Loading summary
Transcript52 lines
- [00:02]
A
Hi there, it's Jennifer Strong, and this week I thought we'd do something a little bit different. 2025 is now behind us, and as we move into the new year, it's a good time to consider what challenges, trends and opportunities will shape the tech industry in the months ahead, and what all of it means for our jobs and families and our everyday lives. I'm joined by some of my colleagues in the field to help answer these questions and more. Freelance journalist Jeff Wilser has written for the New York Times, Wired, Time magazine, and many others. He's also the author of eight books. London based editor Charlotte Gee is the news editor at the MIT Technology Review. And the Wall Street Journal's Robert macmillan writes about computer security, hackers and privacy from San Francisco.
- [00:45]
B
Welcome to the next innovation.
- [00:55]
C
Hey, my name is Jeff Willser. I am a journalist who covers AI and I host the podcast AI Curious.
- [01:04]
B
Hi, I'm Charlotte G. I'm news editor at MIT Technology Review. And yes, I also focus on AI a lot as of the last few years.
- [01:13]
D
Hey there, My name's Bob McMillan. I'm a cybersecurity reporter here at the Wall Street Journal.
- [01:18]
A
Hey, thanks everybody for being here. You know, it's the start of a brand new year and, and we're going to break down what we think will be some of its biggest trends and themes. I feel like we should kick off with agentic AI, which, you know, please roll your eyes if you need to, but also chime in with your own definitions, but basically a software that can make decisions and perform tasks to achieve goals on its own. Right, and roll your eyes because I'm sure all four of us spent most of 2025 talking about agentic AI, so now coming into 2026, but there are so many different things that we could talk about here, from identity to people's jobs to, well, trading. I'm sure we'll get into that. But let's start with cyber. Bob, would you agree that this is one of the key areas we'll be talking about with agents this year, both as potential helpmates, but also as our adversaries?
- [02:07]
D
It's funny, if you had asked me this question six months ago, I would have said this is, it's all hype and all that. The AI hacking stuff just seems to be producing a bunch of slop that's just wasting everybody's time. In fact, I was at one point thinking of writing a story about how the open source community has just been like, overwhelmed with these really crappy, useless reports and it was Just draining resources from people. I heard from a number of sources that it was just like a huge problem. And then sometime around the fall, the models got better and some of the things that started happening were like real, you know, like, very interesting to me. And probably the most public of thing that happened was a couple of months ago, Anthropic released a report saying that China nation state hackers out of China were using their LLM and to do some pretty interesting stuff. And you kind of look into that and it's sort of, I mean, we're obviously pretty far from where the AI is going to hack us or anything like that. But there are, there are basically tool sets that are being built right now. And it's not just by the, you know, by criminals, but by, by penetration testers and red teamers, people who make a living breaking into companies to show them their security weaknesses. They're building tools that allow them to do a lot of the steps that previously might have been done by a, by a human and chaining things together. And basically, I mean, what seems to be happening that's the most obvious, right, in the cybersecurity realm right now is that the cybersecurity practitioners are going to be able to do more for less, right? Like they're going to almost some like. And you see this a lot, I think, all over the AI space. Like it's, it's like they're getting an army of interns that are not like, terrible. And so what the implications of that are going to be is something I'll be watching in the year ahead. Like, it's a little unclear right now, but everybody seems to feel in the cybersecurity space that things are going to move faster, they're going to maybe get a little more out of control. There are going to be areas where there are exploration for bugs that maybe wouldn't have been accessible or easy to, to poke around at before. But now with the help of these agents, there's going to be a broader, what we call an attack surface. So there's just going to be more stuff that could theoretically be hacked. And so I think that when this kind of thing happens, it often gives an early advantage to the Red Team or to the people trying to break in. But what's going to, you know, what, of course, all the venture capitalists are excited about is that they can now build defensive tools that leverage the same phenomenon to sort of protect, you know, to get one step ahead of the bad people. So there's going to be a little cat and Mouse game, everyone seems to think this year.
- [05:09]
B
I mean, I'm actually watching the same thing. It is fascinating. You know, I agree with Bob that there's been a lot of hype around this and a lot of overblowing claims of autonomous hacking. I mean, AI is not, you know, going off and just doing all of the steps of a hack, but obviously it is kind of speeding up the operations. I mean, most scams now are written by AI, so it just means you can push a lot more things out more quickly. I mean, I think that in the world of AI right now, there's actually quite a lot of disillusionment. I think there are a lot of people that have got kind of almost kind of too hungry for progress, should we say, because we've seen so many kind of rapid breakthroughs in that field that I think the fact that the last couple of model releases were not absolutely astonishing has led to a bit of a kind of deflation in what people think is going to happen and questions about will the scaling laws continue to apply? Are we hitting a wall? Can we get better performance in AI? And I think this year is going to be a bit of a crunch year where we'll find out whether the answer to that is yes or no. And then I guess you also have quite a bit of what people call model overhang, where essentially people haven't adopted the technologies widely enough yet for them to have had a big impact. But I would say one thing I wouldn't underestimate is just how flawed a lot of this technology is. I mean, one problem that criminals and everyone else is running into is hallucination. Right? That's a pretty big problem. If you can't rely on these models to actually say things that are verifiably true, that really does limit their utility. So until we can start fixing some of those underlying issues, and I am not clear whether we can, I don't think the impact is going to be quite as astounding as people think.
- [07:01]
A
And Jeff, you cover all aspects of AI too, including crypto AI, which seems likely to be a major trend this year all on its own. But it's also providing kind of a sandbox for AI agents to be used in financial trading.
- [07:13]
C
Yeah, I view it almost as a possible sneak preview we might see for better or worse in mainstream finance. You know, for. As most FOL folks know, for. For years and years, crypto has far less regulation than traditional markets. And so it's a lot more room for experimentation. And so you have. I've speak to CEOs at startups who are creating these agentic AI traders. And they, as opposed to kind of old school bots, the kind of quant bots that just buy or sell based on math, they have more judgments to try and optimize a portfolio. And they are competing, I think is actually a swarm of agentic portfolio managers. Right? So you might have one portfolio manager who has programmed to have a certain kind of risk tolerance for the overall portfolio. Then you have an analyst agent who might cover stocks, analyst agents covering real estate or what have you. And this is all coordinating to try and optimize a return. And this is all still in simulation right now, but actually I believe they're either just started or will be starting soon to be trading real assets. And obviously Merrill Lynch, Fidelity are a long ways away from doing that kind of thing. But I do think it is a possible Wild west, that interesting kind of thought experiment being played out in this kind of real Web three world that might give us a sense of how, um, we could see changes down the road.
- [08:46]
A
Well, for myself, I'm going to keep looking at how agents are not just being deployed by companies, but also how they're essentially being employed as well. And this is to say, also I agree with everything that everybody said so far, including the potential for things to really go off the rails. But when things do go off the rails, how do you know? How do you pinpoint the agents causing trouble? How do you know what's working if you don't provide an identity? Which, talking about employing agents like people caused quite a stink last year. But what the alternative would be is not quite clear yet. So that's something I'll be watching for this year. We also can't really talk about AI without talking about chips, energy, data centers. So, Charlotte, what do you see as priorities to watch for in this space for the year ahead?
- [09:29]
B
I don't think that tech companies considered just how much opposition they would end up facing from local communities, partly because they're seeing increasing electricity bills, but also because people don't really like living next to data centers. I mean, that doesn't sound surprising to us, maybe, but apparently that's surprising to Silicon Valley executives. So I don't think that they really geared up for as much of a fight as they're now going to face. And I also think it's worth noting that this is an issue that's kind of uniting people across the divide. I mean, people are kind of mad about this.
- [10:04]
D
I feel like there's this whole sentiment in the AI world that gains will come, right? Like we're banking on that. That always has always happened in technology. But traditionally with tech we had this thing called Moore's Law which allowed us to increase the number of transistors every 18 months or something like that. And so there was this free ride for like most of my career reporting on technology. And you know, a lot of people make comparisons between the early days of the Internet and AI right now and this infrastructure build out. And I mean we can't depend on Moore's Law. Like everyone knows that that's, that's done. So I'm really curious about, maybe you guys know the answer here, but, but where the, the, the improvements are going to come. They seem to be from software architecture, but it's definitely not going to be from like shrinking the size of the transistor. And I just wonder, should we count on that? Like, is that something that we're going to be able to reliably deliver or not? I really have no idea. Maybe you guys know the answer to.
- [11:06]
A
That one now Bob, I'm glad you raised it because that is what I was going to piggyback off Charlotte to say next. We do have to consider both that and we do need to talk about water. I feel like it's not just the having in your backyard, it's not just the electricity conversation, but we've in the US built a lot of these or are building these data centers in places where there's not enough water in addition to not enough power. And so I do think that'll be one of the key things we talk about in the year ahead.
- [11:33]
C
One thing I think it's, I agree with all of that 100%. I also think that it's tricky though to measure some of these quiet gains that might be seen. Right. And so when I spoke to and worked with a company of regional banks, they kind of walk through all of the ways they have used and not necessarily cutting edge AI, not even agentic AI, but like basic stuff to improve workflows. To cut something that used to take 20 hours becomes two hours. I know it sounds very boring, but I, I don't know where that kind of gain would show up in a survey, would show up in a big economic report. And so yes, I think that we are looking at possible plateauing of like frontier performance and so on. But even if the models never improve one bit as companies and humans figure out how to work with them better, it's possible we could see some actual legitimates that benefits that are just tricky to Kind of measure with receipts.
- [12:34]
B
I just want to say I completely agree with Jeff on that. Like, I mean, it's becoming incredibly difficult to measure AI's progress because benchmarks, the traditional way that we do that are completely broken. I mean, these are sort of the, in theory, kind of scientific measurements of AI's performance, but they're just sort of gamified by the AI companies. They're not particularly reliable. No one really takes them seriously anymore. And it's much more about just on the ground, how are people finding this helpful? And that is much, much more difficult to measure.
- [13:06]
A
It also ties into what I think will be the other key trends in AI. For types of AI or sectors where we're going to see tangible progress. We already really are. One of them, I think will be advances in physical AI this year. That's things like robots and critical infrastructure. We forget actually how little intelligence has been put into robots up until this point and so how much room there is for growth. Medical AI I think is something we also have to continue to consider both what's being used by hospitals, also by doctors in a whole variety of ways.
- [13:38]
C
I think that this is so non glamorous, but hospitals are working to just cut down on red tape with AI. And there is, so, you know, there's, by some, by many standards, there is a medical or burnout kind of crisis amongst doctors and health workers. And to the extent that AI can improve some of the back office work, the boring. Oh, how do we translate this patient visit to insurance code 37B? Like no doctor goes to med school excited about doing insurance code stuff, right? And so I think that from my conversations with folks in the, in the healthcare field right now, like there's almost like two phases. Phase one is still very much kind of this like behind the scenes paperwork stuff. And it's, there are exciting possibilities of diagnosis and scans and, but even potentially like robotic surgeries and advancing on things in that front. But that's, that's still a little more down the road in most cases given the obvious high stakes and liability concerns.
- [14:44]
D
One of the things that I'm interested in this year is just observing society's tolerance and interaction with AI. Right. Like I live in San Francisco here where we have Waymo and a couple of weeks ago Waymo killed a cat and it was like a national news story. You know, I think the Times covered it twice. Like, and, you know, this is something where like a, you know, humans get killed all the time by cars. That's not a national news story. But you know, Because AI was, was, was connected to it then. People really cared about it. And I think, I just feel like, I think today we, there was a story about Gmail, you know, but by default analyzing our, you know, and there's sort of this growing sense like are we, you know, it's like the sci fi movie where like we're what's for dinner? You know, from the, from the entity that's coming. Like we're, we're sort of feeding these AI tools and they're going to replace us, you know. And so it's fascinating to me this is how the world changes, right? Is like these new things get introduced, they seem like really weird and like, I mean people still come to San Francisco and they, I see tourists all the time photographing the self driving cars, whereas I, you know, at first I was a little anxious about them. But I love them because they drive so carefully, you know, like I'll, I'll walk in front of one. No, no worries. Like I'm much more scared about drive walking in front of a human. But there's this kind of societal back and forth that, that's happening right now about a variety of issues relating to AI and just you talking about doctors using it. And I was thinking like, what is our tolerance for mistakes? You know, if a doctor, an AI, you know, nurse makes a mistake is that, you know, like mistakes happen all the time in the medical profession. And you see people sort of proponents of AI arguing that we're going to actually reduce, if you objectively look at it, we're going to reduce the number, but we don't know what's acceptable risk in this, in this world. And we might have different standards from these products than we do from humans. And I just find it fascinating to literally see that negotiation happening on the streets of San Francisco every day.
- [16:56]
C
I think Bob is spot on though for next year that there will continue to be this kind of, we're navigating this new conversation and kind of there might be some fault lines on if people are pro or anti AI and that will become more and more front and center. And kind of Bob's point, I totally agree that there might be a mismatch between the actual objective risk and what feels like risk. Right. And so as folks get like have this, this almost this, this allergy to AI when they hear something has AI in it, it might be that, oh, I'm not going to buy that product because there was a bit of AI in there or I'm not going to see this movie. I'm not going to consume this thing. And other, other folks might be marketing aggressively. AI So I think, you know, we're these. It's still developing our, our societal kind of taste or anti AI ness and I think that will kind of begin to crystallize this year.
- [17:54]
A
Yeah, it's kind of funny, I keep going back. Who knew this goofy short story I wrote from CES a lot of years ago now would still be coming back to me. It's for AI to be useful, it needs to become boring. And a lot of what we're saying here is like, well, these things that are moving are boring, but that's kind of what we want, right? It's like it means it's working. Nobody calls and says, yeah, the same thing happened 200 times today. And it, you know, it just works. But it's fundamentally the thing that moves the needle and it is kind of hard to track to that end. I think other spaces we might see movement this year would be smart construction, right? Digital twins, often drones in construction site can keep track of materials. 20% of fatalities on the job in the US happen in construction. So if we had a better way to model and track all of this, that could make this job a lot safer. It also just means less waste, things can move faster. And it's not new. Like again, boring. Digital twins. What we've been talking about this for a decade, but it's. Some of the promise of this is starting to realize and we're starting to see some acceleration. I think we're both there. And then also within agriculture as well.
- [18:56]
C
I think in adjacent, adjacent spheres that do show some, some real promise and that are actually being implemented are things like in recycling, right? So AI being used to help instead of having just two or three buckets of various recyclable components, say we can now go, I'm making this up. But 9 or 10 or 11, so recycling more effectively. Building management, right? It's these complex office buildings. It's actually there's a million variables going into the proper way to heat and cool and air quality and so on. And so it's AI systems that can optimize that to help building this cool more effectively, which takes less energy. Now, whether that offsets all of the water being used energy, that's not for me to say. But there are, I think, legitimate areas of progress in the physical world. Jennifer's mentioned similar to construction.
- [19:51]
B
The area where I'm seeing like massive amounts of growth and excitement, for want of a better word, is in defense tech. I mean, European countries are currently grappling with a very real crisis of having to potentially fend for themselves without as much kind of assistance from the U.S. and that's that, you know, this is a kind of massive post World War II shift. And then realizing they're going to have to spend a lot more, they're going to have to confront future threats, they may have to, you know, fight against Russia, I don't know. And we're seeing all these kind of threats evolving very rapidly. And yeah, I mean, I think that that's unleashing a huge amount of innovation in military tech, some of which is frankly really quite scary. I mean, I saw something about a drone system in Ukraine that's being developed at the moment that can essentially just follow, hunt down and shoot someone autonomously. I mean, this exists. How do we feel about that? And also, you know, I hate to say it, but we're living at a time where a lot of the institutions and kind of things that we thought would constrain countries from doing things appear to be falling apart. And it was all, maybe it was all a mirage held together by goodwill. That doesn't really, we don't feel like we're in a world where that exists as much now.
- [21:06]
D
How do we feel about robot armies?
- [21:09]
B
Yeah, yeah, I mean it's, it's genuinely, I mean, obviously it's like, well, it depends whose side they're fighting for. You know, I mean that's, I guess that's the answer a lot of people, people would have, wouldn't it?
- [21:19]
A
Bob, what are some of the more interesting things that you think you'll be watching this year?
- [21:23]
D
Well, I mean, we've talked about the sort of offensive capabilities of AI, but we haven't really talked about what I think could be the biggest problem in all of this. Like, we're adding a lot of complexity to the world. And if you write about cybersecurity, you quickly learn that complexity is the ally of the hackers. Right. And so we already have all, we have something called the Internet of Things, which is just a big festering cybersecurity hole. And now we're talking about adding artificial intelligence, which has sort of these like unexplainability problems in it and is also just like the subject of like intense nation state espionage level scrutiny. You, you kind of wonder like everybody's kind of pushing this stuff. Are we really being thoughtful about the security model around these agentic systems? Right. Like, so when I hear talk about a therma AI powered thermostat, I think like, oh, I wonder how that's going to get hacked? Like is it going to be hacked? And we're also seeing that the interface toward hacking is becoming natural language, right? Like the interface is changing from traditional computer like Unix shell type stuff where, which required some level of technical sophistication to something that just anyone can do. So I think that there's, you know, there's this weird debate about alignment that's happening that seems very separate from the debate around cybersecurity. And I think that the, the, the, the AI companies get this like they, they are thinking about it, but if you take the time to produce a model that has the guardrails that will prevent like implementations of it from being hacked, you will lose the race to have the best model. Like I, that's just a, that's just a maxim from like watching technology over the last 30 years I've been reporting on it. So there's, there's like not a great incentive to build a secure product when you could build a fun and fast moving and you know, delightful product. And so I'm kind of worried about that because we have really screwed it up with, with the Internet of things. We have a lot of fast, cheap and out of control devices out there and they're causing a lot of problems for people.
- [23:57]
C
I think that we started to see this year AI and kids and whether kind of companions or whether AI injected toys and we've seen that go sideways. And my guess is we will see more and more concerns in the coming year as these quote unquote companions and we've seen from stats that kids are using them even if they're not supposed to be on our team, they are. At least there's a good chunk of early teens are using these quote unquote companions and friends and we're only just starting to see what the impact of it is. Right? So if, if 9 year olds, 11 year olds are growing up where their best friend is not a real person, how does that impact their development in social skills? And I think there, that's, that's kind of like a, amongst AI nerd circles being discussed, but it's not really broken through in a mainstream like the average parents thinking a lot about this and we're not seeing kind of like a widespread kind of congressional inquiries about this. But I think down the road that feels inevitable this because as the tech improves, as more kids use it, as we see more very concrete harms, I imagine this will kind of snap into the spotlight.
- [25:14]
D
I'm so curious about whether there's going to be like a pushback against tech. Right. I think like over the last five years as a parent I've seen just a growing, just like blah about technology, about being online kids. Kids, you know, my daughters don't, you know, they, they have a lot of skepticism about these big tech companies. They, they realize that the business model is not in alignment with their mental health. And you, you read these stories about people who go back to like pre iPhone type phones and you know, I, I have digital sabbaticals and stuff like that and I, I just wonder if that's, that has any legs. Like if there's going to be just a generation that's just like oh my God, you guys were so brainwashed by tech. You know, we, we, we, we, we don't want anything to do with Instagram or TikTok or whatever, like, you know, because, because there is like a sort of generation Alpha, like what, what are they going to make of these Gen Zs and their, you know, their obsessions?
- [26:17]
C
Well, on one, on one encouraging note, I recently spoke with the head of research at the Alan Turing Institute and research AI in kids and she's studying for, you know, over time, children's and teenagers thoughts on AI. And what she told me is kids are often asking kind of the hard questions and in many ways they're asking, they're, they're more savvy than most of us adults are about asking about hey, how is my data being used? Where is bias creeping to this? So I, I do think kids hopefully are, seem a little better equipped to ask these questions and kind of stand up for themselves certainly more than when I was a silly and naive kid.
- [26:57]
B
In a funny way I think like AI Sloth might be the, the kind of death knell for social media. It, it doesn't feel like you can believe anything you see. And that's, that's really quite profound.
- [27:07]
C
Yeah, I was just going to go there. Charlotte. I think it's really well said and I think that as the, as we see more one kind of irony I think is the word slop has of course become ubiquit ubiquitous. But as it gets better and better, the irony is it looks less like slop and it's actually now pretty good slash, really good and as it evolves will be exceptional and easy to fool everyone. And I, I, I'm very curious about, I think it's surprising almost we have not seen anyone in any effective massive mainstream way fully weaponize this to cause massive societal harm. And it feels like only a matter of Time until that happens. Right. Whether it's like the, the classic example I remember is in a presidential debate, if someone creates this deep fake video of one of the debaters getting saying something untoward off camera or what have you, and that caused the entire perception of the debate to change. And even if we find out that's a deep fake, doesn't matter. Cat's out of the bag. And I'm surprised there has not yet been a true weaponization of slop in deepfakes. And as it gets so easy to make to Bob's point, I just want, I'm kind of just nervous of how will a society react when there's no objective truth anymore, when videos are not, are no longer sufficient. Just we all agree, okay, this happened, this is real. How does it change the world? I find that deeply unsettling.
- [28:35]
A
We all cover lots of different types of tech, right? So I understand that AI is now eating up most of the oxygen in every room that we enter. But are there other things that we would be remiss if we didn't bring up like, quantum or.
- [28:48]
D
Well, we would be remiss if we didn't bring up quantum because it's so, it's so cool. Right? Those, those quantum computers are like sub zero, you know, zero degrees Kelvin. And you know, they, they look cool. They do. Amazing. They're going to crack the encryption that we know. But the thing about quantum that is, is interesting to me just like with AI is like, what is the path toward widespread use of it, Right? Like, it's just so expensive to do quantum computing and the, it's like, kind of unclear whether the applications are actually going to work or even be interesting. I think for like encryption cracking, like, it's pretty cool. But like, right now it's, it's hard to imagine that happening in a widespread way. But you, you know, you, you, you think that the, the nation states probably have a way of, you know, breaking, you know, advanced encryption algorithms and it's going to be super secret, but they might just do it for a small amount of.
- [29:46]
A
Although as I said this, and I'm kind of taken back to early Internet dates, we're like, wow. Yeah. I mean, the reality is that all of the, whether it's synthetic biology or quantum or any of these other things we might choose to talk about, they're going to be, they're going to converge or be integrated with AI anyway, much like the early Internet days when we say, wow, all we talked about was the Internet this year.
- [30:03]
C
Yeah. So I think you mentioned at the top of the show, robots. And again, there's obviously a component of that I think we'll be seeing whether we like it or not. Maybe we'll be rolling our eyes a year from now. And oh my God, there was so much robot talk in 2026. And I think that, you know, that what captures the imagination is humanoid robots, but more functional robots in the development of these that do things like, you know, clean our bathroom or clean our toilet, that kind of stuff might be, I think, begin to emerge and then totally other topic. But I think that's what is AI related, what is not clean energy and nuclear. And I think that it's potentially one silver lining here, at least in my personal opinion. As these AI companies are realizing, oh my God, the energy needs we're going to have, we have to think more boldly about energy sources and developing more nuclear energy. And kind of quietly, the cost of solar continues to go down. And will we see, I'm very curious to see, will we see any real needle movement of continued solar development and deployment.
- [31:11]
B
I was also going to point to kind of climate tech. I mean, there have been amazing advancements. We're seeing kind of different types of nuclear reactors, different types of batteries. I mean, the world has all the tools it needs to address climate change. It's just question of whether we've got the will. So, I mean, like, there are huge positives to take away from that. I also kind of want to point to health. I mean, despite what some people seem to believe, we literally have never been healthier than we are now. At least in the Western world. We're living for longer, we're living more healthily. And you see things like the weight loss drugs which have come out, which I know, you know, contentious, but the fact is they have delivered great results for a lot of people. And we are finding that they're also good at protecting people's hearts. And so I'm really optimistic about the frontier of health and how we can improve that. And I don't want to mention longevity on that list. I know it's getting a lot of funding and a lot of excitement, but I think it's more just like better understanding our diets, our movement, and the kind of basics that. That really excites me.
- [32:13]
A
I feel like that we should talk about jobs, right? We should. We would be remiss if we didn't talk about jobs. But also it's the area where there's not a ton to say, which may be why we haven't brought this up. Like, are any of you seeing giant efforts towards reskilling or anything else that I'm not seeing right now?
- [32:28]
C
I'm a maybe a little more bullish in this room, virtual room here. As far as AI's ability to help businesses. However, I think that means I'm even a little more bearish on impact to jobs. And I actually do think that as this is, as companies figure out how to more effectively integrate even the current level of AI into their workflows, it seems inevitable that it might be that, okay, we're not going to fire anyone, but natural attrition, we don't have to replace Frank and Jill because we've already automated this workflow. Right. That feels like, and maybe down the road when dust settles, when reskilling happens, we'll all be fine as an economy, but it feels inevitable there'll be a very bumpy transition down the road and it's quite euphemism for a lot of jobs lost.
- [33:17]
D
At the risk of burying everything in obscure cinematic metaphors, I think this is not like the Invasion of the Body Snatchers, but more like the Bionic Woman, you know what I mean? Like, I think that AI is, is definitely making me more productive and more powerful at my job. There's no question about is like having an intern, you know, sometimes just at my disposal. Which makes me, does make me wonder how things are going to be for the interns, you know, and the ramping up of skills and the pathway to doing that. But it just seems like, you know, with, as we become more powerful, the world becomes more complex and so there's more complexity to deal with and it's just like a never ending thing. So I don't think the humans are, I don't see them leaving that equation are sort of important but, but they're, they're just going to be able to do more with less.
- [34:05]
A
But I want to ask each of you what's something we have not talked about yet? What is a major trend that you'll be watching?
- [34:11]
C
AI companions. I think that as they get more refined and personalized, they will get more addictive. And I don't just mean romantic companions and that whole kind of weird niche, but even like your chief of staff, the person says, oh, your own personalized Walter Cronkite here was the news yesterday. I think we'll be seeing more and more of that and all of the pros and cons that entails.
- [34:36]
B
I would throw in increasingly weird and wonderful ways of finding trading data for AI and robots, which might sound really Esoteric, but we're seeing things like people being employed by robotics companies to open and close their dishwasher a thousand times. It's, it's a really weird world. We're kind of turning humans into robots so they can train robots so they can do job. So it's kind of a strange time. But that hunt for more data, better data, I think is leading us down some kind of crazy paths, which I will enjoy watching through my fingers this year.
- [35:12]
C
As these AI companies are running out of low hanging fruit of data, there will be this thirst for, for kind of hoovering up more and more data. I think that is going to quietly already happening. That will become more of a story. I go to kind of nerdy AI conferences where literally I'm seeing ball cap saying data is the new oil, right? And so as, as AI companies are, you know, thirstier and thirstier for data, I think that will have all kind of economic consequences. We're just beginning to get our arms around.
- [35:42]
D
Well, as a employee of a publisher, I think I would be remiss if I didn't talk about the way the business of publishing is changing and the, in fact the way the Internet works is changing, right? This, this sort of recalibration to a post search engine link, you know, kind of world to, to one where we just ask questions and get answers. I think that that's definitely affecting people like me. And it's, it's, it's sort of like unclear what the business models in the age of AI are going to be.
- [36:24]
A
All right, everybody, this has been so much fun. Thank you so much for taking time out of your days to join us here.
- [36:29]
C
It's a lot of fun. Really, really enjoyed. Next year we can have our digital twin avatars do this together and we'll.
- [36:34]
B
Be, you know, sounds like a plan. Thanks guys. That was great.
- [36:39]
D
Now is the time for me to reveal that I am an AI. Just kidding.
- [36:43]
A
This guy.
- [36:43]
D
I'm real.
- [36:45]
A
All right, thanks everybody. Thanks for listening to the next innovation and thank you once more to my guests, Jeff Wilser, Charlotte g. And Robert McMillan. This episode was produced by Situation Room Studios. Christine Barata is our executive producer and Sharon Barreiro is our senior producer. Leila Shirawi is our associate producer. Additional production support by Global Science Situation Room. I'm your host, Jennifer Strong. Until next time.