← All TWiT.tv Shows (Audio)
All TWiT.tv Shows (Audio)
Tech News Weekly 417: Smart Home Year in Review
Looking Back At Smart Home News
Loading summary
Transcript65 lines
- [00:00]
A
I'm Micah Sargent from Tech News Weekly and this week I'm joined by Jennifer Pattison Tuohy of the Verge. I kick off the show by talking about a great piece from CNET about the current state of AI health coaches. Then Jennifer Pattison Tuohy has a wonderful look back at the year in smart home news and devices before Allison Johnson of the Verge stops by to talk about how AI image generators have changed over time and how worse is actually better. All of that coming on Tech News Weekly.
- [00:35]
B
Podcasts you love from people you Trust.
- [00:40]
C
This is TWiT.
- [00:44]
A
This is Tech News Weekly episode 417 with Jennifer Pattison Tuohy and me, Micah Sargent. Recorded Thursday, December 18, 2025 Smart Home Year in Review hello and welcome to Tech News Weekly, the show where every week we talk to and about the people making and breaking the tech news. This is our last official episode of the holiday season. We've got a best of coming up, so I am very excited today to be joined by our wonderful co host. It's Jennifer Pattison Tuohy. Welcome back Jen.
- [01:21]
C
Hello. Thank you. It's been a while. Sorry I wasn't here last month. I was in the rainforest so I missed you. But I enjoyed cars.
- [01:29]
A
Yeah, I hope so. I hope so. It sounds like a lovely time. I'm really excited. Jen has worked to put together some kind of year in review stuff, so we're going to have her go second. So I'm going to kick things off with my story of the week. But welcome one and all to this episode. Let's get into it. The promise of AI powered health coaching sounds, well, almost too good to be true. A tireless, especially judgment free companion that can interpret the mountains of biometric data our wearables collect and turn it into actionable advice. But as Vanessa excuse me, but as Vanessa Hand Oriana explores in her comprehensive piece for cnet, the reality is far more complicated. After testing AI health coaches from Google, from Apple, from Samsung, from Aura and others over the past year, she's arrived at a sobering conclusion. These tools aren't quite the game changers they're marketed to be, and the privacy trade offs may not be worth the payoff. It's a story about the gap between technological ambition and human behavior, and what happens when we invite AI into our most intimate health decisions. I found this a fascinating read because I am definitely I fall on the side of sort of. I love to track different things. I have a bed that keeps track of my sleep, but I also have the Oura ring that keeps track of my sleep and my movement. I have an Apple watch that's also keeping track of a lot of that stuff. And I have used Bluetooth connected blood pressure sensors and all sorts of stuff. And I love this sort of quantitative data that I've collected over time. But it is true that a lot of times this stuff exists in its own little silos and there's not a whole heck of a lot that you can do with it if there's not something there analyzing it. So that's where the promise of AI comes along and says, well let's look at trends, let's look at when you do this, then this happens. How is food impacting your ability to sleep, how is stress impacting your ability to function, etc. Etc. So that's kind of what the promise is, but the outcome still not quite there. We're going from though predictive AI to generative AI because it's important to understand that with wearables. For years it was kind of this idea of a wearable AI like high heart rate alerts, fall detection, sleep scores. I would get something that was like oh, over the past six months you've moved a lot more than you did in the six months prior. All of that stuff has been kind of trained to identify a particular type of event. And because of that they can be kind of precise and predictable. They're there to focus on something specific. So they're looking, they're pattern matching to a very specified extent. But generative AI health coaches are using large language models and that makes them more dynamic, more responsive but occasionally can get a little weird because of the confabulations and hallucinations. So I wanted to kind of talk to you. Well actually one, one quick thing I'll say in this piece it was looking at kind of the big, the big companies Google testing AI coach inside of Fitbit built on Gemini. Apple of course has workout buddy for Apple watch that has real time audio, motivation, Samsun, Garmin, Aura, iFit have all rolled out AI features and then meta partnered with Garmin and Oakley to embed meta AI into smart workout glasses. Have you used any of these trends? Do you have like, would you consider yourself a sort of health tracking person given your smart home love? I'm curious if that kind of bleeds into the wearable and health tracking space for you.
- [05:25]
C
I haven't embraced it fully. I do have an Apple watch and I have a sleep number bed so that does a lot of sleep tracking for me. I had in, I had used like I Think it's why things had the sleep mat that I use for a while. So I've done a. I've used this technology more to address problems than necessarily to kind of evaluate my overall health. And that's. So yeah, I'm not, I'm not one for quantifying my health mainly because I don't. I'm a bit lazy about that, about my health. I should probably be better, but I'm a mum of two and I'm a work full time and like my, my priorities, I fall at the bottom which you know, is probably not great, but I have tried. So I do use an Apple watch and I, the AI coach, I turned on and so it has been popping up when I do my occasional exercise and I find it, I haven't found it useful. Yeah, I can see the motivational side being somewhat helpful, especially for someone like me who's not very motivated. I also find what's interesting about the AI and what I think is, is going to hopefully be valuable eventually is that interpretation because having so much data, I mean I have collected a lot of data even though I'm not very active with all of this and it would be nice to have something to do with that data, have someone or something analyze it. But this article kind of pinpoints an issue I think we've seen across technology about generative AI is that its promise has been so much more than its actual current potential and you know, nothing's really panned out the way it's been promised yet. We're still kind of waiting for that magical oh wow, this really fixed a problem for me. And this is something we've seen in the smart home with AI as well. It's like it all sounds like everyone's hyped it up so much like this could do. This could transform your health tracking, this could transform your smart home. This could transform your workflows. And so far it's just a sort of the friction is still there. I don't feel, I really don't feel like anything has broken through in any meaningful way. And as this article points out, it has exacerbated the privacy and data collection issues that we have around technology.
- [07:56]
A
Yeah, absolutely. And here's I, I want to kind of talk about in this piece where they kind of see the, the, the best case scenario where it's doing a good job and it does seem to be that it is like it's ideal is to fill gaps in care. So. Dr. Jonathan Chen, professor of Medicine and Director for Medical Education in A&AI at Stanford says that there is some potential here because of the overburdened, expensive US health care system, and says we already have ways to collect data for people all the time. But even your doctor doesn't know what to do with all that data in the icu, let alone all the wearable data. And so it could help to synthesize this information, flagging warning signs of conditions like hypertension, of course, helping patients arrive at appointments with patterns and insights rather than raw data dumps. So then you could come in and say, here's some stuff that you know that has popped up. Is that helpful to you, Doctor? And in fact, Oriana talked about this idea firsthand. A family member's apple watch detected irregular heart rhythm. Never appeared during routine doctor visits or weeks of wearing a clinical heart monitor. But because the watch is always on and always, well, not always on, but it's on regularly and over a longer period of time, it flagged an episode in real time leading to a confirmed diagnosis. So the ECG and in the heart procedure, it ended in a heart procedure, reduced his risk of a life threatening event. But there's also false positives, there's over indexing, and so you get situations where these little deviations start to add up and they actually don't mean much. So there's a lot there. You know, it's a back and forth for sure. And I think it's, oh, what's happening? I don't even know what's happening.
- [09:50]
C
Speaking of wearables, I just got some smart glasses and it just started talk. They just started talking to me. I don't know if you can hear it or not.
- [09:58]
A
I just barely heard it. I just barely heard it. I thought.
- [10:01]
C
I'm so sorry.
- [10:02]
A
That's okay. I thought it was probably Google device. Yeah, exactly.
- [10:06]
C
No, it was my glass. I just got them. Yes. Sorry for interrupting the train of thought there, but it's a very good point. And I also notice, I think one of the issues there though is when I've ever tried to prevent present any of this information to my doctor, I'm like, oh, my watch said they kind of roll their eyes a little bit. I think there has to be. There's a point where there may be a point where if it becomes more embedded into the health system, that doctors will pay more attention to it. But right now I think it's largely used more anecdotally. I mean, obviously there are cases like this where, oh, wow, actually this did pick something up. Great. But overall it kind of gets greeted with the same reaction from my doctors when I say I looked this up on the Internet.
- [10:50]
B
Yeah.
- [10:50]
A
I was gonna say I feel like I could, and this maybe sounds a little arrogant, but I feel like I could do a seminar or like a sort of class on advocating for oneself in the, in the health space. Yes. And I know that part of I'm a man and that already is going to get me more, you know, get me through the door a little bit more. But I have had situations where a doctor is not listening to me about something. And so I came with peer reviewed articles and other information, was like, here's what I'm seeing and here's. And they ended up, it was long story, but they ended up sending it off to the doctor that specialized in that. And the doctor was like, yeah, I think that is what's going on with them. You should do this. And then the doctor, my doctor was like, oh, okay, yeah, let's do that. And it was what I ended up needing. But I, when it comes to that kind of data that you're talking about, I never say that it came from my watch or it came from. I'm like, yeah, so I did an ECG and it's, you know, to be so there. It's. I hate that you have to be careful about how you're delivering this information. In some ways. At the same time, I can also understand that these doctors are seeing lots of people over the course of the day, over the course of the week, over the course of months that come with so much stuff that may or may not be valuable. And so it's it, it results in kind of a boy who cried wolf skepticism built out of like a desensitization. So ultimately, the way things stand right now with this AI health coach situation is it's the trade off I don't feel like is quite there. The idea of giving up one's privacy in order to have actually good, actionable data is not there. And I'll be interested to see. You know, there have been lots of rumors about Apple really hitting this space more than any of the others in terms of AI health. And given that I keep all of my data in Apple's health app and it comes in from all the different places, from my bed, from my watch, from anything else that I'm using. I don't have one of those toilet. Those toilet things that are analyzing that. Yeah, exactly.
- [13:18]
C
Lots of those. I'm sure we'll see lots of CEs this year too.
- [13:21]
A
Yeah, I bet, I bet. But I think that this is Fascinating. Now we're going to take a quick break. We're going to come back. I'm really excited about this next segment. So much so that I have double booked for Jen the next two segments. But let me tell you really quick. Very excited. Always love to talk to you about Aura bringing you this episode of Tech News Weekly. So here's a question for you. Wouldn't it be great if you could change the photos on your walls every day? And I don't just mean slapping a TV screen up on the wall and having it change. No, I mean what looks like a photograph on your wall but changing every day? You can now with Aura Ink, Aura's all new cordless color E paper frame. Meet Ink, Aura's first ever cordless color E paper frame. Featuring a sleek 0.6-inch profile and a softly lit 13.3-inch display, Ink feels like a print, functions like a digital frame, and perhaps most importantly, lives completely untethered by cords. With a rechargeable battery that lasts up to three months on a single charge, unlim storage, and the ability to invite others to add photos via the Aura Frames app. It's the cordless wall hanging frame you've been waiting for. In fact, Aura just released a new bit of tech where people can, if you give them the ability to can text photos to the frame. And I gotta tell you, I was kind of preparing myself whenever I first heard about this because I thought, wouldn't it be super cool if there was an E ink color frame that I could do photos on? And then shortly after that, Aura, you know, arrived at my doorstep with this. And at the time I was thinking, okay, I'm gonna pull this out, I'm gonna put a photo on it and it's gonna look okay. But it's E Ink, so you know, just prepare yourself. And then I loaded the photo on there and I said, wait, hold on and sort of got close looking in. I thought, this looks so good. I was really surprised. It's a breakthrough in E paper technology. They have special dithering algorithms on epaper technology already has to make these look so good. Ink transforms millions of tiny ink capsules into your favorite photos, rendering them in vintage tones. Serious design innovation involved with a graphite inspired bezel. The paper textured matte around the outside and the glossy glass front make it look like a piece of decor, not a device. You get unlimited free photos. That's great too. You're not worried about some subscription thing that lets you have photos or not have photos. You just download the Aura app and connect to Wi Fi. And honestly, it's the perfect gift gift. Aura, Inc. Is the perfect gift for anyone who appreciates innovative design and cutting edge technology. I mean, I've got this photo of a recent camping trip with my dog Henry in his little backpack. You know, I don't know if he was having, he was, he was having fun sunbathing there. Whether he was having fun on the camping trip, it's another story. His sister loved it. Him, not so much. He's an indoor dog for sure. But sleek, subtle, and stunning Ink blends the warmth of a printed photo with the versatility of an e paper frame. No chords, no fuss, just your memories beautifully displayed wherever you want them. Head to auraframes.com Inc. To see for yourself. Support the show by mentioning us at checkout. That's auraframes.com Inc. Act now as they are offering a limited time holiday discount ends soon. Seriously, you gotta go there even if you don't go for the ink. Aura has some really great frames that have long been lauded as some of the best in the business. So check it out. Auraframes.com Inc. Thank you, Aura, for sponsoring this week's episode of Tech News Weekly.
- [17:10]
C
That was such a cute picture. Isn't it.
- [17:15]
B
Adorable?
- [17:16]
A
He hasn't been on like out in nature nature since he was a young pup. And so I was kind of nervous, you know, how's he gonna handle? And for the most part, he did enjoy it. But he definitely was happy to have that backpack thing. Cause he's like, I'm worn out. So I'm very excited about this next section. I mean, when you're joined by Smart home Mama, you do have to kind of ask, can we please do a look back at the year in smart home technology? So between now and 45 minutes after the hour, we've got all the time to stretch out and talk about smart home tech. Jennifer Pattison Tuohy, Take it away.
- [17:59]
C
Yeah. Well, I was excited to be asked to do this because there's been a lot of smart home news this year. It's been a big, big year in the smart home. And actually one story that has, that I've been covering all year kind of wrapped up this week. So it is quite relevant for a weekly news show. And this is the news of iRobot's bankruptcy, which was, which is big news this week, but actually has been big news all year as we sort of watched this. The original robot vacuum company, they developed the Roomba, which has become literally the Kleenex of Robot vacuums. Everyone calls their robot vacuums Roombas, whether they are or not and which proprietary eponym.
- [18:45]
A
For those who are curious about the term. I love that term so much.
- [18:49]
C
That's a good one. I haven't heard that interesting. Yeah, so they have. They've been facing a lot of competition. And in my world, I spend a lot of time testing robots and robot vacuums. And my house was getting more and more full of robots as more and more were being released and poor. And Roomba has really struggled in the face of Chinese competition. Also some competition from other. Other places, not just China, but China has become the dominant space this. The dominant brand space in this space. Sorry, a lot of robots coming from China, in short, and Roomba is. Irobot is an American company based in Boston, Massachusetts, and it was the original manufacturer for the Roomba. And the Roomba was really the first commercially successful robot vacuum. There was one before them, but it didn't go anywhere. So what happened just in the last week is that they. After the whole year, they've basically been warning that they may go bankrupt. They. This year, this week did. But during this year, they also released eight new robot vacuums. A whole new line, actually nine or ten in the end. But they started out with eight new robot vacuums in March. And it was like, wow, okay, maybe this company's gonna turn it around. But it's. It's been really kind of a tragic tale to follow this year, seeing this once very successful, innovative company that everyone loved. Roombas are, you know, the star of many cat videos on the Internet. I've not been able to get. My greatest tragedy as a smart home reviewer is that I've not been able to get my cats to ride on any of the robot vacuums that I have. But that's. That's a side note. But, yeah, seeing this company sort of fall has been really quite sad. However, there is good bad news. If you own a Roomba, it's not going anywhere. Unlike the Nito robot vacuum, which also shut down completely this year and no longer works with the app. IRobot says it's going to continue. There is going to be no disruption to your Roomba. It has actually been bought by a Chinese manufacturer, which is really a kind of bad end to the circle. And one of the things that really caused iRobot's biggest problems this, well, in the last couple of years was its attempt to be bought by Amazon was blocked by the EU and the American regulators, or at least they said they were going to block it, we never actually got to the point where it was, went through any formal process, but that really caused the company to struggle in the face of both its competition and its lack of being able to sort of innovate while it was under this scrutiny of looking at trying to be bought by Amazon, a lot of people were not happy with the idea of, of Amazon owning the most popular, if not the most used robot vacuum at that point. Because by that point roborock, the Chinese, one of the big Chinese competitors, had pretty much overtaken Roomba as the most popular robot vacuum sold.
- [22:09]
A
So yeah, it's been, it was also, if I remember correctly, the most data collective company too. Right. I mean they had the, the home scans before anybody else and I thought that was a big part of why Amazon, they were worried about Amazon.
- [22:24]
C
So one of the, so that was one of the reasons that Amazon wanted to buy iRobot, I believe. Personally, I don't. They never specifically, they never said this publicly, although iRobot did say one of the things that was part of our offering for Amazon was our iRobot OS. So this was what Roomba was doing a little differently than the rest of the competition is it was trying to create a sort of smart home platform, so not sort of help understand your home for your other devices in your home to also know where they were in your home to give context. And this was something that Amazon was really interested in. But yes, to your point, privacy advocates were concerned that Amazon would have this data about your home. Now Roomba does map your home, but you, and they keep, they say they keep that data private, but yes, they do have a lot of data on you and that was a concern amongst regulators about privacy. So yeah, there were a lot of things that lined up against Roomba, but actually a story that I just published this morning, talking to the newspaper Newish CEO, he's been there a year and a half, Gary Cohen, he says that the company is going to continue as is and will continue to produce new products. So 2026, this will probably be a story for as well. Wow. But yeah, it's, it's, it's really sad that iRobot's been around for 35 years. So. And it's really, it's a new company as of, you know, next year. So this was sort of the end of an era for the smart home, I think, which was if you can't.
- [23:58]
A
Beat a them, join them by the end of it.
- [24:00]
C
Exactly. That's kind of what happened. They completely retooled their line and now that, to me, they just kind of look like clones of all the other robot vacuums out there, which is a very ironic because all the other robot vacuums are started out as clones of Roombas.
- [24:15]
A
Wow. Yeah, that. Yeah, that's kind of a weird. Yeah, it's like we're doing what you're doing and then they're like, oh, but now we're doing more and you're not doing.
- [24:24]
C
Not doing more. Yeah, yeah, like robots, these. I mean, we've. I've talked about this on the show many times this year. Like, robots have got. Robot vacuums got crazy this year. We got arms, we got legs, we got all sorts of weird things because it is, It's. It's one of those rare smart home products that has gone really mainstream. You know, everyone has. Not everyone has a Roomba, but everyone knows what a robot vacuum is. You know, they know whether they want one or not. But a lot of, you know, regular people, not just techies, are really interested in the robot vaccine vacuum space because it does. It has a great utility and a great purpose for your home. Kind of like video doorbells as well. Like, you can see the point of having this kind of technology in your home and even if you're not into tech, so to speak. So, yeah, it's been an interesting space to. It's been interesting space to follow this year. And I've. It's also kind of an interesting, I think, trend to onto my other stories of the year, which is centered around AI Shocker. Because I came to the smart home just like it has to every other space here. But what I. The sort of the connection and I'm losing my hat. There it goes. Bye hat.
- [25:37]
A
Bye hat. Goodbye Christmas.
- [25:41]
C
I've still got some sparkly things on here, but there we go. The other. The connection here, I think is this tension we've seen this year between what tech companies think we want in our smart home and what consumers actually need, are ready for and potentially will embrace. So I think this is something. And this is what happened with, with iRobot, I think, is its founder, Colin Angle, was incredibly innovative and tech forward, but very much pushed his vision in. Onto the product and, you know, spent a lot of R and D and resources into making something that he thought was going to be really impactful as opposed to. To just making a really good vacuum cleaner. And, you know, what consumers want really in the smart home are products that do their job and do it well. Not that Roombas didn't do that, but I think quickly the competition Sort of outstripped them with some just basic functions. And this is what I'm seeing also with AI in the smart home. So the biggest news really in the smart home this year was the launch of the generative AI voice assistants. SO plus arrived and Gemini for Home arrived. Sort of that one. That one's sort of still rolling out. But just, just in the last few weeks, I think more people have been getting these, getting Gemini for Home on their devices. So this is. SO plus was, though, arguably the. The largest launch this year. It was first announced in 2023, so it's taken a while. And then they announced it again in February of this year. Finally started to roll out to most people in the summer. And then they launched new Echo products, which also was pretty big news in this space. New Echo shows and a new Echo Dot Max and a new Echo Studio. And the Echo Max is purple, which was definitely my favorite thing about it. As you can see. I like purple. And this, this has been a sort of ongoing news story. I feel like it's the thing I've written the most about this year because there's a complete reinvention of the voice assistant. And I'm sorry for everyone whose devices I have been triggering. I will try, I will try to stop doing that. But yes, it's now instead of this command and control system that we have had for the last decade where you can just say, hey, A, turn on the light or hey, A, what time is it? And get a response. If you try and do anything more advanced, generally you wouldn't get much back. Sometimes, you know, it'd be like, I cannot do that or I'll have to look it up on the web. Whereas the new A and the Gemini for Home is much more conversational. So you can have a conversation back and forth with the assistant. You can ask it to do something with more, more variety or with some pauses or ums and ers, or change your mind mid sentence and it still will catch up and respond and do what you asked most of the time. This is the flip side of what has been a good experience in many ways. It's also been. I found it's caused some quite a lot of issues with smart home devices. It's like the technology, again, we're pushing this exciting new technology into people's homes when it's not always exactly what they need. It's sort of broken some of the functionality of what the old devices used to do. So the old A used to be very limited but fairly reliable. The new A I Find can do so much more, but also struggles with sometimes just basic things. I had a really interesting conversation with both my voice assistants last night asking them about the weather, which is something that they do have done very well for most of the past decade. And neither of them could agree on what the temperature was and neither of them told them me the right temperature. So it was like, what am I doing here? Am I like a better tester for these big companies? I think I am, yeah.
- [30:03]
A
That's the thing.
- [30:04]
B
This is the year of being a.
- [30:06]
A
Beta tester, unfortunately, across the board. Yeah. And it's, it's, it's frustrating to me, particularly as you're saying this idea that we will compromise on what was once arguably quality for the idea that you may at one point in the future have this very smart system. I don't think that the trade off is worth it if I can't do a timer or figure out the weather. Yeah, because you want to make it so that this thing can have a conversation with me. I don't want to have a conversation.
- [30:45]
C
No. And that's, and that's the thing, I think, I mean, to be fair to the companies, they have been rolling this out in early access so you don't have to do it. Although for Gemini, for home, that is eventually going to be unreversible or irreversible from what I understand. Whereas with a, you can say, hey, I want to go back, give me back the old one. But yeah, it's like the potential and the promise is what they're betting on. And what with Alexa, there are a lot of new things that it can do that it couldn't do before. Like I've written about, I love the new cooking experience. So much better than the old one. There's a lot of good coming with it. But if they can't do the basics, yeah, we're losing what we had. And that I'm, I'm finding frustrating. And I'm not the only one. I've seen a lot of users complaining about basic things that they used to rely on their devices for no longer working. Now, hopefully the companies will fix this, but it seems to me that they're all sort of being attracted by the bright shiny star that is the future and not really focusing on the present and what we need in our home.
- [31:49]
A
That's it. It's just this hand wavy sort of thing. Future. Now we do need to take a quick break. We'll have to do like a 10 minute speed round when we come back because I know there's loads more to talk about, but I got to take a break here to tell you about our next sponsor bringing you this episode episode of Tech News Weekly. It's our friends at outSystems, the number one AI powered low code development platform. Organizations all over the world are creating custom apps and AI agents on the Outsystems platform. And there's a good reason for that, because Outsystems is all about outcomes, helping teams quickly deploy apps and AI agents and deliver results. And the success stories are endless. Here are some examples. They Helped a top US bank deploy an app for customers to open new accounts on any device. Delivered 75% faster onboarding times Helped one of the world's largest brewers deploy a solution to automate tasks and clear bottlenecks. I don't know if that is a pun, but it should be. If it's not, which, well, an intentional pun anyway. Brewers and bottlenecks anyway delivered a savings of 1 million development hours and even helped a global insurer accelerate development of a portal and app for their employees, which delivered a 360 degree view of customers and a way for insurance agents to grow policy sales. The Outsystems platform is truly a game changer for development teams with AI powered low code teams, custom future proof applications and AI agents at the speed of buying with fully automated architecture, security integrations, data flows and permissions. With Outsystems, it's so easy to create your own purpose built apps and agents. There's really no need to consider off the shelf software solutions. Again, OutSystems number one AI powered low code development platform. Learn more at outsystems.com TWIT that's outsystems.com TWIT thank you OutSystems, for sponsoring this week's episode of Tech News Weekly. All right, as I mentioned, we are joined this week by Smart home Mama herself, Jennifer Pattison Tuohy of the Verge. And we're doing a year in review of smart home news and tech and we've got eight minutes, eight minutes left to talk about the rest of it. Oof.
- [33:58]
B
Can we do it?
- [34:00]
C
I'll be quick. I'll be quick. Okay, so have to. Obviously matter was a big story again this year. It did feel like this was the year it started to come into its own. And I think one of the things that really identified that was ikea. So Ikea launched a entire new smart home line for its Dirigia hub and that line is all thread and matter compatible, which means you actually don't need to Use the hub hub with it which is something again, you know the smart home and hubs has sort of been a, an ongoing issue for people. And so now with matter that's sort of been. The promise is you don't have to have proprietary hubs in your home. And Ikea, one of the largest furniture globally recognized furniture manufacturers that has now pushed further into the smart home, committed to thread and committed to matter was kind of a shining moment for the smart home protocol which is aiming to make interoperability in the smart home easier. Make everything just work again. One of these going back to that theme of, you know, what do people actually want their smart homes to do versus what technology companies are trying to convince them they need in their smart home. And Ikea kind of delivered some of the really simple, easy, hopefully easy to use because I haven't got to try them yet. I see a couple of your listeners have got their hands on some so I'm excited to get my hands on some here soon. I can't pronounce the names of any of them but they released, they released smart plugs, smart switch dials, they release motion sensors, leak sensors, all the kind of core basic elements light bulbs which are coming later this year and they are, are also inexpensive like four to five dollars for these sensors. Light bulbs are going to be around $8. A few indoor air quality monitors are a little more expensive because there's a bit more tech in there. But inexpensive interoperable smart home devices that work with any smart home platform that supports matter. So Apple Home, Amazon, Google Home, Samsung SmartThings Home Assistant. It's like Christmas came early for the smart home. This was, this was a big, a big shift for the matter standard I think and we started to see a number of these, the big companies in this space sort of fully commit to matter this year. Which actually leads me to my other point that I want. Other news story that I thought was a pretty big one this year which was Hugh Phillips Hue, the Smart Life lighting company. It also fully support supported Matter with its brand new line. Came out with much less expensive smart bulbs. It's always been a big issue with Philips Hue. They're very good price, very expensive and they came out with a line with new bulbs that cost as little as it was $25 for a four pack. So really inexpensive. What's the trade off so that I have. The main difference they said is it doesn't have have as strong light color sorry as strong color matching. So if you have the high quality Hue bulbs and you Have a number of them in your room and you choose a certain color, they will all match perfectly. Whereas the less expensive ones won't have as much range and they, they, they won't have the color matching color sync they're calling it. They will, but you're, I mean to the untrained eye you are probably not going to know the difference, right? And then they are also have slightly lower dimming ranges. So for the more expensive ones. So if you like that really dim subtle light as opposed to a little bit more bright on the lowest dimming level. So not a huge, not, honestly I don't think a huge trade off considering the price. So you know, you, you can put the expensive ones in your very sort of crucial light fixtures, whereas and you can have the inexpensive ones in, you know, the cupboard in the, the hall or your closet or your laundry room. And again, just makes the smart home more affordable, more accessible. And these all work with thread and matter, so you don't even need the Hue bridge. You can just connect them straight to whichever platform you like. But if you do get the Hue bridge, this is one of the new stories of the year I was really excited about. Hue launched a new system called Motion Aware which uses so cool, cool, it's really neat. It uses the radios, the zigbee radios in the bulbs to communicate between each bulb so that you can act, they can act as motion sensors. And you no longer need little white boxes around your house. If you want your lights to turn on magically, you can just use this new Motion Aware tech which is powered by a company called Ivani. And I'm really interested in the advances we're seeing in sensing technology in the smart home. This is something that I think we're going to see a lot of new next year. We're looking forward to 2026 where more ambient sensing in our homes will make smart homes start to work a lot better for the average user as opposed to someone that is willing to spend hours setting up these little devices and, and getting everything working. So this was something I was really interested to see the launch of because it's been something they've been working on for years. There have been a couple other brands. In fact their sister brand Wiz has a, has a technology the same that uses WI fi and wireless sensing with WI fi has been, has been a thing for a while too, but it's never been very reliable. But we're starting to see AI is actually helping to sort of fine tune this kind of in sensing technology. So I think we're going to see a real uptick in reliability and use cases for this type of, of technology going forward. And the final thing, because I know I have like two minutes left, it was big news is cameras finally came to matter.
- [40:08]
A
Finally.
- [40:09]
C
They haven't actually yet, but they will soon. So that was exciting. This was something that everyone had been wanting since the smart home standard launched. It's like video doorbells, security cameras, as such an important part for a lot of people when they start to put a smart home together and not being able to have this connectivity of what can be some of the more important elements of your smart home with matter was, was seen as a big barrier. Now you can connect when the platforms support it and when manufacturers start to produce cameras that support matter, the idea will be that any camera that you buy that supports matter, you can live stream and use with the cloud service or local storage service of the platform of your choice, which would just be, I mean, it will make things so much simpler when you're buying and setting up smart home devices and you know that you want to use Apple Home or you want to use Amazon, Alexa or. But unfortunately none of those platforms have committed supporting matter yet. But, but fingers crossed that will be happening soon. So that's a quick roundup of my year in the smart home maika. What stood out to you from that stuff? Was there anything in there that you were most excited about?
- [41:26]
A
I was just going to say yeah, to kind of round it out with you. Seeing all of the products start to pivot toward AI more has been an interesting aspect, but I think the coolest thing for me absolutely was Philips Hue saying, look, we're going to take what is for the most part existing technology and do something cool with it. Because you did have to get an updated bridge if you had an older one. I got the updated bridge and set it up and it was funny because I was just curious. I just wanted to see how it worked. And my partner came downstairs and he said, why? I just walked into the room and the light turned on in the room and I didn't. Did you turn that on? I said, oh, I was doing a test and it worked. It worked, it worked. It's a really cool technology. Like, it's, it's just really neat. It's really neat. So that more than anything stood out to me because I love the idea of adding value to stuff that we already have. That's. I think that's great. So. Jennifer Patterson Tui, there's so much that's going to be going on in the new year. You're getting ready for ces, so people will need to stay posted and know what's going on. Where do they go to keep up to date with all the great work you're doing?
- [42:35]
C
Yeah, come check. Check me out@theverge.com you can follow my Click on my bell line and follow me and you'll get all my stories. And I also will be doing a lot of social videos on the Instagrams and the tick tock from CES this year. So if you want to keep up on all the hot new gadgets, that's those are places to find me. I'm the I'm at Smart Home Mama on most of the socials and Happy Holidays and great to hang out with you again, Micah and look forward to seeing you in January.
- [43:04]
A
Yeah. Yes. Always such a pleasure to get to chat with you. We're gonna have to do like a bonus episode sometime where we just spend the whole time talking. There's so much to talk about and I love getting to talk about smart home stuff with you. So thank you. Happy holidays to you and yours and we'll see you soon. Bye Bye bye. All righty, folks, we are gonna take a small little break before we come back with Allison Johnson of the Verge, who's joining us to talk about how, how and what is going on in the world of AI image generation. But before that, I have a suggestion for you. Because it's the holiday season and you might be thinking, what do I get the person in my life who has all the things they already want? Maybe they're like me, where they do a lot of research. They say, you know, I don't really need you to get me a stand mixer, because when I get a stand mixer, it's going to be the best one I can buy. And if you buy one and it's not the one that I want, then I'm going to go, oh, this is awkward. Those kinds of folks. Well, I think you should consider a Club Twit membership. Twit TV Club Twit is where you go to sign up or sign up someone else. And when you join the club, you gain access to some awesome benefits. It's all of our shows with no ads, just the content. You gain access to our special feeds. We've got three different feeds. One for bits and clips, one for our coverage of live news events, coverage and commentary of live news events, and one that has the different shows that we publish on the network or in the club, which include my crafting Corner that has my D and D adventure I ran recently and things like coffee time and photo time and so much more. So if all of that sounds good to you, plus access to the members only Discord server, then join Club TWiT, Twitter, TV Club TWiT, and we thank you. All right, we are back from the break, and as I promised, we are joined by Allison Johnson of the Verge to talk about the state of AI image generation. Welcome back to the show, Allison.
- [45:11]
B
Hello. Good to be here.
- [45:12]
A
Good to have you. So I was perusing the interwebs as it were, looking for some interesting pieces, and I came across yours. And I thought, you know, I think I may have noticed a little bit about what you're talking about, and I think you did a fantastic, fantastic job of kind of digging into this and understanding what all is going on in this space. So I was hoping that in talking about AI image generation and the current state of things, could we perhaps roll back the clock a little bit and have you tell us about the early days of AI image generation? What were those images like? I'm laughing thinking about them. How did the technology evolve from tools like Dall? E to where we are now?
- [45:56]
B
Yeah, Dall? E is a name we probably haven't thought of in a little while. Yeah, just maybe three or four years ago, having AI generate an image for you was. Was kind of new in a novelty. And the images you would get from, like, OpenAI's tools were very low resolution. They got, like, steadily better over a few years. But there were always tells, you know, some of the classic things like too many fingers or just weird details that didn't really make any sense, and sort of a general kind of glossiness. I don't really know how to sum it up very well, but kind of a glowy perfection that you don't really get in the real world or from the tools we use to capture the real world. So that kind of brings us to the present day, where that is quickly changing, I think.
- [46:57]
A
Yeah, absolutely. I shared an image, John, for you to bring up, if you can. This was actually from Mid Journey, which also had its own, you know, kind of early days before we got what we have now. And I had asked it to let me see if we can't get this pulled up. I had asked it to create an image. What did I say? I said, gummy worms perform a religious ceremony at Stonehenge. High detail, realistic, and first and foremost, there's already a story in the way that you once prompted these things. You had to do it In a very specific way it was. Was all. And then this is the image that it created. A does not really look like gummy worms, barely looks like Stonehenge, but at least you can sort of pick that up. And the whole thing is just odd. And not. Now you type something like that in and it's getting pretty close to what you imagine. But in this you kind of write about AI generated images settling into kind of a. I love this. A bland middle ground that's kind of technically correct, but then lacks character. Can you describe that looks like. And maybe why. Why it happens?
- [48:18]
B
Yeah. And one place I notice this particularly is there's all kinds of tools, especially in, in Google Photos pixel phones to edit your images with AI. And it can be simple stuff that's like taking things out of a scene or you can go full on AI generated rainbows and things like that. And I messed around with these things and just noticed a lot that when you take the imperfections out, you know, the extra people who are maybe standing in the background or little details that you, you didn't like, things just sort of look bland. You sort of can edit your way into a stock photo basically. And I see that a lot with some of these, particularly like the earlier image generators that, you know, they're. They're fed datasets and they are asked to make predictions based on the. What they know. If you know what is a rainbow, it has ingested photos of rainbows and it will make a prediction and give you something based on what it has seen. But I don't think they can really surprise us or kind of give us something interesting or new or slightly imperfect in the way that like human generated art can, which is a whole doctoral thesis, I guess.
- [49:40]
A
Yeah. For real. Now Google did recently update its image generator VEO3, which I already get confused between the Nano bananas and the VEOS of it all. Can you tell us kind of the approach there? Why is embracing imperfection actually an improvement which has sort of become, I think, the thesis of your observations in your new piece?
- [50:05]
B
Yeah, and Veo or VO3 is the video generation part. It's all kind of just wrapped up in Gemini now. But Nano Banana and nanobanana Pro is the image generation piece. So the thing we started noticing in nanobanana Pro came out maybe a month ago, we were noticing and people commenting on Reddit were kind of like, whoa, these images look super realistic. And you don't need to necessarily prompt it in a way that's like make it look ultra realistic, make it look like it came from people taking a photo of their friends or something like that. And when you were like wholesale generating an image and saying, show me a scene from the Washington State ferry of downtown Seattle, it gives you something that looks like a smartphone photo. It has these kind of unmistakable qualities. They're really kind of over sharpened. The exposures can be sort of skewed bright or a little bit darker depending on the scene. All kinds of weird little things that I notice because I look at smartphone photos all day, but I think kind of signal to us like, oh, this is a real photo.
- [51:32]
A
Because it has that particular look that makes sense. Now Adobe Firefly has a feature, lets users dial down stylization of AI images. And I think for me at least Adobe doing something is when I look at sort of tech journalists and the way that they cover things, then look at sort of the creative space and the way that they cover things. When Apple does something, there's sort of a blessing that gets placed on that concept. And for me, when I see this in the creative space, when Adobe does something, there's a blessing that's placed on it and people go, okay, now it's serious, now it's for real. Tell us about the Firefly feature and kind of like, who is, who is it designed for? Is it designed for the, the professionals who are, you know, attempting to create sort of content? Is it for the everyday user who has an ado? I mean, I, yeah, who's it for? What does it do?
- [52:40]
B
I think you're right that Adobe is sort of the gold, excuse me, sort of the gold standard. Definitely a tool for professionals, students I guess, and people who are creating imagery for websites or promotional materials, that kind of thing. So Firefly is their tool to generate AI art images and you know, you, you ask it to create something, maybe a photo of two people stand posing together on a street and kind of the default output. And I think this is true of other tools you get with chat, GPT or even Meta has that kind of glossy, artificial look to it where you're sort of cued like, oh, this looks a little too perfect and a little too nice to be just a snapshot of two people. But Firefly does have some adjustment sliders and you can tone that all the way down. It's called something like visual style or intensity. Meta has a similar setting and it brings that gloss all the way down. And what's interesting about Adobe's tool in particular is it'll look more like a photographer took a picture. It's a little less smartphone Y but there's kind of like maybe there was some off camera flash and it all kind of makes sense with who Adobe is targeting, maybe the data it was trained on, but they do have their spin on, make this look less like AI art and more like something that just happened in the real world. And of course it's completely fake.
- [54:31]
A
Yeah, yeah. Have you heard the Good news? The 2026 TWIT Audience Survey is live. Yes, it's that time again. We use the survey to help us understand a little bit more about what you like, what you want, what you don't want, what you think of our shows. It's super quick, super easy, and it's open now at TWiT TV survey 26. So please take a moment to fill it out before mid January. That's when it closes. Thank you so much for helping us grow. Now we've kind of talked about this throughout, but there's this kind of parallel that you draw between AI image generation and computational photography because obviously there's a lot going on behind the scenes. Even without throwing AI into the mix in this way, this generation based AI. Can, can you tell us a little bit more about, as you were putting together this piece, this sort of coalescence of these worlds of computational photography and AI image generation and kind of do you think that's going to affect the way that we perceive images going forward?
- [55:41]
B
Yeah. So computational photography, of course, is kind of associated with, with smartphone cameras. They're very limited in, you know, their image, the image sensor size, the lenses are small, so they have to kind of do these computational things to make up that difference versus the quality of a, of a bigger camera. So these are things like merging multiple photos and frames and kind of trying to optimize for a small screen. And that has kind of led to what we think of as the smartphone photo look. So maybe shadows are raised a little bit and details kind of pop. That all comes from the smartphone just sucking up a lot of data every time you're pushing the camera shutter and putting it together in a way that, that tries to recreate the scene not necessarily in the way that we 100% saw it, but the scene that it thinks we're trying to capture. Which feels like a loaded statement.
- [56:50]
A
Yeah, yeah, exactly. That it thinks. Yeah, exactly.
- [56:53]
B
Yeah, yeah. So things like if, if there's somebody's backlit, it'll raise the shadows because it assumes you want to take a picture of that person's face, you know, so it's not. Our phone cameras are already not 100% just like faithful to whatever is in front of it. It's already doing a lot of processing. It's making a lot of judgments about exposure and detail and color and all that. So there is a little bit of a strange area you get into with these AI generated images. And I think one way that it is finding you can create something that looks realistic is to mimic that particular computational look.
- [57:44]
A
Yeah. Now, you did speak with the developer of the Halide camera app, and this I think is sort of going further along this line. Can you just tell us from his perspective the way that things are headed and if you sort of of feel in your own look at all of this, that that is an accurate representation of the way things stand?
- [58:07]
B
Yeah. So this is Ben Sandofsky. He's part of the team that founded and develops the Heled Camera app. It's very popular app for iPhone photographers. And, you know, I contacted him to kind of be like, hey, I'm seeing this with Nano Banana Pro. Do you think this looks like phone photos? And he was like 100%. And he had an interesting observation in that, along those lines of the computational photography look, the AI can mimic that, and that sort of sidesteps the Uncanny Valley thing of it doesn't have to reproduce something that looks like the real world, that looks 100% perfect, because in a way that gives it away. All it has to do is be able to get really good at mimicking the way that we capture those things. And that kind of sent me on a spiral.
- [59:09]
A
Yeah, absolutely. Yeah. Like, okay, I think the last thing that I would love to hear about is because we have the C2PA and you can tell us a little bit about what the C2PA is, the content credential standard, and this being a potential solution for distinguishing real AI, excuse me, real from AI generated images. What is it? Who's adopting it, what needs to happen for it to actually work? And then the last thing I would ask for you is, is do you think it will actually work if even enough people or corporations adopt it?
- [59:55]
B
Yeah. So C2PA is a coalition of a bunch of different companies like Google and Adobe are all involved, and they've been tasked with, you know, finding a way to label photos and the provenance of images so that people can tell what's AI generated and what's not. And it's a very hard thing to do for many reasons, but they've been chipping away at this. So it's. It's kind of the standard we have, and it's very Much a work in progress. So content credentials, they're cryptographically assigned to a photo, and so they, they cannot be changed. It'll say, you know, this photo was created with a camera, or it was created with a camera and edited with AI. That metadata can get taken off if you change, you know, the format of a file or mess with it in certain ways. But the crucial thing is that it can't be. It can't be changed and tampered with without being marked as such. So it's, it's sort of. It's not perfect. It's the standard we have. Google, I think, has taken some real strides in the past year adopting it. They apply it in their pixel cameras. Every single picture you take with a Google Pixel 10 phone will get a label as taken with a camera or created with AI or edited with AI tools, and that will follow that photo wherever it goes. So the other piece of that is, how do you show those credentials? And that's kind of the, the messy next step that we're in. You know, Instagram has some, you know, you can label something as AI. There was a time when they were putting automatic AI labels on, and then it was tagging things that were just, you know, edited with, like, Photoshop tools that were not just generated with AI. So it's a little bit of a confusing. There's some confusing cases there that need to be solved for. And I think that's kind of where we're at is this uncomfortable middle ground between we have some tools, they're not widely adapted, adopted, and we need the platforms to get on board with good ways of showcasing them. And I think we're just kind of in a spot that's ripe for, like, confusion and misinformation. I think we're going to get to a better place, but it's going to be weird for a while.
- [62:47]
A
I do appreciate the optimism there. It makes me feel a little bit better about the situation, and I hope that's the case. But I hope we all, in the meantime, keep that healthy dose of skepticism every time we see a photo and just go, this may or may not be real. Or now video too. This video may or may not be real. It's all very important. Allison Johnson, always a pleasure to get to talk to you here on the show. If people would like to stay up to date with the work you're doing, where should they go to do so?
- [63:15]
B
You can go to theverge.com, you can follow me on threads or Instagram.
- [63:19]
A
Allisonjoe1 wonderful thank you so much for your time. I appreciate it.
- [63:24]
B
Thank you.
- [63:25]
A
Alrighty folks, that brings us to the end of this, the final full episode of Tech News Weekly of 2025. As I always say, thank you so much for being here. Appreciate all of the support this year, all of the wonderful chats in the club Twit Discord and so much more. And of course a huge thank you to all of the guests who joined me this year, all of my wonderful co hosts this year. Shout out to Jake Ward who joined us as well. If you would like to tell other people about the show, Word of mouth is the thing that helps us the most when it comes to getting the podcast out there. Have them head to TWiT TV TNW to subscribe to the show and audio and video formats. I already talked a lot about Club Twit during the show Twit TV Club Twit is where you go to sign up, so be sure to check that out if you want to follow me online. I'm ichasargent on many a social media network where you can head to Chihuahua Coffee, that's C H I H U A H u a Coffee where I've got links to the places I'm most active online. Be sure to check out my other shows including iOS today. Hands on Apple, Hands on Tech. Thank you John Ashley for everything that you do to make the show possible. I quite literally could not do the show without you and it means a lot. Thank you to all of the other folks Burke behind the scenes who helped make the show possible as well. And I do believe that is it. That is all. Goodbye to all of you for 2025 and boy howdy, we ready to hit the ground running in 20. 20.