Tech News Weekly 455 Transcript

Please be advised that this transcript is AI-generated and may not be word-for-word. Time codes refer to the approximate times in the ad-free version of the show.

Mikah Sargent [00:00:00]:
Coming up on Tech News Weekly, Jennifer Pattison Tuohy of The Verge joins us this week and we talk about Google adding MCP support. For you AI nerds, you know exactly what that means. Plus, the request for Apple to add a smart home computer. Then we dig into Apple's reference image and Victoria Song of The Verge joins us to talk about how health age is a fake metric and a little conversation about being aware of AI journalism scams. Stay tuned for this episode of Tech News Weekly.

Mikah Sargent [00:00:46]:
This is Tech News Weekly, episode 455, with Jennifer Pattison Touhy and me, Mikah Sargent, recorded Thursday, September 17th, 2026: The Trouble with "Health Age". Hello and welcome to Tech News Weekly. This is the show where every week we talk to and about the people making and breaking the tech news. I am your host, Mikah Sargent, and later in the show, we will be joined by Jennifer Pattison Tuohy of The Verge. I needed to rearrange things today because we've got a great guest coming up at 15 past the hour. But before we get to that, I wanted to share a story to kick things off today. About a topic that I think I hear about more than any other when it comes to those among us who aren't the biggest fans of generative AI.

Mikah Sargent [00:01:43]:
And that topic is, of course, AI slop. Because if you've searched for anything in the last year, perhaps space trivia, some pop psychology, or even a vegan recipe, then there's a real good chance that you have landed on a website site that is owned by 3 Australian brothers that you probably never heard of. Well, at least until now. Maggie Harrison Dupree, John Christian at Futurism spent months, months investigating Brown Brothers Media. This is a company that's run by Justin, Lachlan, and Brendan Brown and has quietly assembled a network of at least, as far as we can tell, or as far as Future, uh, Futurism can tell, 24 websites drawing an estimated 64.7 million visitors a month. Of course, if you count that, that is more digital readership than quite a few popular online sites, including NPR. It's more than CBS News, more than Forbes. And how many employees do you think they have in comparison to CBS News and Forbes? Turns out It's about 12 employees.

Mikah Sargent [00:03:04]:
And that's because, as you can imagine, none of the writing is done by people. And frankly, some of it isn't even attributed to people who exist. We've got fake NASA scientists, a fake retiree whose age changes month to month, real writers who lost their jobs, and then a search engine that doesn't seem especially bothered. So let's dig into what we're working with here because boy howdy, it is, um, a, a tale as old as the time of the existence of AI, generative AI. Uh, so Futurism, in looking through this, found more than 50 fabricated writer personas, and that's across the entire Brown Brothers Media network, and again, invented in many cases. Space Daily, this is an estimated 10 million page views per month. There are articles written by a Dr. Katherine Chen, who was described as a former NASA Jet Propulsion Laboratory systems engineer who, quote, spent 15 years designing autonomous systems for deep space missions.

Mikah Sargent [00:04:20]:
Alongside her is, of course, Dr. James Whitfield. who was identified as an aerospace medicine researcher at the European Space Agency. And I think you can imagine where this is going. Neither of those people exist. NASA did decline to comment, but did say it, quote, does not verify past employment or affiliations with the agency. The ESA was a little blunter, saying, Mr. James Whitfield is not currently employed by the European Space Agency, nor do our records indicate that he has ever worked here in the past.

Mikah Sargent [00:04:55]:
There's also a tech site. Oh, hey, it's called Direct Message News. And Rachel Vaughan was on there giving mental health advice as a Dublin psychologist with a master's in applied positive psychology from Trinity College Dublin. The British Psychological Society, as you can imagine, had no record of her. Trinity College Dublin told Futurism that, quote, applied— I love this because immediately when I read applied positive psychology, I thought, What? Who?

Mikah Sargent [00:05:28]:
Where? And yeah, Trinity College Dublin said applied positive psychology isn't a degree or even a course that the school offers. And then Futurism took her headshot and traced it to a stock photo in Pixabay. And here's the messy thing about it. They didn't even go as far as to pay for it. It was a free stock photo in Pixabay. I laughed not because I'm happy that this grift is happening, but simply because it's ridiculous, the sort of, I think, earned hubris that exists in being a grifter. I mean, they're called con men because they're confidence men, right? They've got confidence. They instill confidence in you.

Mikah Sargent [00:06:22]:
So why would they think that they need to do more than get a free stock image? It's too easy. to make these online and, and to sort of put this together. So here's an interesting thing about sort of tracking across these various sites. Sometimes the fake writers that were involved seemed to have struggled with a bit of, of forgetting who they were, because Marlene Martin, who is a retired English teacher— for those of you listening, I'm doing big scare quotes— on VegOut was used to— or excuse me. Yeah, yeah, that name, Marlene Martin, was used to publish thousands of first-person essays about loneliness and about purpose. Her retirement age in one was 66, in another 64, and in another 61. So she's retired at various ages. Perhaps the Social Security Administration needs to look into that.

Mikah Sargent [00:07:26]:
Her own age moved around too. She was 68 in January, 69 in February, 71 in March, and then went back to 70 in April. And this was in an essay about her very real husband's death. Sarah Mitchell, lead editor and a clinical psychologist at Silicon Canals, wrote a sentimental piece— that's probably canals— wrote a sentimental piece about her baby's first word being Mac. which is the family dog's name. This is so depressing, all of this being made up. Weeks later, in a piece about the psychology of dog owners, she wrote, I don't have a dog. I should probably admit that up front.

Mikah Sargent [00:08:09]:
There are, of course, more examples of the, the different writers forgetting their own history. And You know, here's the thing. It's not just a matter of completely made-up stories. In some cases, like in February, VegOut had an investigation that was credited to a writer called Jordan Cooper, and it claimed that the author bought a 4-pack of Apple AirTags, superglued them inside plastic containers, and sent them on their journey through America's recycling system. And then There was an accompanying photo of a bottle with an AirTag inside that showed sort of an Apple logo that was all garbled. And of course, it seems like it was AI-generated. The investigation, it was a near copy of a 2024 story by CBS 8, which is a San Diego station, and tracked a water bottle, a yogurt container, an orange juice bottle, and a wipes container doing the same method. VegOut listed water bottle, orange juice jug, yogurt container, and a cleaning product bottle.

Mikah Sargent [00:09:21]:
And then of course, never mention CBS 8. There are AI detection systems. We've talked about those in the past and how I say to use those with caution because they often aren't going to get things right. But Futurism did add that they did run an AI detection on the text, and the system gave it a 100% AI-generated result with no sign of human involvement. So what is going on here? Why take all the time to set up this infrastructure and make this happen? Hmm. Well, let's talk about the business model. Brown Brothers Media buys publications that have real editorial identities and old Google Trusted Domains. That's the important part.

Mikah Sargent [00:10:20]:
And then they take those and they flood them with AI content. The portfolio grew from just like less than, fewer than a dozen sites in 2023. In 2024, went to 17. In 2025, they had 24. And they seem to be adding even more this year. BBM, as it were, acquired the Law Dictionary— oof— Science Blog, and the Space Media Network. Those brought in Space Daily and more than 20 sister sites, including Mars Daily, Solar Daily, and Nuclear Power Daily. BBM republished thousands of old Space Media Network articles as current news, which gave headlines like iPhone 5 rollout draws big crowds worldwide.

Mikah Sargent [00:11:15]:
That was posted June, and of course was about a product from 2012. A 2025 press release credited the traffic to organic reach through Google Discover and Search. And of course, the brothers are linked to AI content scaling startups Mosquito Fleet, Tippa Fleet, and Cognitive AI Systems Limited. These companies that do the AI content scaling are registered in the British Virgin Islands, the UAE, They are— it's really troubling, to be frank. One of the sites that was sold, the VegOut site, Maggie Ortlib was the original founder of the site and sold it. She said she started it at 24 and said on Instagram, unfortunately, the new owners turned it into just another AI-written content website. So not really the direction I had in mind. I can imagine that's the case.

Mikah Sargent [00:12:21]:
There's so much more to this, like so much more to this. This Futurism piece is well worth reading to support. But what I want to get at is that these companies are— this, this company in particular, BBM, is scooping up And using reputation of former human writers and creators, and then making money, of course, on the ads. Now, here's the thing. Google spam policies do prohibit what they call scaled content abuse. So using generative AI tools to generate a bunch of pages but not adding value, and then creating multiple sites with the intent of hiding the fact that they are doing this AI content scaling. And Futurism did send questions in. Google, at that point, did choose to deindex several BBM sites that included VegOut and Silicon Canals.

Mikah Sargent [00:13:25]:
Weeks later, unfortunately, back in the search results. So it's messy. It's especially messy when a company that is running Like the, the world's search engine is also using AI and generative AI and shipping generative AI products and trying to find a balance between the two and figure out these policies together. And so that's really what this conversation, I think, needs to be about is how we handle this future where generative AI results in sites that have a lot of content that may or may not have accuracy, that may or may not have real people behind it. And what, what should be the responsibility of the individual to check the sources of the site? What should be the responsibility of the individual user to check if their AI systems that they're using are sourcing from these other sites that are created by AI? It can get real messy real fast. And I think that that's something that we're still trying to reconcile for sure. But I know that we're all working on it, right? And Google is actively updating its policies as we continue to see this roll out, as we continue to see more generative AI content hitting the web. So again, go check out that Futurism article I linked to.

Mikah Sargent [00:14:55]:
We linked to it in the show notes, and so you can check it out there. But It is time for us to take a quick break before we come back with my interview. All right, we are back from the break, and I am excited to be joined today by the wonderful Victoria Song of The Verge. Welcome, Victoria.

Victoria Song [00:15:14]:
Hi, thanks for having me here.

Mikah Sargent [00:15:16]:
Absolutely. So I was scrolling through my various sites and I saw my brain plastered on the screen and I said, You, you get it, uh, because you put a piece out on The Verge called Your Health Age Is Fake. And I am so excited to talk about this because as much as I am a sort of quantified self person who loves to track all the things and is, you know, wearing an Oura Ring right now and has my Apple Watch on the other wrist and, uh, has my bed tracking me, uh, there are some metrics that just tick me off is what I'll say, because this is a rated PG show. And so I would love to hear Victoria Song. Can you start by telling us, um, what is— what is health age, and why is, um, your Whoop age your villain origin story?

Victoria Song [00:16:16]:
Yeah, so health age, or biological age— there's a lot of different names for it, but a bunch wearable and health tech companies are essentially giving you a score for how well you're aging these days. It kind of coincides with a trend towards longevity being a big wellness trend right now. And I think it's meant to be kind of a helpful visualization of how lifestyle factors that you have control over— so things like sleep, regular exercise, or the stuff that you're eating could contribute to your body aging better than average compared to your chronological age or worse than average compared to your chronological age. So that's just like the overall concept that we're talking about. And Whoop put out its version of it, I think, with their last update with the Whoop 5.0 and the Whoop 5.0 MG. And it told me I was 41.8 years old, so like 42? I'm not that. It was like about, at the time, 5, 6 years older than I actually am. And I was just like, oh, rude.

Jennifer Pattison Tuohy [00:17:28]:
That's so rude.

Victoria Song [00:17:29]:
And yeah, I was not particularly happy about it as someone who like thinks a lot about my health, tries to live a generally healthy life. It was a— I was a little, a little peeved. I'm not gonna lie.

Mikah Sargent [00:17:44]:
Yeah. I mean, and for me, there's also an aspect of anxiety that comes with it where I go, what if it is right? What if it knows something I don't? What if the way that I'm feeling and the way that I'm actually doing are 2 different things? And that gets a little scary. Now, something that I thought stood out, you were in the Steve Jobs Theater when Apple announced Health Age, and you said you had a little bit of a no moment. When Apple said, let's do Health Age. So I'd love to hear kind of what went through your head in that moment. And more importantly, does— well, not more importantly, but as importantly, does Apple doing this mean that more of the industry is going to follow suit and try to calculate your health age?

Victoria Song [00:18:39]:
Yeah. So when I was in there, I was just like, no, not another age that I have to contend with. That's basically what was going through my head at the time. You can kind of see it in The Verge Live blog if you go look back, that Allison, my coworker, had just basically been like, there's an audible no coming from Vee in the theater. And it was mostly— the interesting thing is that a lot of the health tech and wearable companies have already kind of implemented a version of this. Like, Oura has one, Whoop has one. Apple is just catching up here. One thing I will say, though, is that Apple has an immense amount of scale.

Victoria Song [00:19:20]:
The Apple Watch is the OG health tracker in the sense that they're the first ones to put EKG into their device. I know Fitbit came ages and ages before. I'm not saying that there weren't others that came before it, but it was one of the first smartwatches to be like, If you wear this, it'll have an impact on saving your life, right? Or that kind of ethos in there. And they have these commercials every year at the keynote where they show that the Apple Watch can save your life. It can influence a great amount of health change. So for them to put that in there, I do think it's significant. For the Health app to be getting a major redesign this fall, that's also quite significant. So part of me was just sitting there going like, oh God, here we go.

Victoria Song [00:20:09]:
here we go. This is going to be kind of shown to a lot of people who are not necessarily big self-quantifiers. It's going to be shown to a lot of kind of laypeople, so to speak. So that's kind of the swirl of thoughts I was having at the moment.

Mikah Sargent [00:20:29]:
Absolutely. Yeah, I think that there's something to be— I remember reading a piece after Apple released previous features that showed an uptick, right, in people visiting the doctor for their concerns. And it's like, it's a tough thing for me because when that first— when I first read that, I thought, oh brother, like how, how silly that people are getting one notification and then they're thinking, ah, I got to go in. But then I thought, you know what? I think that our health system should be able to support people choosing to go in to see if their health is being impacted. And so it becomes kind of a balancing act of, in my own brain, of, of where I stand in terms of democratization, so to speak, of—

Victoria Song [00:21:22]:
Mm-hmm.

Mikah Sargent [00:21:23]:
It's almost like the John Deere movement to not John Deere, but opposite of John Deere movement to be able to fix things yourself. And I think that we should have access to our own firmware. But I think it's a matter of packaging how it is. And I think I feel, and it seems like you do too, that health age is not a great package for it. You said that Whoops— not Whoops, but Whoop says you're 42. Oura says that your heart is actually 2 years younger. I've also gotten really good feedback from Oura on my health And then compared it with some others and thought, who is telling me the truth? Because Oura says I'm doing all right and all the others are like, bro. Withings says that your vascular age is a year older.

Mikah Sargent [00:22:07]:
Garmin says 36.5. There was a mirror at CES who said that your facial skin is 26.

Victoria Song [00:22:14]:
Hey!

Mikah Sargent [00:22:14]:
Hey! And then your bed said that you were 38.7 and trending older. Would you like to share your age? I think you did at the beginning now. And then what is it like waking up to that many competing verdicts about Your body?

Victoria Song [00:22:29]:
Um, I am 38. I think I'm doing pretty great for 38, actually. Um, and I was basically like, when I look at these scores in the morning, I'm just like, are we, are we serious here right now? Because it's not, you know, all of these companies have their own formula as to what they weight and how that impacts your overall score and age. And some of them will weight certain factors more than other factors, but it's also just kind of depending on when you test it and what stage of life you're in. Like, I've been a runner now for over a decade, right? A regular running that has made me a fitter person now at 38 than I was as a teenager at 18. Like, my cardiovascular health now is like objectively better. I couldn't run a mile when I was 18. And in my 30s, I've I've run multiple half marathons and I regularly go running.

Victoria Song [00:23:25]:
It's just a fact of my life. So, you know, do— does the health age really kind of measure my health age now, or is it measuring the whole progress of my life? It's not doing that, right? It's not telling me that I'm actually healthier or fitter now than I was when I was 18 in terms of like cardiovascular fitness. Was I lifting weights at 18? Absolutely not. I do now. not as regularly as I probably should because I don't love strength training, but I do engage in it because I know it'll help me be a better runner, right? I don't eat as many French fries as I did when I was 18 and had a great metabolism.

Mikah Sargent [00:24:02]:
Mm-hmm.

Victoria Song [00:24:02]:
My metabolism at, you know, midlife is not what it was. So there's like different things based on the stage of life that you're in. And, you know, one thing that I've heard from readers is, You know, you go through big life events, you go through periods of great stress, great grief, great joy. Those will all impact your health in different ways. There's different seasons of life. A person who is a new parent is going to have different abilities to eat, sleep, take care of themselves than a person who doesn't have any kids or whose kids have flown the nest. So, you know, like these kinds of scores, they don't Tell you the whole picture of what your health is, because health is a holistic thing. So it can be a snapshot, it can tell you areas of your life where you maybe need to put a little more effort, but it's not like a whole picture that you should be like, haha, I'm younger than my chronological age, whatever.

Victoria Song [00:25:05]:
Yeah, that's not— That.

Mikah Sargent [00:25:07]:
That is honestly, I think, part of it. We'll kind of combine a couple of questions here because I think that the 2 words shame and ego play big roles when it comes to, to health age. I think there was even talk of bragging about one's health age. And then, as has been my experience, I know that VO2 max is important. And when my Apple Watch says That's the bro moment. Bro, you need to work on your VO2 max. I feel an immense amount of shame and sort of don't feel motivated to improve it because I just, you know, feel I have failed and that life is slipping away. And I'm 34, I think, right now.

Mikah Sargent [00:25:54]:
Yes, that— oh no, I will be 34 at the end of this year. And so, like, life is not slipping away. I've still got quite a bit of it left. Knock on wood. And so it's hard to kind of deal with that. But maybe the ego side of things, when we think about the people who are doing the quantified self a lot of the times, and we talk about this sort of longevity playing a role, that seems to be something that is, is a big part of this. You talked about— and now I'm missing it, but there was a part where you talked about the guy who's— Oh, yeah, there it is. Brian Johnson's post about aliens, not being able to tell him apart from his teenage son.

Mikah Sargent [00:26:39]:
What does that kind of culture do to the rest of us normal humans who are just looking at our smartwatch for some info?

Victoria Song [00:26:47]:
Basically, Brian Johnson, he is kind of— he's famous online for being a guy who is, you know, trying to defeat death. And he's a person who is trying to optimize himself all the time. And it was really funny because while I was writing this, I found a post of his where he was talking about like an alien wouldn't be able to tell that I'm 49 because I have like the biometrics of an 18-year-old. Like an alien would have a hard time telling me apart from my teenage son. And I was looking at him and I was like, no, I don't think they would. And not in a mean way. You know, like, I thought of it more as like, what are we trying to say here? Like, he looks pretty decent for his age, but you're not 18. Like, your life experiences are completely different.

Victoria Song [00:27:42]:
So I just— it was just very striking to me to be like, why are you discounting the experience that you have and the things that have gotten you there?

Jennifer Pattison Tuohy [00:27:52]:
Mm-hmm.

Victoria Song [00:27:52]:
It's not better to be 18 than 49 per se. Like, I don't think health age is necessarily an evil thing, you know, but I am sick of how if it's younger, it's better, and if it's older, you should feel what way about it. Like, I shouldn't feel what way about the fact that, you know, I downloaded the dev beta of iOS 27 and it said that my health age is like 0.3 years higher than my actual age. And that's because I haven't been taking as much— I haven't been taking as many daily steps because I'm going to all these tech events in September. And I'm not sleeping quite as regularly because I'm going to all these tech events in September. So I'm doing pretty great in the sense that it's pretty aligned with my age, I guess. It's one of those things where I think there's a genuinely helpful intent behind these features, but the way that it's communicated and the way that our culture feels about age is not always going to be a super helpful metric there. So that's just kind of how I feel about the shame and the ego side of it.

Victoria Song [00:29:08]:
I do know that, like, you are going to get some insufferable people going like, hey, my age is younger than my, uh, health than my biological age. I'm in great health. And then some of them are going to go to the doctor and find out actually they have high cholesterol.

Jennifer Pattison Tuohy [00:29:25]:
Ooh.

Victoria Song [00:29:25]:
Or some of them are going to— that's something that a reader shared with me actually. Or, you know, some of them are going to go to the doctor and they're in great health, like Bryan Johnson actually, and find out that they have autoimmune gastritis, which is a thing that he has publicly shared that he has. There's factors that you can control in your health, and doctors will tell you that it's sleep, regularly, exercise regularly, eat healthy most of the time. And that's very good advice to follow. Those are things in your control. But on the flip side, there are things that are not in your control. Like genetically speaking, we are all predisposed to certain things and not disposed to others. We all have heard of somebody who was like super healthy, did everything right, was in the prime of their life, and then died in like a freak accident.

Jennifer Pattison Tuohy [00:30:11]:
Yeah.

Victoria Song [00:30:12]:
There's kind of like a grain of salt you have to take with all of these metrics and not get super obsessed with them. Because again, they're proprietary metrics. They're weighting certain things heavier than the other. I could probably get a lower Whoop age if I just lifted a lot.

Mikah Sargent [00:30:30]:
True.

Victoria Song [00:30:32]:
Sure. And that would be good for me. That's not a bad thing for me, right? It's just sort of like, it's not the whole picture because my metabolism is a part of the picture. And that's an invasive metric where I have to get a blood test to kind of measure that out. It's not taking into consideration the diet that I'm eating. What if I eat a bunch of fried chicken? What if I'm a smoker? But like, it's not everything. It's—

Jennifer Pattison Tuohy [00:30:55]:
Absolutely.

Victoria Song [00:30:56]:
Yeah.

Mikah Sargent [00:30:57]:
It's only one part of the picture.

Victoria Song [00:30:58]:
It's one part of the picture, right.

Mikah Sargent [00:31:01]:
The last thing I'll ask, and it's a quick one, because you mention it, you know, your wearables convinced that your health was in a death spiral, or had you convinced that your health was in a death spiral. Your doctor was going, you're doing great. When people have that happen to them, what's your advice on, on what to do in those situations? Who or what do you listen to, or do you try to get them both on the same page?

Victoria Song [00:31:27]:
Um, I think you more or less check in with a doctor, have a conversation about it, and then check in with how you feel. All of these things are just tools in a toolbox. You are the ultimate person who knows your body best. And like, the whole point of a wearable is for you to be more in touch with your own body and your own health. There are some doctors out there who are really great and some who are— I mean, I've had to fire a couple of doctors in my life because they were not really listening to me or helping me on my health journey. I felt not listened to. Like, I think a lot of people have that experience. So it's just whatever makes you feel most confident that you are kind of feeling your best at a certain point in time, but you do have to be like, your wearable is a tool, it's not a doctor.

Victoria Song [00:32:19]:
Your smartwatch is not a doctor. There are times where you're gonna have to go to a doctor and listen to established science and medicine, and it's a conversation that needs to be had. I think anything that encourages discourse is good.

Jennifer Pattison Tuohy [00:32:34]:
Mm-hmm.

Victoria Song [00:32:36]:
There's no clear— this is like a cop-out, but there's no clear answer. It's just you kind of have to develop a sense and a knowing of like, okay, it's great that this says my health age is younger. I still need to go to the doctor and get my physicals and see if there's anything there. If there's something in my body that feels wrong, I do have to go see a doctor and talk about it. I just don't think you can cut out either per se. You just have to take everything with a grain of salt and check in with yourself. So yeah.

Mikah Sargent [00:33:10]:
Well, Victoria, I wanna thank you so much for taking the time to join us today. If people would like to follow you online and keep up with the wonderful, wonderful work you're doing, where are the places they should go to do so?

Victoria Song [00:33:19]:
They can check me out on theverge.com. I have a newsletter there called Optimizer where I talk about all this stuff every week, uh, and you can find me on on all socials at, uh, @vicmsong. So I just keep my handle the same everywhere for ease of convenience. So yeah.

Mikah Sargent [00:33:37]:
Wonderful. Thank you so much for your time. We appreciate it.

Victoria Song [00:33:39]:
Thank you. Thank you for having me.

Mikah Sargent [00:33:42]:
Alrighty, folks, we're gonna take a quick break before we come back with stories of the week and the wonderful JPT, Jennifer Pattison Tui, in just a moment. All right, we are back from the break, and as I promised, joined this week by the wonderful JPT, Jennifer Pattison Tooley. Welcome.

Jennifer Pattison Tuohy [00:34:00]:
Hello, hello, hello. Happy to be here.

Mikah Sargent [00:34:03]:
Happy to have you. So it's time for our stories of the week, and you got a big one from Google, uh, and maybe a little chat about Apple too. What's going on in the world of Google Home?

Jennifer Pattison Tuohy [00:34:15]:
Yeah, so, um, whilst everyone's talking about how, uh, AI will kill us all, I have a way that AI might kill you in your sleep.

Mikah Sargent [00:34:23]:
Yay!

Jennifer Pattison Tuohy [00:34:25]:
Basically, what has happened is that Google Home, which is Google's smart home platform, has added MCP integration. So this is Model Context Protocol, which basically means that you can now use any third-party AI agent with your Google Home. So you can take your Google Home instance, you can plug it into Claude, OpenCLAW, It also works with Google Antigravity, Hermes, or basically any AI agent, agentic platform or framework that supports MCP. So it's not locked in. You can use it with anything. Obviously, Google mentions Google Antigravity as one because that's their agentic framework. Claude, obviously something that consumers and users might be interested in using. OpenCLAW was sort of at the forefront of the sort of smart home integrations when it launched.

Mikah Sargent [00:35:22]:
Mm-hmm.

Jennifer Pattison Tuohy [00:35:22]:
more recently. So this is a big move for Google Home. There are a couple caveats, but basically what it means is that you'll be able to use the intelligence of your AI agent to digest, analyze, and control your smart home. So this is— yes, I could now use Claude to turn my lights on, but much more sort of importantly, I can actually use Claude to understand the context of my smart home. So this is things like actually being able to understand— well, a couple of the examples that Google have given that you might be able to do is cross-camera analysis. So if you have a number of cameras connected to a Google Home, using an AI agent, intelligence can sort of look through the cameras. And you can say to it, hey, what did my kid do when they got home? And it could look through like 40 different cameras and say, okay, your kid arrived at this time, then they went and had a snack, and then they went to their room. Um, you know, if you follow your child that closely.

Jennifer Pattison Tuohy [00:36:26]:
Um, then also like, how many loads of laundry did I do this week? It can access your smart washing machine and, and look at the data. And how many times did— how long did I leave the lights on? You know, what could I do to reduce energy use? Like, it can actually take the data, analyze it, and give you feedback could help do things like optimize automations, um, help you with more complex tasks that like your smart home interface, especially Google Home, which whilst it's got better recently has been quite limited when it comes to setting up automations. Um, so help you set up more complex automations. Specifically, the thing I'm most interested in would be troubleshooting, helping because it can see the whole picture and potentially see where things are going wrong, design dashboard for you so you don't have to use Google's web interface or Google's app. You could create your own, but still running all of Google's, um, running on, on Google's platform. So it's a really interesting move, um, from Google, who obviously is, is one of the more open platforms. The biggest caveat is it requires Google's Premium Advanced— Google Premium— Google Home Premium Advanced plan.

Mikah Sargent [00:37:40]:
Oh my word, what a name.

Jennifer Pattison Tuohy [00:37:41]:
Which is an expensive one, $20 a month, $200 a year. This does get you video recordings stored in the cloud, Google's AI features. But yeah, you can't just sadly use this to plug into Google Home and suddenly not have to pay Google anything. There are other platforms that use this, like Home Assistant has MCP, I actually used Claude to vibe code my Home Assistant a year or so ago and wrote a piece about that. Me too!

Mikah Sargent [00:38:08]:
Me too!

Mikah Sargent [00:38:09]:
Isn't it great?

Jennifer Pattison Tuohy [00:38:10]:
It's so great.

Mikah Sargent [00:38:11]:
Literally yesterday I said, I want a dashboard for my office and here are some things that I wanted to be able to do. And so now all I do is click one button and it turns all of my lights orange or blue or purple. I've got a thing that just turns off my studio lights. I can see the temperature and humidity in the space. I love the MCP access to Home Assistant. And so when I saw that this was coming to Google Home, I thought, that's great too for people. Because I think you're right. People at first think that it means I can talk to Claude and have it do these, like turn on and off the lights on my behalf or something.

Mikah Sargent [00:38:55]:
And it's like, no, it's about much more powerful instances because otherwise it's a little wasteful of Yeah, you want the analysis.

Jennifer Pattison Tuohy [00:39:03]:
You want to use the intelligence to actually help create a more intelligent home, like, so that your home can actually start doing things somewhat more proactively, hence the potentially kill you in your sleep joke. But yes, the idea here being that there is so much potential once you have the data and the analysis and the ability to analyze the data from your home, and you can connect other things, obviously, to your AI agent. So you can pull in all aspects of your digital life and have a much more complete picture and set up much more interesting automations that are not just about turning lights on and closing shades, but combining everything. And this is something, you know, this is what the smart home has been working towards. This is sort of what Amazon's Alexa Plus was promising when it first launched, although it hasn't really gotten there at all. And Google Home has Gemini for Home now, which again, We kind of thought it would do this, and perhaps this is a precursor to Gemini for Home getting smarter because Google does a lot of this, you know, opens these platforms for developers. I think this really is more of a developer play.

Mikah Sargent [00:40:09]:
Mm-hmm.

Jennifer Pattison Tuohy [00:40:10]:
I think there are people that will, there are consumers that a lot of your listeners that will probably want to play with this, but I think this is really more about developers building, being able to build more powerful instances, apps, automations on top of the basis of using Google Home as an infrastructure. Because Home Assistant, as great as it is, big fan, isn't that consumer-friendly yet. It's getting a lot better.

Mikah Sargent [00:40:36]:
Right.

Jennifer Pattison Tuohy [00:40:36]:
But Google Home is much more so. And, you know, it's a lot easier for people to use and get set up and started with. And now, you know, if you're a developer that has a product that integrates with Google Home, you can use this sort of integration you can use the MCP, you can develop something more powerful, more interesting, and potentially more profitable because I'm sure there's money at the end of this tunnel for most people. That's what, you know, that's what was happening with AI. AI is getting much more expensive in the smart home. Case in point, Apple Home. We've just found out how expensive Apple Home's Apple Intelligence plans are.

Mikah Sargent [00:41:16]:
Yeah.

Jennifer Pattison Tuohy [00:41:16]:
And we're talking $60 a month.

Mikah Sargent [00:41:18]:
$60 a month, which is— that's— who?

Jennifer Pattison Tuohy [00:41:24]:
Why?

Mikah Sargent [00:41:24]:
Why?

Mikah Sargent [00:41:25]:
Why do you think that? Like, what's going on there? What?

Jennifer Pattison Tuohy [00:41:27]:
I'm hoping it's a mistake and that, like, next week they're going to say, oh, hang on, we forgot to tell you.

Mikah Sargent [00:41:34]:
Because it seems so out of touch.

Mikah Sargent [00:41:35]:
I know.

Jennifer Pattison Tuohy [00:41:36]:
So out of touch.

Mikah Sargent [00:41:37]:
You pay more or less— excuse me, you pay less than that for so much more from other places. And that's what I'm struggling with.

Jennifer Pattison Tuohy [00:41:47]:
It's the bundling. So they bundle— Apple Home bundles everything into iCloud. Cloud Plus. And what's happened now with Apple Intelligence and Apple Home, which just officially launched with iOS 27, is whilst you can still use HomeKit Secure Video for much cheaper, I think you can get unlimited cameras on the 6 terabyte plan, 5 cameras on the 2 terabyte plan. And that's, I forget exactly how much that is, but I think it's like, it's not $60 a month. The $60 a month plan is 12 terabytes of storage. But now with Apple Intelligence, if you want Apple Intelligence that's activated on your cameras, you need to pay for either the, uh, 2— 6 terabyte plan, um, or the Apple One bundle plan gets you 2 cameras. No, I can't remember off the top of my head.

Jennifer Pattison Tuohy [00:42:37]:
One camera for $20 a month, 2 cameras for like $30 a month, and then 5 cameras for $60 a month along with increased storage. But yeah, 5 cameras for $60 a month is an awful lot. So which brings me to my other story from the week, which was I think Apple needs to make a smart home computer because we're seeing the benefits of AI in the smart home. There are a lot of great things you can do with it, but we do not need frontier models to do this type of data analysis.

Mikah Sargent [00:43:09]:
Right.

Jennifer Pattison Tuohy [00:43:10]:
To turn our lights on and off, to do some orchestration in our home. to tell us it's a dog at the front door, not a squirrel. We can do this type of stuff on device with more limited, more simple, potentially less capable models. So what I saw at IFA last month— yeah, a few weeks ago. Time, who knows?

Mikah Sargent [00:43:29]:
It flies by.

Jennifer Pattison Tuohy [00:43:30]:
Was a new sort of category of smart home brain. So it's an NVR, it's a NAS, it's a smart home controller hub, So Thread, Matter, and then it has a local AI. And you run your cameras, your smart home devices, you store your files, your music, your photos all on this one hub. And I was like, what if Apple made this and made it an Airport Express? It would solve all the problems of the smart home.

Mikah Sargent [00:44:00]:
It would.

Jennifer Pattison Tuohy [00:44:00]:
So, and then the idea here, we wouldn't have to be doing, using private cloud compute. We wouldn't have to be sending data to the cloud. To run our smart homes, they could run their models on device, which is what they do currently with HomeKit Secure Video on Apple Home and on the, I mean, on the Apple HomePod and the Apple TV. But those aren't, those current models are obviously not capable. And this is potentially where we may see Apple's tune change around this when we get a new Apple TV or a new Apple HomePod that potentially has more capabilities. We may not have This is all speculation, but we may not have to be ponying up for the 12 terabyte iCloud Plus plan in order to get these features of Apple Intelligence. But we shall see. Apple, the famously open company, has not shared what their plans are.

Mikah Sargent [00:44:50]:
Let me ask you this. Somebody watching this who heard a lot of what we've talked about and went, what?

Jennifer Pattison Tuohy [00:44:58]:
Yeah, sorry, a lot too much information all at once here.

Mikah Sargent [00:45:03]:
And I think sometimes times for, for us, you know, because we've got the terminology underneath, it's— and I, so I think it's more about sort of impact that it could have on the average person who in their home, they've got some lights and they've got, let's say, uh, one camera and their thermostat is smart. Um, what's one example of how this Google Home change could help them? And then what's one example of how Apple's updates with HomeKit Secure Video and Private Cloud Compute could be of service to them?

Jennifer Pattison Tuohy [00:45:48]:
Yeah, so, so say you have 3 or 4 smart devices, a thermostat, maybe a video doorbell, a smart light, and MCP might not be in your future, but if you were going to go give it a try, and where this could potentially lead with Google Home bringing these types of capabilities to the main platform and to everyone down the line is that it could— So we already have with Google Home like presence sensing. So your camera can say, you know, to tell your system that you've arrived home, or your phone can tell your and thermostat that you've arrived home and adjust the temperature, turn the light on, maybe turn off the camera because you don't want to be recorded once you've got home. So there's some of this already, but that relies very much on, you know, certain elements all falling into place. You have to have your phone on, it can't be out of battery. The facial recognition on your doorbell camera has to be correct and know that it's you instead of your UPS guy, which my Nest camera always thinks I'm my my UPS guy. So in terms of having a more intelligent brain in your home, it could— it can take all the cues from each device and, and analyze them and take out potential conflicts or doubt. Like, okay, this— the door said it was the UPS driver, but her phone did arrive home, so it is Jennifer, and we are going to adjust the thermostat. Like, it It just adds that extra layer of context to your home, make things work hopefully more reliably, but also more contextually, like understanding what's happening throughout your home.

Jennifer Pattison Tuohy [00:47:29]:
In the long term also, it can do things like— Amazon already does this, but it says it'll have a hunch that everyone's gone to bed, so it'll turn the lights off, turn down the thermostat, lock the door, even if you haven't told it to do that. That kind of context is really valuable in a smart home, but to date, again, relies on too many variables and unclear cues that in theory, an AI that can see a complete picture of your home would be able to act on reliably and hopefully not fry you in your bed at night. For Apple HomeKit users, the Apple Intelligence features that have arrived include something that both Nest and Ring have had for a year or so now, but text descriptions for video cameras. So this means instead of just getting an alert from your Apple HomeKit doorbell camera that an animal was spotted or a person was spotted, although you do get facial recognition in Apple Home today, but it will give you very specific things. For example, there's a cat sitting on the porch and Brian just walked away from the front door and the UPS driver arrived. So it'll give you a whole description of what happened. happened rather than you having to click on an alert that says, you know, someone's at the door, and then you have to watch the video to see what exactly happened. So this kind of tech, this AI-generated text description, is, is using the Apple intelligence to analyze the footage and provide you exactly what happened in a text description.

Mikah Sargent [00:49:00]:
Mm-hmm.

Jennifer Pattison Tuohy [00:49:00]:
Um, and it's, it's very useful for glanceable, like, oh, this is what happened, I don't need to worry, or, oh, okay, that's bad, I do need to take some action. And where you can see this going with Apple Home is similar to what I was describing with Google Home, is you're feeding that context rather than just a person, specific context about a package, an animal, and which person into your smart home so your smart home can be more proactive and more reactive. And that's ultimately where the smart home's going. I'm just— I would really like to see it all happening locally.

Mikah Sargent [00:49:33]:
Yes.

Jennifer Pattison Tuohy [00:49:33]:
And not in clouds that we have to pay $60 a month for.

Mikah Sargent [00:49:36]:
Amen. All right, we're gonna take a quick break and then we'll come back with a quick little story of the week from me about— it's another side of AI and making sure that we can verify when our images are our own. All right, every few weeks, the question of whether a photograph shows something that actually happened gets a little harder to answer. But Apple Well, Apple has its idea for a solution to this. On September 15th, Apple's security research team published a very long technical post on Apple Reference Image, the verification feature that arrived alongside the iPhone 18 Pro. And Malcolm Owen at Apple Insider called it fairly a mammoth effort. The short version is that Apple is trying to certify a photo starting at the camera sensor itself before the operating system ever even gets to touch the pixels. And along the way, Apple takes some pointed shots at how the rest of the industry has approached this problem.

Mikah Sargent [00:50:38]:
There's a lot that kind of is included with this and, you know, some design choices that I think are really interesting. First and foremost, Apple talks about the industry standard, which include the C2PA. Apple says, Attach provenance metadata after capture and certify the history of edits from that point forward is an approach that is vulnerable to compromise at any point in the editing chain, and a viewer has no way to detect such a failure. This is the other problem that Apple had with this. Says C2PA can also create privacy risks for photographers working in dangerous conditions by tying the image to a public identity, either to a particular device or to an individual. And so it says that Apple believes that no other commercially available photographic provenance system meets these strict requirements of limiting those concerns, which is why when I first saw this and I thought, now hold on, there's already an industry standard underway. Why is Apple not using that? It's got buy-in from Adobe, from the Associated Press, from the BBC. And then there are a bunch of camera makers who have already shipped bodies that sign in camera.

Mikah Sargent [00:51:54]:
So So why is Apple doing its own thing? Now we at least understand what the company says about why it's doing its own thing. Because this does provide proof without needing to have sort of a byline. It should, as Apple says, it should not be necessary to forego anonymity in order to prove image authenticity. And so all that needs to happen is that an outside observer Cannot determine whether any pair of reference images were taken by the same device, but still be able to understand that this photo has sort of an original version. And real quick, we'll get into the technical aspect of it. Apple refuses to sign at either end of the normal pipeline. So signing raw sensor values gives you something that you know people can't see, and then signing at the end of processing, which is what. C2PA and other systems do leaves the software chain open to attack.

Mikah Sargent [00:52:53]:
So Apple does sort of a split. Reference mode reboots the camera sensor into a specialized secure capture mode that signs pixel data immediately at capture and then blocks the sensor's firmware from touching it afterward. So it's a hardware-enforced assurance that the operating system receives pixel data exactly as the hardware sensor captured it. It gets signed separately by the Secure Enclave, and the defenses are part of the factory process because the sensor will generate its own, uh, special key on first initialization and then never releases the private half of that key. The Secure Enclave generates an attested key of its own, and then both get bound together in the device manifest. So if you pull a sensor out of one iPhone and drop it into another, Well, then that pairing no longer validates. So it is truly a, a, a full from the factory go of things. And it is very complicated.

Mikah Sargent [00:53:56]:
This, this part is especially interesting. The secure digital negative on the phone isn't a viewable photo, which is something that I misunderstood. To turn it into one, the device uploads it to private cloud compute. That runs the, uh, it's called— I can't even pronounce it. It's, it's like demosaicing. Uh, it's a very weird word. Uh, tone mapping and compression that the iPhone would normally do on device. And it does so, um, to append only cryptographically tamper-proof, uh, transparency.

Mikah Sargent [00:54:33]:
And I'm trying to find the part now. There were a lot of notes that I was trying to take for this. Um, But essentially, you have to send it to Apple for the photo to get verified through, uh, PCC. And over time, um, it will, uh, continue to work because Apple did a post-quantum, uh, signing step here. The final signature is a composite that combines RSA-3072 and another— it's called MLDSA-87, but these are just numbers for us. Um, It, it does it so that it cannot be even in future systems that you know would rely on quantum compute an ability to sort of break it. As far as what how this how this is sort of impacted for you on the end, it's an opt-in camera mode. It's on the main camera sensor of the iPhone 18 Pro and 18 Pro Max.

Mikah Sargent [00:55:30]:
The secure digital negative is stored on device in DNG format and linked to the ordinary processed photo from the standard pipeline, and it can sit there indefinitely, and it can also be shared. in the undeveloped state. And, um, that once the negative has been successfully developed into a reference image, it's automatically moved to the deleted photos folder. So the user can recover it or delete it immediately. Otherwise, it's purged after 30 days. So—

Jennifer Pattison Tuohy [00:55:58]:
The negative is purged, not the reference image.

Mikah Sargent [00:56:00]:
Yeah, exactly.

Jennifer Pattison Tuohy [00:56:00]:
Exactly. So once a reference image is created, that is as valid as the negative?

Mikah Sargent [00:56:06]:
That's my understanding. The digital negative is the The actual evidence in the system.

Jennifer Pattison Tuohy [00:56:10]:
And then it goes to the cloud compute to make the reference image. And then, but that's also verifiably correct.

Mikah Sargent [00:56:18]:
Exactly.

Jennifer Pattison Tuohy [00:56:18]:
Authentic.

Mikah Sargent [00:56:19]:
Yeah. Again, Apple is very careful about privacy protection as a step in this. And I think that those extra sort of filters and applications do make it a little confusing.

Jennifer Pattison Tuohy [00:56:35]:
And I guess my biggest question is, who is going to use this?

Mikah Sargent [00:56:39]:
Yeah, that's what I'm wondering that too, because it's an option.

Jennifer Pattison Tuohy [00:56:42]:
It's not on by default, right? You have to have the newest cameras. I mean, I guess we assume this is going to be on all cameras going forward. I don't know.

Mikah Sargent [00:56:50]:
That's my understanding. Yeah, I imagine.

Jennifer Pattison Tuohy [00:56:52]:
But right now it's obviously going to be very limited. Um, I mean, it sounds great, but it also—

Mikah Sargent [00:56:57]:
Who's going to turn it on?

Jennifer Pattison Tuohy [00:56:58]:
Who's going to turn it on? Like, and I mean, if you know you're you need to take a verifiable photo, wouldn't you just use an actual camera rather than a smartphone camera? Because isn't like a negative, an actual negative or an actual physical negative, physical negative, or a physical or an SD card with the cam, with the image on? I mean, that's just as verifiable, right? If, because the software, if you take something directly off an SD card.

Mikah Sargent [00:57:23]:
Oh, that's right. Because it has, I'm sure that there's someone somewhere who could tell us that there are special instances where you could pull something off of an SD card without the SD card knowing. But yes, I think for, for most situations, there is metadata attached to transfers and copies and edits and stuff like that that would have an impact. So you're right that there are other ways.

Jennifer Pattison Tuohy [00:57:46]:
There are other ways of doing this. If you know you need to take a photo that can be verifiably— that this is original, there are probably other ways. But now Apple's just given you an another way.

Mikah Sargent [00:57:57]:
Yeah.

Jennifer Pattison Tuohy [00:57:58]:
I mean, and, you know, the nice thing, the thing that's encouraging about this is to have that, because that is a concern that everyone has today about how images can be manipulated. And, you know, my son's in college and I'm always, I'm like, be careful, don't let people take pictures of you because you never know what they're going to do with them. And, but this isn't solving that problem yet because it's so niche and so small, but I guess it's the beginning I think so. Yeah, that's laying the foundations so that going forward, this may be less of a problem because someone will be able to say, okay, this is what you saw on the internet, but actually, here's the original photo, and I can prove this was the original photo.

Mikah Sargent [00:58:37]:
And that's the workflow, so to speak, that I feel will be going forward and kind of what I thought this was going to be in the first place. In the same way that I can open up a photo on my computer, right-click and choose Get Info, or the equivalent on Windows, or open it up in some sort of metadata viewer and see so much about it. I felt like this was going to have sort of a metadata place where you'd see the, this is the unedited version of this photo. Um, but I can also see how that, if somebody wants to share a photo of themselves and they've done a little bit of editing to remove a blemish or something, the thought that anyone, if they chose to share it, could download it and go, oh, the original had this. Let me take a screenshot and show. I could understand why Apple doesn't want to, you know, go that far with it. But I, yeah, I do imagine a future where you, where one could download a photo that someone has shared and you would be able to on device, you know, get a little check mark or something that says this is an unedited original or a little pencil icon that shows, you know, this has been edited. That would be nice to have because there are some false positives when it comes to AI detection.

Mikah Sargent [00:59:53]:
My engagement announcement, infamously, when I posted it on Instagram, it had the little AI warning above it. And that was simply because in Photoshop, I used the wrong extension feature to the crop. Just, I needed a little bit more space at the top. And I did not use old school tools. I just like quickly clicked checkbox, which of course uses AI generation. So when I made this little bit at the top, it, that part was AI generated, upload the photo, and now everybody thinks that I'm— that I was lying about getting engaged at the time. So like those false positives, I think, get in the way too. It can be complicated.

Jennifer Pattison Tuohy [01:00:34]:
Yeah, it's—I mean, and hopefully this is— if this is something— it sounds incredibly complicated, but if it's something that other phone camera manufacturers can emulate and we can start to have an option going forward and maybe it becomes the default as opposed to just something you turn on. I think there are other standards out there though. So, but it sounds like Apple was— Apple's version is very much trying to adhere closely to like its privacy pledges and making sure that an image isn't necessarily always gonna be tethered to the person that took it.

Mikah Sargent [01:01:09]:
I think, which I think is valuable. Especially when it comes to journalists. Apple does seem to do a good job of protecting journalists working in Yeah. In environments where that's necessary with some of the tools that it's released in the past. So I think this is just one more way to kind of do the verification without it going, oh, so this dictator can go, oh, that journalist is who took this photo of me. Go get them. In any case, I want to thank you so much, Jen, for taking the time to join us today on the show. If people would like to follow along with your work, where are the places they should go to do so?

Jennifer Pattison Tuohy [01:01:49]:
Yeah, you can find all of my work at theverge.com and you can follow me on the socials @SmartHomeMama, also @JP2E on the Threads, no, the Bluesky, sorry. But yeah, great, great to chat and I will take the rest of the week off from Twit 'cause I've been here for a while.

Mikah Sargent [01:02:11]:
Yes, do, do take off the rest of the week. Thank you, Gemma, we appreciate it.

Jennifer Pattison Tuohy [01:02:16]:
All right. Bye.

Mikah Sargent [01:02:18]:
Bye-bye.

Mikah Sargent [01:02:19]:
Alrighty, folks. Tech News Weekly publishes every Thursday at twit.tv/TNW. That's where you go to subscribe to the show in audio and video formats. If you'd like to follow me online, I'm @mikasargent on many a social media network, or you can head to chihihua.coffee. That's chihihua.coffee, where I've got links to the places I'm most active online. Be sure to check out my other shows on the network, including Hands-On Tech, Hands-On Apple, iOS Today, Some of those published today, one of them publishes on Sunday. It's for you to figure out. And of course, stay tuned for October 1st where I launch my new show, Hands-On AI.

Mikah Sargent [01:02:54]:
Very much excited about that and what we'll be doing there. We'll hope to see you again next week for another episode of Tech News Weekly. But until then, goodbye. Bye.