You probably know it, yeah? I have a website for a lovely three-party league. Uh, and then third time we get that sort of team. I obviously see what our goals would be. So, aside from all this uh, i might have a number of lists actually got ten great tips on the list, and we've got thirty five other things. 00:05:37 Speaker 1 I was quite inspired by the story of Rajkumar. I'm going to analyze the article that came out in the newspaper. Currently, uh, in Georgia now, uh, we've been developing these structures. Okay. And I said, "I'm just going to do a few things on our annual highlights." Oh, yeah. Thank you. 00:06:09 Speaker 1 For a subscriber to be along. So, yeah, I'll then go through the process we go through here at TechCrunch. Right? We get, we all get together, all the reporters and editors to dump all their ideas into a big pitch document, sort it all by category, and then we have multiple meetings, arguments, email exchanges about what which ones are going to make the list. And again, we're looking for a good mixture of like. 00:06:39 Speaker 1 World, not just for AI. For example, And this gives us something to go back to the twenty fifth edition of it. It started in two thousand one, and all our members are there for our educators. And over the years, we've got some things right that you say. I'll give you a couple of examples again, just to show that what we're doing kind of. um. with an early adoption trend from technology. We're going to need cryptographic protocols that are protected against attacks from future quantum computers. And, in twenty twenty four, the U S National Standards Institute standards technology released some algorithms, which they said were now quantum proof. Just in time, there are loads of new developments happening in quantum with Google, Microsoft, and others getting closer. 00:07:42 Speaker 1 So, what's so cool about them? We can just go into these cells and then revert these cells to create any other type of cell. Any other type of cell in the body. And they only put it on this in twenty ten. So, We need to test the treatments for epilepsy and Parkinson's disease as part of those trials for this technology. We just saw a couple years ago about the same thing. 00:08:13 Speaker 1 So, it's not so more than into the place in the next. And then, after they up in twenty fifteen, but it's total less. I know only like three people on the house here, actually more than my colleagues with this. And then six years ago, we put satellite constellations on our list. We've spent enormous satellites in orbit that certainly tripled through mass and drop of being doing from Starlink constellation. For that, it's a nightmare for astronomers. So, you see these pictures from Hubble that show the night sky and those fine lines from the starlight, otherwise kind of breaking with you. You don't hear them on five or so much. And so much more confidence in this, But we'll do it if you can get one, of course, here through actually dropping clouds. And then like he doesn't look stuck out. 00:09:13 Speaker 1 And then, What they do is a quick kind of presentation and suggest what the business could have done to reverse changes in people's history. And so, They kick out a few ideas from the last twenty - five years of things. We got wrong that they want to start applying. So, in front of our future might be on some other people's holidays. In particular, this one called "Language," it was actually a really cool idea. The camera back to save that about. 00:10:16 Speaker 1 And we find companies to incorporate your unique gesture code in products and services. It works well, but privacy and security challenges. And then, lastly, the project Blue. Of course, I have a softball project, which is called the Global Look of the Air Showing Products. You're glad to see them in your product. You can try out the solution. So, put a list of things that are being supported by the project. You can see that we shut down in twenty twenty one because people were pushing their issues to such environments. So, What we say sometimes, it's not so much about things you have here at your project as you need to do. All right, I'll keep you waiting for a couple more minutes. What's all of them to do? Because we have this big argument as well on the other list, maybe it's along there. 00:11:16 Speaker 1 I'll tell you what. So, first, they're male competitors. So, as you know, there's a lot of support for men the Commonwealth Games. There's. Now just any good support for men getting through the trials and a shelf of them to go on. 00:11:45 Speaker 1 So, really cool. Um, I think it's much more complicated, but we're still a few years away from seeing these in FDA approval. So we thought maybe let's just wait on them for a little while. Next up, world models. I didn't think they'd be controversial ones because this is incredibly hard to do. There's some big name AI researchers working on the idea, Including Yann LeCun, who got that early in the story with FAI Week. And then he. And we're also trying to set up our own world models, companies to follow that. But, a few of the startups is working on it as a way of systems that learn the underlying rules of the world from observation, like baby learning about gravity. And a lot of anticipation about what these things can do. What's will model represent? I feel is making it much easier for robots who can practice us and learn stuff about the world without having to be programmed in. But what are we working on? 00:12:48 Speaker 1 Before we actually see them in the real world, but seeing what's contained in so many many next ones. And then the last technology I didn't make a cut was quantum personal. As you are aware, it's getting incredibly hard to know what's real and what's not online. You know, The other day, we saw stories about agents that went out to the surface of other systems that humans entirely on its own. Also, there is AI voice clones there are, Characters that we bring about, but basically interestingly called from humans. So we need to know, we need to know a lot more about how people are online. How do they act? How do other people work with them? They should be posted on this kind of thing. But normally if I have a single approach to doing it, And no one's actually using them yet because there's just a whole lot of them not coming back on me. All right, I've teased you enough. 00:13:48 Speaker 1 Uh, let's dive into the potential of these four technologies. Ready? We go now. All right, first up: sodium-ion batteries. Most batteries right now that you want to get home or in your car for going to be are based on lithium. Lithium is expensive and hard to find ; causes of supply chain complications in the world. Sodium-ion is everywhere; it's mainly we can. 00:14:18 Speaker 1 Thank you. Being far cheaper to make, and more reliable. Okay, the batteries themselves work basically the same as lithium ion. Energy by shuttling electrons, ions between two electrodes. These batteries are much safer. They go Uh. And he worked in Indian and Cree villages. Hello. Lithium, uh, most comes from three countries: China, Chile, and Australia. Some of it is much much easier to find. And, what I've got to say is that we need to be thinking about this issue. And I think there's some kind of a problem with the cost of lithium recharge batteries, which can be massively transformative. But actually, um, Lithium - ion batteries won't go back ;. They can't store as much energy as lithium - oxygen because they're about but always going to be probably more. 00:16:21 Speaker 1 Some of those, I know it's probably one of the bit crowd. A lot of you know a lot about already, but I'll tell you why some of this is done. Talk about impact and scale really major major part in the service workflow. I think we're getting engagement because that world survey in twenty twenty five found that sixty five percent of workers are using these tools weekly. Ten thousand eight surveys late last year, what needs to be done? 00:16:53 Speaker 1 The first time I was in the market, probably about five years ago, I was working for a number of companies. A lot of the work that we were doing at that point was around AI companies. Full code came on here, Although it's continued to break out and expand into full code more fully last year than it did last year by a product. Probably used to be either masters or still maybe still be out there. 00:17:23 Speaker 1 System, we really from these existing problems. Fantastic, and I was so excited about it because you know, A term called " pull of benefits, ", where Mark and they go to a store and build websites to apps using code. It's kind of just coding, and coding is also thought for code work. It's kind of the same thing, but it's like the more general stuff like research and documents or information. What it came about was that problem. We're watching to see how people are using code. We noticed they're using it for non-coders, so we kind of applied it to mostly new lines as well. And essentially, it's add-ons to core work: web design, HR, finance, legal work. And by the way, core work was pretty much entirely done by code. So I'll give you a difference in the making now is actually a very difficult one to make because this is moving so fast. There was an influential study. 00:18:53 Speaker 1 How serious you are. Most of the time, I think it's clear that many of those. uh. because um. thirty three of those who have seen it, but I don't think we'll see any of those. But there's a lot more than fifteen days out there. It might be true from some of the other possible errors that we've seen earlier, or able to communicate them anyway. All, we know is that software development has changed forever and has not gone back. That's just all possible. 00:19:25 Speaker 1 So, conventional nuclear plants are expensive and very difficult to build. The last one that's built in the US, in Georgia, is twenty twenty four, nearly a billion dollars above budget and finished seven years behind schedule. And next is all going to be powered by H D rating cooled water. 00:20:26 Speaker 1 Deploy a new kind of reactor, or just go back to the next smaller, more regular power. That isn't what we do. Another sort of cooling water coolers like molten salt or metal. These reactors can run at much hotter than hot water, and that helps us to help move things around high temperature heat in the hot spot. And. 00:21:26 Speaker 1 I see where the is a really good kind of multiple reactor. It's cool, And I like the kind of new one, which can do world's first modular reactor come alive. Okay, I think we can have some more. All of these are technologies that are in the sector for other technologies, actually go to critical action machine data centers as well as these new markets. 00:21:57 Speaker 1 And if you're going to go to the last sport, for example, for the next time we see active, you'll have a back-to-back. It might happen in Colombia. It's been mentioned in football tournaments and programs getting through. Next one: AI hacking again. I probably want to show this one that probably only you know how the world is coming back here. People are increasingly using it. 00:22:27 Speaker 1 Um, some people, some companies can use chat like Chat GPT. We'll also dedicate AI to process portfolio management through platforms like Parallax and by direct recognition. Give you a sense of how popular it is: seventy-two percent of U.S. teams use AI for content creation. Pretty incredible number, and in this particular case, 00:22:58 Speaker 1 I'm going to do it. I'm going to do it. I'm going to do it. 00:23:29 Speaker 1 Thank you. 00:23:59 Speaker 1 So, The way to find out how to help themselves is by providing a briefing here and advice on how to get around with parents. On the app, you can have some better information than their parents can now. So, we've got there. That's one of a number of courses. 00:24:29 Speaker 1 So, it's important to go out and let them know that they're not alone. 00:25:02 Speaker 1 And after this, the church failed. General to carry on and walk them through each step process. We'll see some regulation in China, some regulation in California and New York, and I think more protection and safety measures from higher block. 00:25:32 Speaker 1 I'm going to show you a few examples of how we can use this approach to help people. 00:26:02 Speaker 1 Caused by single letter of DNA. And they've given a new form of the gene editing called CRISPR, called base editing, so that's actually rewriting the initial letters of bases in our genomes. And in fact, CRISPR is an amazing Nobel Prize winning technology. Currently you can reach or shut off genes which cause problems. But now base editing means we can rewrite entire genes and restore their function. It's also got to be safer than the rest. 00:26:32 Speaker 1 And you will get these kind of like generic or patterns. This very treatment was on just a unique case condition that you know probably never be used again. There is, it's a baby, his name is KJ, And his doctors will be monitoring the progress for the first two weeks and then two years. So four platforms to count really well so far. And the group that treats this wants to start off. 00:27:08 Speaker 1 Of course, not a fortune, but two or three dollars to create. And that has perhaps come down over the course of these years, as we've moved out from this narrative that we keep making it just like two or three dollars per page. However, These kinds of search engines could potentially queue up thousands of pages for reading at once. If they're particularly useful to the brain in use. It's only really our people who are so about. 00:27:47 Speaker 1 Thank you. The future, And we're actually seeing some ancient like six species in the dawn of the possibility today. And so, a lot of people have been working on these mice. They're called "wooly mice" because we edit genes into mice. 00:28:47 Speaker 1 And actually, do that. And after not just tweak the matter's own genes, but we did here, but insert African human genes for those species. We've actually got fossil now to create the child of us, speaking in the last 12 or more years ago. We're going to extract DNA from novel bones and use base editing. Full of cell fragrance, then a pulling of cells with three active panels, and they're called luminous. 00:29:46 Speaker 1 And also, was keeping them in undisclosed locations. So, obviously, work with people who are going to steal them. She's obviously studying right now. Yeah, have to do it. So, each of those animals has twenty genetic changes across fifteen genes that are associated with a larger brain size and facial features and really amount of sunny and white fur. And. 00:30:17 Speaker 1 Possible, I don't think it's straightforward. High violence while they're playing more the kind of practice of this one, like or this complete rejection of execution. It's as much philosophical discussion as it is practical. I'm not going to consider that speaking you just by late on the backboard, I would have to. But maybe society and this is really just a really. 00:31:20 Speaker 1 Is and it wasn't a sort of pushing samples of the character. And it's not just animals, although they are the easiest version of it. Similar work happening in plants, you know, That you try and make it. So the plant there is putting ancient genes into more crops, like barley or wheat, to make them hardier against pests or climate change or extreme weather. They've already got. 00:32:01 Speaker 1 This not only a wonderful essay, but also one of my favorites. Next, let's take the third possibility. This an extraordinary note to keep down how our elements work. That's pretty weird because eventually the elements are actually beautiful and interesting. You write it down on a little level ;. We don't know exactly why you should have anything essentially for maths. Maths is too complex to fill in the parameters. 00:32:35 Speaker 1 I, why? And hopefully, AI and our topic are all the real leaders in this. For it, I am talking to them when we're looking at a version of that on some of this. So you may come a bit to the whole response with real people. 00:33:34 Speaker 1 Like a door, or objects like chairs and tables, even more abstract concepts like disobedience to exist inside the model. And, they found that it might start to make thinking about the Golden Gate Bridge or show images of it, or even the Bay Area as I mentioned. They found with this version of code called "they can dial that up," and they made "Golden Gate Code." And yeah, kind of summing things up is then came up. 00:34:05 Speaker 1 To it into everything, Even if it's not, even if it's not. Sixty nine, ninety five. And after how we got that I said, oh, I, just you do this pretty much, how any of us will do it. So the mental arithmetic, but fine cool uh when they, when they looked closer into looking at what it was doing, they found it was nothing like what I said at all, because basically splitting into two. 00:35:06 Speaker 1 Similar numbers, and it came to the conclusion that the number was probably ninety two ish. And then at the beginning of European numbers, nine six, and knew that any number like that had to end with five. So if you had a number ending with five, it had to be ninety two ish. And, that's how I came down to ninety five. ;. It would have to be correct. But, it's not how you would expect us to go. Ultimately, we have X ray techniques, which are magnificent for observing things here. So we're able to identify what elements are really there. 00:35:36 Speaker 1 Not. 00:36:07 Speaker 1 The first way you use space stations is to kind of take shapes. The very first one we launched earlier next year by a company called Back, with an artist from Europe, what we would like to call "take shape number one." And the other two: there's one called Axion Space, they'll have inflatable space station. Sierra Space, partnered with Bezos's company Blue Origin. They won't just be for science ;. They could be anyone with money who want to go up. 00:36:37 Speaker 1 I'm not going to do it. If you want to do it, please go ahead and tell me what's planned. You can tell from some of the marketing material that I've picked up that what they're looking for is more compliance. Some of the imaging presentations are pretty rough, and they're not exactly what we would call ideal or optimal for our purposes. This picture here just shows a few features of the fracture :. It creates this kind of like a bladder, and these mucosal flaps that kind of hold you against the wall, and then you look at the top. 00:37:07 Speaker 1 And then you also have another approach, which is the kind of attack mode, which is how hard it is to get into space. It's a little bit more powerful. Russia, which is currently a partner in the ISS, is making its own space station too. India is also making one as well. And China already has its space station ;. It's not very old ;. It's only a few years old. Okay. All right, next one: NBO storage Now this is potentially. The most controversial one on the list. And I guess there's a thought that not all these technologies are necessarily good things. Just because it's a breakthrough doesn't necessarily make it better for society as a whole, with bad outcomes. So, For decades, people have gone through IVF ; they'll take tests to screen embryos for genetic disorders or chromosomal abnormalities. Now here we think this is the risk of inherited. 00:38:06 Speaker 1 And what improves pregnancy success rates. And I think most people agree that's good and fair and right. What about using the same technology to test characteristics related to a future child's appearance, behavior or intelligence? Feels a little bit different. But it's happening. A few startups now are promising more than they can in the way of doing just that as you sift through the embryos. And this is a whole genetic. 00:38:40 Speaker 1 A certain IQ or brown eyes, and then it's parents pick the best one. These tests are a little bit iffy as well as those sort of general iffy feelings that we've been getting around the whole thing. These tests are really probabilities for these kind of traits, but usually these traits are a mixture of lots of different genetic elements, not just one single faulting gene like for the diseases that I mentioned before. It's also incredibly expensive. And the data they're taking are fifty days, and people are watching the test. And then they say, "Hey, look, We already make thousands of decisions that influence our children's future and schools and tuition neighborhoods." Is this different? Do you think it's different? 00:39:39 Speaker 1 Seventy four, thirty seven percent of Americans would be happy with screening for intelligence. But okay, technology is on the rise. The most we get back in the bottle now. Embryo score of a trait like intelligence might I know what you're talking about? Parasite, nuclear warheads, And one of those ads in the New York subway about picking building your next baby. Um, as they PTTP, I just know is offered over one hundred and seventy clinics in the US. And I just feel AI takes so this is on our list, but it's an entirely new class of this project. So, this tool is good or bad may have. 00:40:37 Speaker 1 A lot of data centers have been built for people to commute. We've been committing hundreds and hundreds of billions of dollars in capital, especially in these new ones specialized for AI. AI warehouses with specialized chips, new kinds of cooling systems, all the purpose of running and training AI models. OpenAI, Google, Amazon, Meta are all pouring hundreds. 00:41:07 Speaker 1 The consulting group estimated that fifty three trillion dollars spent on AI. The largest is going to require more than ten volts of electricity from a small city's work. My son and I are making bold investments in semiconductor companies, and Mark Zuckerberg has taken four or four shares in big semiconductor companies. Most of the world's natural gas is about the critical energy infrastructure to go out for around thirty hundred. 00:41:42 Speaker 1 Now, last year, Our AI reporters team looked through some excellent reporting that we were shortlisted for the National Media Awards on this question of just how much energy can AI chatbots use. We've got a little short video to kind of give you a bit more context on this. You've probably heard that AI uses a tremendous amount of energy, and then it produces a tremendous amount of carbon emissions as a result. But how much energy is that exactly? What happens when we query an AI chatbot or generate a video? And what happens as those queries pile up? To find out, let's start at the bottom and look at the impact of an individual prompt. Then, we can work all the way up to a nationwide network of data centers. Companies like Google and OpenAI keep energy usage figures closely guarded, but by looking at open-source models, we can make some informed estimates. Asking, a simple question to a small AI model might use the same amount of electricity as running a microwave for eight seconds. 00:42:42 Speaker 1 Not even long enough to reheat last night's dinner. But larger text models can use more energy than AI models that generate images. Video creators can be the most energy-intensive of all. Creating a five - second video with AI might use enough electricity to run a microwave for over an hour, but the high-definition videos that you see from leading AI companies likely use. 00:43:11 Speaker 1 Now, let's look at what happens when we put everyone's individual queries together. OpenAI has said that ChatGPT receives about 2.5 billion queries each day. All of the responses to those queries are generated in a data center, And data centers are often powered by particularly dirty forms of energy like natural gas and coal. Over the course of a year, 2. 5 billion daily queries could emit the same amount of carbon as driving more than 200 million miles. Miles to power AI. That's enough to circumnavigate the globe over eleven thousand times. Experts predict that in twenty twenty eight, The electricity being used to power AI could generate the same emissions as driving over three hundred billion miles. Over one thousand six hundred round trips to the sun and back. Ai is emerging, not just as a technology, but as a societal factor with its own carbon footprint. So where does. 00:44:12 Speaker 1 Where's that energy going to come from, and who's going to pay for it? What will it mean for our planet? To learn more, go to technologyreview dot com slash energy ai. That was James O'Donnell, one of our reporters who worked on that story. Yeah, and it's pretty eye opening what they found. So be scared. And actually the economy can not tell us, About how much uh how much they were uh. 00:45:12 Speaker 1 There's also a big question about who ultimately is taking the real approach that they need. Experts say many Americans think that maybe it's better to increase their income, but I think there'll be no relation to data services. 00:45:42 Speaker 1 Somebody, I haven't actually disagree with somebody being on the bookcase. Finally, It's always a good sign if you want to make it into your own story. It proves that you're really going to help educate your children about the future. It's always moving based on what we know now. Things can change, in fact, based on our experience they will certainly will So now is your turn. If you pull up slide over every year we invite our readers. 00:46:14 Speaker 1 That will be called the "elementary breakthrough." So, I'll talk to you about what we mean by that. And if you go through my next sort of presentation, so here we go. But, we can also do what she may be aware of : any kind of chatbots or systems now have reasoning models built in. We're going to talk about all that. Artificial rules : machines can make a rule on, which is probably something humans could learn in days or over weeks. 00:46:44 Speaker 1 You know, robots are companies like figure. I think robots about war, something like music. I don't really find that. I don't know the fact that we're looking for a way to do this. That's the thing with any kind of robot, right? Actually a lot more than I thought. There's actually a lot happening in the world of robotics right now, particularly with AI and generative synthetic data for that. And using LLMs to do things like what I showed you there on the world around us. So, we are kind of moving towards this world where we can actually have these systems that can understand what's going on around them and make decisions based on that. 00:47:44 Speaker 1 So we actually had an idea. They were like, "Oh, why are we doing this?" I mean, they asked us. So what we are now ten days from our project going live right now. We're walking in line at our own end of the next month. And that's got a big audience who can actually see it, But also just kind of take it to make sure that I was responsible for this and including five three. 00:48:13 Speaker 1 I am going to go out and look at my whole day. I am going to hang out and have a chat. So, yeah, that's what I've been doing. It's really fun about some really cool things. 00:48:47 Speaker 2 Ah, for you he only gave one idea of some things, 00:48:51 Speaker 3 Right? Yeah, it was interesting like that because when it's going fast, he changed two. Changed two? Of the ten or eight that were there, he brought in two new ones. Two new ones. Two babies and actually, he revisited the story of what he said about Colossal there. 00:52:14 Speaker 2 Once they.