Tech News Weekly 394 Transcript

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

0:00:00 - Mikah Sargent
Coming up on Tech News Weekly. I, Mikah Sargent, am joined by Amanda Silberling of TechCrunch. We talk about X working on a program where AI chatbots will generate community notes. Then we talk about how the Supreme Court has ruled on age verification for sites that could potentially harm minors. Afterward, I talk about what Microsoft is working on when it comes to diagnostics in medicine and diagnoses, before rounding out the show with Andrew Langston of CNET, who tells us about his time with the Nothing Phone 3. All of that coming up on Tech News Weekly.

This is Tech News Weekly, episode 394, with Amanda Silberling and me, Mikah Sargent, recorded Thursday, July 3rd 2025L: Why the Nothing Phone 3 Stands Out. Hello and welcome to Tech News Weekly, the show where every week we talk to and about the people making and breaking that tech news. I am one of your hosts today. My name is Mikah Sargent and, as is the way I am joined across the internet by the wonderful Amanda Silberling. Welcome, Amanda.

0:01:27 - Amanda Silberling
Hello from the other side of the internet.

0:01:30 - Mikah Sargent
It's so weird how it works. It's just a series of pipes and or tubes.

0:01:35 - Amanda Silberling
I am closer to the holiday weekend than you are.

0:01:38 - Mikah Sargent
You are, and now I'm thinking is a pipe a tube, or is a tube a pipe, or are they both each other?

0:01:47 - Amanda Silberling
I don't know, ask Mario.

0:01:49 - Mikah Sargent
I'm just thinking like the Mario pipes. If there's anyone who knows, it would be Mario. So, for people who are tuning in, this is not a show where we discover whether squares and rectangles matter, but instead it is a show where we talk about our stories of the week. To kick things off, Amanda, tell us about what you are bringing to the table today.

0:02:14 - Amanda Silberling
Yes, so I'm bringing to the table an article about how X the Everything app as you can only call it Twitter, or X the Everything app can't call it X, x the Everything app is piloting a program that lets AI chatbots generate community notes, which I find this interesting because it's a way in which, like, I'll just call it X for the sake of brevity instead of actually everything out. But um x has kind of like started a trend in social media companies of their community notes feature, which tiktok and meta and I believe youtube have also piloted, like similar programs, which is basically like when you go on x and you see like someone will tweet um, did you know that Mikah's favorite fruit is an apple? And then the community will be like well, actually known famous podcaster Mikah said in an interview that his favorite fruit is a strawberry. So that's you know. Got to correct that. Obviously, the actual context are about about like politics and things like that. Um, but it's very much like human driven, volunteer driven.

So the idea of including ai in it I think is interesting, where the knee-jerk reaction is like why are we just further taking humans out of the content moderation system, which it can be dangerous to have just AI content moderation without it being checked by humans.

But then there is a research paper where these researchers that work on community notes were basically showing how this might work, where it might be that the AI note writers are proposing notes for posts and then they get read by the humans and then, like, you still have that human oversight.

But then also you get into questions of, well, what ai are people using for this? Because you they are planning to do it in a way where you can like plug in something through the api. So then what happens? If you have something like the overly sycophantic open ai model that was like telling everybody like this is a great idea, no matter what, where, then you end up with, like, are the community notes going to be written from a specific perspective or, more likely, to be agreeing with what is in the actual content of the post itself? Question of, can AI do what humans can do? But if this is being done through an open API and a bunch of different AIs can be contributing to this, are we going to end up with problems like just all of the community notes are about like South Africa or something.

0:05:22 - Mikah Sargent
Yeah, I mean, that's the big thing, right Is so okay.

Here's the deal.

In my opinion and I got to do that because this is my show I think that there's something to be said for a more neutrally trained AI being able to provide context and community notes, in this case, for things that people maybe it's things that people like regularly get wrong.

For example, dogs are allergic to chocolate. No, dogs are not allergic to chocolate. They simply cannot process the two nervous system stimulants that are in chocolate, which are caffeine and some like something, something. They're not allergic to it. They just can't process those things and therefore in their body they build up and it causes them to have strokes and seizures and da-da-da-da-da, that kind of thing where it's well-known, you know fact-based, scientific-based knowledge that could provide for context that maybe a human like is just not going to go and look for every single time that someone's posted about dogs being allergic to chocolate and, you know, respond to it, although I'm sure there's somebody out there who would. So, given that, I think that there I don't necessarily have an existential problem with AI partaking in providing context, but we know that Grok, in particular, is trained in such a way that it is. It's not great in terms of what we've come to expect from AI and speaking with sources familiar with the matter, aka someone who works in Grok training, oh, you got to hook me up with sources.

I specifically have been told that they and others are mind boggled by the fact that Grok does something that other AI systems don't do, and that is that Grok will sometimes ignore its own training and other AI bots that exist online. Other chatbots will not ignore their training, but instead will be convinced and sort of hacked right. We've seen that before, where it's sort of jailbroken right, but that's because the human being is working toward figuring out a way to get it to go around what it's doing. No, grok will just do that.

So I don't want it fact-checking, I don't want it figuring out what needs to be done. I get this idea of volunteer overload, particularly in the new way of X, the everything app, where there aren't as many daily active users and there aren't people, I would say, as devoted to fact checking as there once were. So in that way, I understand wanting to do something to, in theory, improve upon it, but so much of it feels like it's not improving upon actual facts, and that's the problem that I think we all have is that we can't even agree on what a fact is. What do?

0:09:07 - Amanda Silberling
we do. Yeah, I mean, I think that this is also interesting because AI and human content moderation working in tandem is not a new thing. Like this is how. Like this is how social media has been operating for like decades I don't just as long as I can remember but the way it would work is, let's say, like you are on facebook and you post a sexually explicit image because you're just doing some things on facebook and then the facebook ai visually detects that is something that should not be on facebookcom, and then it can flag it. And then, if it is confident enough, then it'll just be like nope, we are taking that down. Or if it's like more dubious, then it goes to a human moderator, who then is like yep, that should not be on here. But then, um, facebook also, or meta generally, is not using human moderators in the us in the way that it used to.

But then this, the way that the ai works with something like using it on community notes, is different, because it's like it's not trying to recognize, like is this an image or are these certain words that are often used in contexts that violate platform rules? It's looking at like someone wrote that dogs are allergic to chocolate based on my training data, do I know if that's true or not? And then you might end up in a situation where, because this is such a commonly stated thing, maybe the ai will assume, based on its training data, that it's correct that dogs are allergic to chocolate and then not go to, like the one source from, like vetsorg I don't know, that's not a website, but the, uh, the, the conclave of veterinarians discussing chocolate. They might not use that as the source and be like.

0:11:14 - Mikah Sargent
Actually, all of these reddit posts about chocolate and dogs are not quite correct yeah, I, well, and you've, you've really touched on the big thing there, right, the fact that training builds on itself, and that's the big issue when you have human beings doing most. You know it's in pilot. Um, I just, there was a time where a pilot meant we're going to try it and if it doesn't work well, or if it makes, if it's inaccurate or it's bad, then we'll stop doing it. But I don't have faith that this version of the social media network has any interest in approaching things from that perspective, right, like I don't know what the goal is anymore at the company.

0:12:22 - Amanda Silberling
Yeah well, they are piloting it for now. We'll see if it rolls out more fully. But yeah, I just thought it was kind of interesting too, because I kind of come in from a bias of being generally in favor of humans. But there are times in which AI is helpful at making things more efficient or doing things that humans don't want to do, which for something like content moderation. This often ends up being like using facebook as the example again, like you have like low-paid workers that are just being shown pictures of like violence over and over again and their job is just saying, yep, that's violence and that is very like mentally draining and bad. So it'd be cool if an AI can do that. But then also you're putting like something very important in the hands of a technology that is not extremely reliable with this sort of thing.

0:13:22 - Mikah Sargent
Yeah, yeah, we. You know I don't spend a whole lot of time on X-Everything app, but I will be curious to see as the pilot probably inevitably becomes a regular thing how the AI does in terms of fact checking. And again, particularly when it comes to Grok, we do need to take a quick break here. As I mentioned, joined this week by Amanda Silberling.

And I want to tell you real quick about our pals at Zscaler, who are bringing you this episode of Tech News Weekly Zscaler, the leader in cloud security.

Hackers are using AI to breach your organization. Ai powers innovation. It does. It drives efficiency, but also helps bad actors deliver more relentless and effective attacks. Phishing attacks over encrypted channels increased by 34.1%, fueled by the growing use of generative AI tools and phishing as a service kits. Organizations in all industries, from small to large, are leveraging AI to increase employee productivity, with public AI for engineers with coding, assistants, marketers with writing tools and finance creating spreadsheet formulas. They're using it to automate workflows for operational efficiency across individuals and teams. They're using it to embed AI into applications and services that are customer and partner facing and, ultimately, to move faster in the market and gain competitive advantage. Companies they've got to rethink how they protect their privacy and protect their public use of AI to defend against AI-powered attacks.

Jason Kohler, chief information security officer at Eaton Corporation, leverages Zscaler to embrace AI innovations and combat AI threats. He says data loss detection has been very helpful for us. Chatgpt came out. We had no visibility into it. Zscaler was our key solution initially to help us understand who was going to it, what they were uploading. Traditional firewalls, VPNs and public facing IPs expose your attack surface and are no match in the AI era. It's time for a modern approach with Zscaler's comprehensive zero-trust architecture and AI that ensures safe public AI productivity, protects the integrity of private AI and stops AI-powered attacks. Thrive in the AI era with Zscaler Zero Trust Plus AI to stay ahead of the competition and remain resilient even as threats and risks evolve. Learn more at zscaler.com/security. That's zscaler.com/security and we thank Zscaler for sponsoring this week's episode of Tech News Weekly.

We are back from the break and it is time for my story of the week. The Supreme Court's ruling in Free Speech Coalition v Paxton has upheld a Texas law requiring age verification for access to pornographic websites and in doing so, may have opened the door to sweeping changes in online privacy. Texas's H 1181 law demands that any site with more than one third adult content verify the age of every visitor, framing it as a child protection measure. But critics argue that this decision invites intrusive surveillance, forces users to surrender anonymity and risks chilling access to lawful speech, not just for teens but for everyone. So let's talk about what the law actually says and what the court ruled.

Texas's bill HB 1181, requires commercial websites with at least one third quote material harmful to minors. So that's kind of a broad definition to verify the age of all users. Now the Supreme Court did uphold the law, ruling that this burden on adult users is a tolerable trade-off for protecting children. You gotta protect the youth. Justice Clarence Thomas equated it to showing ID at a liquor store, quote the risks of digital age verification aren't meaningfully different from showing your ID at a liquor store, but there are some privacy concerns. Age verification requires users to upload sensitive personal data like government-issued ID or biometric information, unlike liquor store ID checks, where you show your ID, they look at the date or they even type in the date, or sometimes they scan your card to get the date. This data could be stored by third party verification vendors, which, of course, are ripe for breaches, surveillance or misuse. I mean, look at the different places where, right now, the tech companies don't want to do the work themselves in terms of verification, so they hire a third-party company. I know that if I go to, I believe, the social security website or the IRS website and log in, I have to use a third-party tool to verify myself that I am actually me. Not even the government wants to build its own verification method. These companies that are for-profit organizations aren't going to want to take the time in many cases to do this. That's the argument that's being made Now. The Verge notes that there are currently no federal regulations that currently protect this kind of user data, but the Electronic Frontier Foundation says quote age verification systems are surveillance systems. The personal data disclosed via age verification is extremely sensitive and this is the big thing Unlike a password, often cannot easily or ever be changed. So with it being upheld in Texas and setting that precedent.

There's concerns about what critics are calling the end of online anonymity, because the law effectively forces everyone, not just minors, to identify themselves before accessing legal adult content. Users without government ID, especially marginalized, undocumented people, could be locked out of accessing these sites. The ruling could also normalize and this is the big thing digital ID checks across a wider range of online content. The Verge says we know legislation limiting adult content has chilling effects, even when the laws are rarely used, because, once again, that's where it starts, but it can build out from there and that's where those kind of chilling effects come in right. The law targets any site where one third of content is harmful to minors but again, kind of that subjective standard, and the big thing that another big thing that critics are worried about is that if that is the definition content harmful to minors then website operators will probably err on the side of censorship in order to avoid legal risk. Legal risk and, depending on what state you're in, LGBTQ+ content may be considered content that is quote harmful to minors, which would result in people not having access to resources that they would otherwise have access to, and that, of course, comes with its set of concerns, of course, comes with its set of concerns, so this is kind of. You know where we are right now.

I wanted to ask you, Amanda, just in general, like, have you and I'm not talking specifically about these kinds of sites, but I've been to lots of different sites where there's some sort of age verification process involved and where things used to be as simple as just putting in your date of birth and then they go well, we've done our job. Things are getting a lot more complicated than that. I guess it's a time for us to harken back to the early days of the web, when you just trusted that somebody was putting in or not even necessarily that. They trusted that you were putting in the right date, but that that's all their job was. That was all their responsibility was was just to go OK, you told us that's what we need what you said.

0:22:09 - Amanda Silberling
I've only really encountered that sort of thing with, like trying to apply for like tsa pre-check or something and using um, like idme or one of those third-party sites to verify your id.

And yeah, I mean, I think that when people hear about these age verification laws, they might think that it's the same thing as, like when I was 11, and I wanted to post on the club penguin forums and I was like, haha, I'm 13.

But it's basically, in order to determine that someone's an adult, you have to, or to determine that someone is not a minor, you have to determine that they're an adult. And then that just means that there's potential for a lot of content, not just pornographic sites, to be age-gated, which, as you mentioned, there's language around the quote-unquote one-third of a website being, quote sexual material harmful to minors. And if you look, throughout history there have been a lot of different ways that calling something sexual material harm for harmful to minors, like that material like maybe today that is like actual pornography, but in the past maybe that was like the existence of a happy gay couple going about their business and in a non-pornographic way. And there's a really big difference between just like gay people exist versus like just age-g. Like you like, you can kind of make the argument. If you are, you can make the argument based on the wording of the laws that, like basic resources about, like sex education, could be, uh, limited.

0:24:01 - Mikah Sargent
Yeah, and that, I think, is is my main concern with this. For sure, I and I mean again, I think, a concern that a lot of people have. You know, we've seen how this has played out in other ways already. Some adult sites, like Pornhub, have previously geo-blocked access in Texas over similar laws. I wouldn't be surprised if once again, it happens. Other states are expected to introduce or expand similar laws, and, especially now that it's been upheld by the Supreme Court, we're likely to see that even more, and without those federal privacy protections in place that were once there, the ruling may accelerate.

This is the concern the erosion of online anonymity for adults under the guise of child safety. This is something that has played out. We've seen in many places. It's not just about pornographic content, but as a whole. We've seen this protect the kids, and again, this is the thing Everybody wants to protect the kids, and no one is against protecting the kids, and so, because that is so central to what any reasonable person wants, it is also a good way to push forward legislation that could otherwise maybe use a little bit more examination, and I think that's what we're seeing play out here. So it's really it's troubling, and you know one thing that I would love to see happen.

So Apple is one company that has been working for a while on these digital IDs that you'd have a driver's license and that it's a digital version and you can open it up in your wallet app, and that's great.

But the good, the cool thing about it is, in states where it's accepted, you can take your phone and tap it against the device that is checking for your age and the only information that is shared is yes, this person is over 21, or no, this person is not over 18.

It doesn't even tell the system what your actual birthday is, which is actually a lot more privacy protecting than even handing over your ID as we have it now, because the person at the cash register can see your address. They can see other things on your card. So this idea of becoming even more private, that all it's saying is a nod or a shake based on what you've been doing it's very good at being able to protect one's privacy, and so if we could see that as a means of you know, doing this age verification right, but age verification that is as simple as yes or no, as opposed to these systems right now where, yeah, you're having to upload. In some cases you're having to record video of yourself. You are it's, you know it's a liveliness check. All of that is just troubling and I now worry, with this legislation passed at the highest level of our legislative body, how this goes forward. So it's unfortunate, yeah.

0:27:35 - Amanda Silberling
And I think a fundamental difference too and like the example that you mentioned about like apple working on some sort of digital id versus um like uploading your id to another website is that when apple does these sorts of things, like face id and other like biometric data verification, they're doing that on your device and it is not going through the cloud to apple.

And this is also something that, like the company that runs porn hub, has been arguing that like they think that on-device verification is a better uh answer to the problem of making sure kids don't see stuff they shouldn't see, while also not, like making the whole internet surveilled. And it's also interesting how their own response to this has been that in um, I believe 24 states so far have uh legislation similar to the Texas law that was discussed in the supreme court, but in those states uh pornhub literally just like isn't operating because they don't want their users to have to upload their ID into something, and I think that kind of also goes to show that it's like this company is surrendering so much of its traffic and so much of how it makes money.

0:29:01 - Mikah Sargent
Yeah, absolutely. And I think that when it comes to that, as we see, because ultimately there's still taboo, of course, around this and so it becomes more difficult, I think, to argue and get a lot of support behind it, because people might not want to come out and say you know, stop banning Pornhub or whatever it happens to be, you'll get some people who do that, but it doesn't get that huge group behind it, right? That's not going to be the case when it comes to some more mainstream although I'm sure Pornhub is rather mainstream, it's just no one's going to say that but more mainstream and acknowledged sites that may also get, have this, get in the way and cause this, this drop in traffic. So I think that there's going to be pushback. It just hasn't hit the sites yet that will get that kind of roaring collaboration as they all band together Les Mis style and storm the barricades.

Amanda Silberling, I want to thank you so much for taking the time to join me today to talk about these great stories of the week. Of course, folks can head over to TechCrunch to check out the work you're doing, but where else might they go, including giving a little listen to your podcast?

0:30:25 - Amanda Silberling
Yeah, aside from TechCrunch, you can find me I'm mostly on BlueSky at Amandaomg LOL, which is a handle that makes me laugh every time I say it. And I also have a podcast called Wow If True, which is about internet culture and is co-hosted with my friend, the author, isabel J Kim. So kind of a bit of sci-fi in there as well, but mostly internet culture and just general the weird underbelly of tech.

0:30:57 - Mikah Sargent
We love the underbelly. Thank you so much, Amanda. We'll see you again soon. Thank you All righty. We're going to take a quick break before we come back with one more story of the week, and then I've got a great interview for a very interesting phone after that.

This episode of Tech News Weekly is brought to you by CacheFly.

For over 20 years, CacheFly has held a track record for high-performing, ultra-reliable content delivery - serving over 5,000 companies in over 80 countries. At TWiT.tv we've been using CacheFly for over a decade, and we love their lag-free video loading, hyper-fast downloads, and friction-free site interactions.

CacheFly: The only CDN built for throughput! Ultra-low latency Video Streaming delivers video to over a million concurrent users. Lightning Fast Gaming delivers downloads faster, with zero lag, glitches, or outages. Mobile Content Optimization offers automatic and simple image optimization so your site loads faster on any device. Flexible, month-to-month billing for as long as needed, and discounts for fixed terms. Design your contract when you switch to CacheFly.

CacheFly delivers rich-media content up to 158% faster than other major CDNs and allows you to shield your site content in their cloud, ensuring a 100% cache hit ratio.

And, with CacheFly's Elite Managed Packages, you'll get the VIP treatment. Your dedicated Account Manager will be with you from day one, ensuring a smooth implementation and reliable 24/7 support when you need it.

Learn how you can get your first month free at cachefly.com/twit. That's C-A-C-H-E-F-L-Y dot com slash twit.

All right back from the break and I've got one more really fascinating story for you.

Sometimes, when I find these stories and I'm unable to get the researchers involved, we move on, but this is just so cool that I still wanted to take the time to talk about it. So Microsoft yes, microsoft has unveiled a new AI system that tackles some of medicine's most difficult diagnostic challenges and this is the big thing outperforms human physicians at both accuracy and, as you can imagine, cost efficiency. The model called the Microsoft AI Diagnostic Orchestrator, which is like MAI-Dxo it's M-A-I hyphen D-X-0, isn't just answering multiple choice medical questions. It's actually simulating the step-by-step decision-making that doctors use in real life. Tested on cases from the New England Journal of Medicine, MAI-Dxo solved 85.5% correctly, which is more than four times the success rate of a group of experienced doctors. I think this has long been the promise of outcome when it comes to protein folding, finding new medications, finding new ways of interacting with the body, but what we have not seen, I think, is a real and true use of the contextual awareness that AI can provide to better provide an answer to these medical questions.

Not too long ago, I remember watching every episode of House. I was sick and I needed something to do, so I was working on, in fact, the little house that's behind me. I was working on that and watching house, and the cool thing about house is you have a group of doctors and you're taking in this data. Here's what's going on with the patient. Here are the symptoms. Here are the things that they've experienced in life. Here's where they live. Here's all of this stuff. Let's figure out what's going on with the patient. Here are the symptoms. Here are the things that they've experienced in life. Here's where they live. Here's all of this stuff. Let's figure out what's going on with them.

And when AI started to become this generative AI craze that we have now, there was this immediate excitement that we saw, where many experts and those of us who aren't as much experts but are certainly enthusiasts said, okay, this is just like a diagnostician, the person whose job it is to take in data points and provide an answer. And this is very exciting because, where a human being can only have so much that they're able to pull on at a given time, as long as the resources are there, the AI can kind of grow and grow and grow and have more and more information and use all of that contextually to provide an answer. Now here is the interesting thing about how Microsoft approached this. Rather than doing static quizzes, which we've had in the past, in terms of testing the medical accuracy of one of these systems, microsoft built the sequential diagnosis benchmark, otherwise known as SDBench. It's a simulation of real-world diagnostic problem solving using complex case records, again from the New England Journal of Medicine. The benchmark allows the AI to iteratively ask questions, order tests, weigh, cost and revise its reasoning before reaching a diagnosis. Now I would argue that I want to leave out the part where it weighs the cost of things, because let's just try to get the answer. But it does that too, and yet was still more accurate and saved more money than a group of doctors. Now, each requested investigation, of course, incurs a cost. So it really talks about and bears in mind those real world healthcare expenditures.

Mydxo, it's not a single model, but here is what makes it unique it's what they call a virtual medical panel, because its job is to actually orchestrate multiple AI agents with different specialties and strategies. So it is working to guide each step of the diagnostic process. It does follow-ups, does test choices, does diagnosis, does cost checks, does self-audit and its job is to turn these different language models into a virtual panel of clinicians. So it goes to the AI systems that we know and provides the proper information, it gets back recommendations for tests, it looks at the cost for these different tests and provides answers based on what the actual outcome of the case was, and then they benchmark it against real doctors. So it was 304 cases from the New England Journal of Medicine.

Mydxo, paired with OpenAI's O model, correctly diagnosed 85.5% of cases. Now this was compared to a control group of 21 practicing physicians in the US and the UK, and this group of human beings averaged just 20% on the same tasks and with higher diagnostic costs. So, as it says, mydxo delivered both higher diagnostic accuracy and lower overall testing costs than physicians or any individual foundation model tested. So that's important to bear in mind too, because you may be looking at the option to say, okay, I'm going to take all of the information that I have about this case and I'm going to pop it into one of these large language models and look for what it suggests that we do next, what its responses are, et cetera, et cetera, et cetera. That's fine, but by working with multiple language models, it again is sort of building out its own system.

A good question from Burke in the chat, who asks were the cases cherry picked? That is something that, as you might imagine, microsoft is not going to directly point out, but that is something that's worth looking into. It was published in a peer-reviewed journal and so, in that case, you know you do have at least that to count on. But again, that would be something to look more into for sure. So the AI's decisions were evaluated not just for correctness but for test efficiency, which, of course, is something that we have to bear in mind as human beings working in real healthcare systems and work to avoid over-testing, which can help to regulate cost.

So let's talk about how this could be used. Going forward, microsoft says look, we're still in the early stages of research. This isn't FDA approved, it's not clinically deployed, no one's out there using this system, but they hope that AI can augment consumers so that you know, we might be able to just upload a oh, I've got this weird rash. Take a photo, upload it and get some information that you could then take to your doctor and say hey, I ate this and I put this on my skin and this has popped up. Here's the context you need to help me figure out what's going on here. My recommendation don't say I asked chat GPT what's wrong with me and this is what it said. That's not going to turn out well.

What's wrong with me? And this is what it said. That's not going to turn out well, and clinicians of course can use this in routine and complex care. It of course needs to be tested on more typical, everyday medical cases. So, burke, it sounds like these were particularly difficult cases. If you're getting published at this point in the New England Journal of Medicine as a medical case, it's likely that what you're getting published at this point in the New England Journal of Medicine as a medical case, it's likely that what you're working with here are more mysterious cases, the house cases, as it were. So I think in that sense it is a bit of a cherry pick, because you're looking for kind of mysterious, hard to understand cases. But they say AI could empower patients to self-manage routine aspects of care and, of course, equip clinicians with advanced decision support for complex cases. So let's see it move from static benchmarks to dynamic clinical reasoning and from individual AI chatbots to this orchestrated system that's built for real world deployment.

I remember a time where I had something going on that I didn't know what was going on and my doctor at first thought it was shingles. And I knew it wasn't shingles because it wasn't just presenting on one side of my body, and I had to do. I figured out what it was At the time. Ai was not where it was today, so I didn't use that, but I figured out what it was through a bunch of research I did myself and literally had to link to journals to share with my doctor and say I'm almost certain that this is what's going on and this is the medication that, despite not you know on label, would help with this. And my doctor sent it off to the dermatologist, who was like, yeah, I'm pretty sure that is what he's got going on. We should give him this medication and it fixed it. And not everybody has the time and the resources and the wherewithal to be able to do all of that work, and so being able to have a system in place that could help with that, I think, is a great thing. So go check out. We'll, of course, link the blog post about this. I think it's very exciting. I am I'm definitely keeping an eye on this and seeing how it goes forward. I'm definitely keeping an eye on this and seeing how it goes forward because I think that, in a time where we already have a medical system that is so overburdened. This could be a helpful thing and may result in healthcare not being as expensive as it is as well, which is also a great thing. So that is what Microsoft has just published.

All right, we will be talking to our next guest in just a moment.

All righty, we are back from the break, and that means it's time for our guest Joining us from CNET. It's Andrew Langston. Welcome to the show, Andrew. You had the opportunity to review what I find to be a fascinating device, the Nothing Phone 3. And I read your review and thought it was really quite interesting, so I wanted to talk to you about it. You call the Nothing Phone 3, and I think the company does too kind of the company's first true flagship. So can you talk about what makes this phone stand out as a flagship compared to its predecessors?

0:45:44 - Andrew Lanxon
Yeah, so that's very much the language that they're using at the moment. I was down at the launch event on Tuesday so I've been able to spend some time with it, but there's a lot of testing and things to be done before I'm going to be necessarily agreeing completely with their marketing messaging. But their flagship claims really, kind of as it always does comes down to the specs involved On nothing's previous phones, like the Phone 3a that it launched back in March, it's going for lower-end processors, lower-end camera units, all in the name of kind of finding that balance between performance and price. And so, while previous phones were maybe between $300 and $400, this one is starting at $799. Now, that's obviously a lot more money, but it is packing some extra power under the hood, let's say, with the Snapdragon 8S Gen 4 chip, better camera units so it says and even some interesting tech with using a new battery tech silicon carbon battery which it claims will hopefully just give it a little bit more power and longevity.

0:46:53 - Mikah Sargent
Interesting. So one of the things that I think sets nothing apart is its design language across the whole system. I would love if we could start by talking about the Glyph Matrix, this dot matrix display. How does it kind of evolve nothing's quirky design language, and what kind of evolve Nothing's quirky design language, and what kind of information or interactions can users expect from it?

0:47:15 - Andrew Lanxon
Sure, yeah, so Nothing's. Previous phones they've all had this sort of light-up interface, these scattering of LEDs on the back that the company's called the Glyph, and previously they've basically been a little bit functionally useless. They just sort of flash if a notification comes in, but otherwise there's not a whole lot of meaning there. And with the Nothing Phone 3, they've essentially clustered those together and put it into this little dot matrix display that sits in the top right hand of the phone. Because it's an actual little display it can actually show things, little pictures, little bits of text and stuff. So it's functionally a lot more useful because it can actually show maybe the time, maybe it will show the name of an incoming caller, but it's also thrown in what it calls the I think the Glyph Toys, various little kind of mini games that you can actually use to interact with the phone.

But one of the ones in particular that you can actually use to interact with the phone, but one of the ones in particular that sort of stood out to me was a way of playing spin the bottle just using the glyph toy on the back of your phone. Now maybe it's something that's sort of lost in translation, but for me. Spin the bottle was a game you play at teenage house parties with a group of people, but they're very much like oh no, if you're splitting the bill with adults at a restaurant and that's what you'd use it for, I'm like OK, if you say so, guys.

0:48:39 - Mikah Sargent
Yeah, that's interesting. I mean, yeah, I kind of want to talk about that in general. Are these sort of differentiators in a crowded market? Do you think it actually offers long-term utility? And maybe just kind of in general, how do you feel about these kind of here's what we're doing, that's a little bit different and we've added an extra button. Do people from your experience end up sticking with that and using that, going forward and expecting that going forward? Or maybe nothing is just the brand that you go to to get an extra button or get a spin the bottle glyph toy.

0:49:21 - Andrew Lanxon
Yeah, I mean I'll be fascinated to know if being able to play spin the bottle from the back of your phone is the thing that really separates this out. I'd be surprised to be quite honest, but I do think that there is an element of these things and the previous lights being kind of gimmicks but I do still think that it does separate them out. It is something now that the company is known for doing, and at a time when most phones being launched are essentially the same grey rectangular slabs. Certainly, as a journalist observing the industry, it's nice to kind of finally have something a little bit different to say, something that I haven't written about in previous reviews. So it's keeping me happy, if nothing else.

In terms of long-term longevity of these tools, I think some of them are actually quite useful. You know the basics of simply having the time on the back of your phone or being able to see an incoming caller when your phone is face down. Certainly that could be quite useful. They've also opening the SDK to third-party developers and their community. So I imagine maybe things like maybe a countdown to when your Uber is going to arrive, something like that, would potentially be quite useful. So I think really it's going to have to be a case of wait and see see how people get to grips with this kind of thing and, of course, see whether it makes much of a difference in their sales versus their competitors, I suppose.

0:50:47 - Mikah Sargent
Absolutely. Now, one of the features of the phone is a triple 50 megapixel camera setup. That's interesting, and, of course, improvements in image processing, which I think for many a modern smartphone is the differentiator in many cases, more so than the hardware, but how the images are processed. How do these changes stack up against the flagship competition from Apple, samsung, google, and how does it stack up against what nothing has offered up to this point?

0:51:20 - Andrew Lanxon
Yeah, I mean, on paper it looks pretty good. They're saying all the right things. It's obviously, as you've rightly said, it's not just about the megapixels, it's about the software. It's about how it is then integrated deeper into the phone, and so they have said a lot about how they've improved their ISPs. They've also talked about using large image sensors and wider apertures so it can gather a lot more light than its predecessors.

Now, that's great, because those are are things that make tangible difference to the photos that you take. It makes a big difference to the quality of your night mode photos, and that's absolutely something that people are going to be looking for how it actually stacks up against the iPhone 16 Pro, the S25 Ultra I'll be honest, I'm very excited to find out, and I cannot wait to put it side by side against those. It's too early for me to be able to have done that just yet, but I have hopes that it will do well. I would be surprised if we've got a new phone camera champion on our hands. The previous ones have been a little bit of a letdown camera performance wise. I think it's certainly one of the areas where they've maybe tried to squeeze a little bit more off that bottom line in order to charge lower prices. The cameras have been okay for the money, but certainly you wouldn't buy it if you were especially keen on your photography. So definitely looking forward to seeing how the flagship model compares to the more budget options.

0:52:51 - Mikah Sargent
Absolutely. Now I'm curious to hear how Nothing handles software, in particular this time around, because that's one of the things that I think kind of sets these different phone manufacturers apart when it's outside of what Apple provides. Did you notice any changes in Nothing OS that enhance the phone's personality or perhaps make it more practical to use or less practical to use? Are there many changes in Nothing OS with the launch of this device?

0:53:25 - Andrew Lanxon
There's not a whole lot visually from previous models. What I would say is, if there's anyone that is kind of considering this phone and has never used a Nothing phone before, it is a very different experience, simply because the company puts on quite a strong skin, as it were. It changes a lot of the visual interface. It relies on this very monochrome look. The apps themselves, even things down to Chrome and YouTube and other Google apps, have been turned all black and white with this very stark look, which, on the one hand, is kind of cool.

You definitely have an interface that looks different to any other Android phone. It looks different to the iPhone. On the other hand, it's kind of annoying because you are just looking at just black and white apps, and I don't know about you, but for me, I really need those visual cues for me to be able to pick out certain apps, and I spend half the time when I'm using the phone just going hang on. Which one was this? Is this one the gallery? I'll load it up. No, that's not the gallery. That's like the SIM tool or something. So I find it very confusing, but it's certainly an interesting look and I think that is probably something that people will enjoy looking towards.

0:54:39 - Mikah Sargent
Absolutely. There's mention of AI integration, which I think you can't make a modern phone and not mention AI integration somewhere via nothing's essential space. What is that essential space and how does it factor into the day-to-day experience of using the phone?

0:54:56 - Andrew Lanxon
The essential space essentially is a repository for your. It's basically a brain dump space as you have your stream of consciousness throughout the day. Your essential space is kind of where you would put things down, kind of like in a journaling tool. It launched it on the phone 3A earlier this year. I had a player around with it when I reviewed that phone and I actually think it's a good use of this kind of like semi-AI sort of thing. There's a hardware button on the phone and as you press that in you can record a voice note. That voice note then gets saved to your essential space along with maybe a screenshot. So let's say you're browsing the web, you found something you want to buy. You can kind of take the screenshot and record a little voice note with it, or you could take meeting notes, or you could take a photo of something and it will save it in that essential space and then try and give you a summary of of that later on.

Um, so this is sort of ai light, in that it's the ai's. They're kind of picking out some of the key features. Um, it is baking ai a little bit more, uh, deeper this time and there's the I think it's called the essential recording um. That's basically um does voice transcript and will try and pull out the key points of, let's say, a meeting that you're recording in order to give you, firstly, reminders of that but also actionable elements if that's maybe booking a meeting or buying something or anything like that. So it's kind of interesting to see how it's doing this as a busy journalist having a very easy access voice notes tool that will summarize things, Give me the reminder later on. I'm kind of into that.

0:56:41 - Mikah Sargent
Yeah, I mean I got to agree. I think that's kind of a cool feature. I'd like to have that on my phone. Now with the final question for you here, with a price point starting at 7.99, with that snapdragon 8s gen 4 chip, which is a mouthful, and a promised seven years of security updates, do you think that this nothing? Phone 3 hits the right balance of premium features, uniqueness and value uh, I mean, that is a.

0:57:11 - Andrew Lanxon
That is the question. I think they have priced it very cleverly. It's actually at the same price as, I believe, the base Pixel 9. So it undercuts the Pixel 9 Pro. It undercuts the iPhone 16 Pro, but not by a huge amount, so it definitely will still feel premium. There will be that feeling that you are buying a more premium product simply because of that price point. I am expecting it to have good performance, certainly a lot better than its predecessors, whether it is competing with the incredible benchmark scores of the S25 Ultra or the OnePlus 13,. We'll have to wait and see, but it should do pretty well on paper and for me, I think it's really going to come down to the camera performance as well. Cameras are pretty much the main way that many of the top phones are differentiating themselves, and if OnePlus, if nothing, genuinely wants to call this a flagship, it needs to be competing with imaging at a flagship level, and here's hoping it will.

0:58:18 - Mikah Sargent
Here's hoping Well. I want to thank you so much for taking some time to join us today to talk about your review of the Nothing Phone 3. Really appreciate it. Of course folks can head over to cnet.com to check out the work that you're doing. Where else could they go if they want to keep up to date with the stuff you're publishing?

0:58:34 - Andrew Lanxon
You could find me on Instagram with at Battery HQ. That's where most of my stuff goes. Otherwise it's all on cnet.com.

0:58:43 - Mikah Sargent
Awesome. Thank you so much for taking the time today. We appreciate it. Thank you, alrighty. We have reached the end of this episode of Tech News Weekly.

The show publishes every Thursday at twit.tv/tnw. That is where you go to subscribe to the show in audio and video formats. Of course, I'd love to remind you all about our wonderful club Club Twit at twit.tv/clubtwit. When you join the club, you can join monthly or yearly, and joining the club gets you access to some pretty awesome things. First and foremost, you gain access to our awesome Discord server, a fun place to go to chat with your fellow Club Twit members and those of us here at Twit. You gain access to the Twit+ bonus feeds. Those feeds include our clips and bits. You also get our news events and you get our special club shows and I think there's like more than a week now of content that exists in the club, if not more content that exists in the club, if not more, and all of that is yours as soon as you join the club. So it's a great way to like have a huge back catalog of really fun stuff. And, of course, you gain access most importantly for many of you to the ad-free versions of all of our shows. So if that sounds good to you, head over to twit.tv/clubtwit.

If you would like to follow me online, I'm at Mikah Sargent on many a social media network where you can head to chihuahua.coffee that's C-H-I-H-U-A-H-U-A dot coffee, where I've got links to the places I'm most active online. Be sure to check out my other shows, including Hands on Apple as well as Hands on Tech, and if you've got time this Sunday, be sure to join me Sunday morning 11 am Pacific time, as we record for the month of July for Hands on Tech. We'll have a great time doing that and look forward to seeing you there. And, of course, you can also watch iOS Today, the show that I record with Rosemary Orchard every week about all things iOS and iPadOS. Thanks so much for being here, thanks so much for tuning in, and we'll catch you again next week for another episode of Tech News Weekly. Bye-bye.

1:01:01 - Leo Laporte
The tech world moves fast and you need to keep up for your business, for your life. The best way to do that twit.tv On this week in tech. I bring together tech's best and brightest minds to help you understand what just happened and prepare for what's happening next. It's your first podcast of the week and the last word in tech. Cybersecurity experts know they can't miss a minute of Security Now every week with Steve Gibson. What you don't know could really hurt your business, but there's nothing, Steve Gibson. What you don't know could really hurt your business, but there's nothing Steve Gibson doesn't know.

Tune in Security Now every Wednesday. Every Thursday, industry expert Mikah Sargent brings you interviews with tech journalists who make or break the top stories of the week on Tech News Weekly. And if you use Apple products, you won't want to miss the premier Apple podcast Now in its 20th year MacBreak Weekly. Then there's Paul Theriot and Richard Campbell. They are the best connected journalists covering Microsoft and every week they bring you their insight and wit on Windows Weekly. Build your tech intelligence week after week with the best in the business. Your seat at tech's most entertaining and informative table is waiting at twit.tv. Subscribe now.