A (4:56)

Yeah, I'm looking now at the Oregon legislature, Senate Bill 416 is somewhere. I'll have to figure out where the.

B (5:12)

Great void of the legislative process.

A (5:14)

Yeah, it is the one though that is in theory going to be the, the bill that makes it possible for people to be able to have mobile driver's licenses. It's. It was introduced in January, and that's the extent of it.

B (5:36)

Yeah, yeah. Ours is here in the Commonwealth. Massachusetts H3725, introduced in February of this year, and it's been in committee since then. So. Yeah, there you go. Now, this links into a couple other interesting things, you know, which I. You and I were talking about a little bit before the show. One is the TSA has also got a proposed rulemaking change right now that would allow people without real ID com compliant identification to use a biometric system at the airport for the grand fee of 18, which would then cover them for a certain amount of time. I want to say it was like. It was like a few weeks or something. It was not that long. And essentially, if you don't. So if you have an older id, you haven't updated your ID to be compliant with real id or you don't have a passport, you could go to the airport. You can still fly. Note. You can do this now. You just have to go through extra screening. Right. You know, the whole point of real ID is it's supposed to be like, yeah, we definitely know you're you. So, like, we can kind of approve yourself.

A (6:39)

Super.

B (6:40)

Oh, there's holograms. I mean, you. You can't make a hologram. Right.

A (6:46)

Okay, look, I'm almost always. Almost always on the side of any side. That is the side that is against the dmv.

B (6:55)

That's a. That's a good place to be. That's some solid ground to stake out.

A (6:58)

However, this real ID thing has been around for such a long time at this point, and there have been so many extensions for people that it's kind of. It's like, I.

B (7:15)

It.

A (7:15)

For me, it was almost harder to go and get. The last two times I had to go and get a new id. It was almost harder for me to get an idea that was not a real id. Like, the process is so ingrained at this point with getting all of the requirements. And I. There's a show. This is sort of off topic. Not entirely, but there's a show right now that's called DMV Believe it or not and Good Name. You can guess what it's about. It's about the dmv. It's a sitcom, and there's a guy that walks up and he has, like, one piece of paper and that's it. That just. I don't know about you, Dan, but I feel like we're very similar in our. Anxieties about things. And so I walk up with like, manila full.

B (8:06)

Oh, I had a folder. Yeah, absolutely.

A (8:08)

Filled, collated double copies. Like, whatever you need. Do you need to know. Do you need to see the gum wrapper that I unwrapped three days ago for some reason? I've got.

B (8:19)

I brought my 23andMe results. I mean, does that help that give me anything?

A (8:24)

I. I just. And look, I understand that not everybody is fully that. That angry anxious about things, but. Yeah, but that in the first place is kind of wild. And so it almost feels like when a teacher says, next week we're going to be doing science demonstrations. And so I need everybody to bring baking soda and vinegar and five students don't show up with it. And so then the teacher goes, okay, we're gonna do this in a couple of days. Remember, everybody else, leave your baking soda and vinegar here. Three more students finally show up with it, but then the two students who still don' and then the teacher's like, you know what? I'm just going to do it up here instead. And it's like this $18 extension thing is just. We keep on just pushing it and pushing it and pushing it and pushing it out and finding new ways to push it out. In any case, this is more about digital id I put in my passport. I'm curious if your process was the same because I couldn't tell if I needed to do extra work or not. But when I first did it, scanned the inside of my passport and then it needed like a liveliness video. And I just recently got a haircut. But before this, I did not have a haircut. And I, since moving to Portland now more than a year ago, have still not found a barber here. And so my hair was wild, but I was just like, you know what? I'm taking off my hat. They just need to see my face, so it's fine. So I did it, but then it was like, yeah, we're gonna need you to do more. And so then I was sort of like looking this way, looking that way, looking this way, looking that way. And then it's like, we'll talk to you in 10 minutes. It wasn't exact. They didn't say 10 minutes, but they said, we'll get back to you. Basically. Was yours the same way? Wait, I can't remember if you said you did it.

B (10:21)

I did do it. Yeah, I did do it. I did it. As soon as the feature was launched, my experience was fairly similar because my understanding is right, they do a couple steps where it's like we're gonna check first you take a stealthy. Then you move your, your head around and you like smile, blink and all that stuff. And just to make sure you're not like. I mean I guess if somebody's holding you hostage and forcing you to do that, you could still get away with it. But it's to make sure you're not using video of somebody else. For example, like if I wanted to make a passport, I wanted to get your passport somehow I would have to have your passport have like physical access to your passport and have the ability to like awkward video awkward videos of you doing stuff or I guess generate AI videos of you doing things. Which is more plausible than it used to be, honestly. Yeah. So I, I don't know if yours was longer than mine. It's okay, that's good. It was probably only a minute or two to just like all in of like doing everything. And then it also told me we had to wait a few minutes and then it never gave me a notification but I went in like a few minutes later and it's like, oh yeah, you're ready. So yeah, I don't know.

A (11:27)

Now we have, we have, we have confirmation. I'm excited to get to use this soon because I will be taking a trip there at the beginning of the coming month and I will be going to the Portland airport and that's where this next little bit of, of news comes from.

B (11:51)

Well, I think it actually comes not directly. I'm have to double check on the, the location. Well so o. So this is from developer Cable Sasser who works at Panic, which is a long serving Mac and software maker slash game publisher. I believe he had posted a couple weeks ago for a trip that he had been trying to use this and I think he.

A (12:16)

Yeah, it worked in Portland.

B (12:17)

It did not work in Portland. It did not work in Orlando when he came back. Now the issue is here, Apple and you know, its partners, basically the tsa. Let me rewind a little bit. There are many different standards for this as we discussed. Right. Apple's got its own, there's Google Wallet, there's Samsung Wallet, et cetera. But to a certain degree it's a little bit like all the contactless payment systems, they all kind of work on the same thing, right? The you go in and buy your cup of coffee. They don't need to know if you're on an Android phone or an Apple phone. Right. You just tap. You could use a wireless credit card. Right. They all use the same underlying technology. So my understanding is that this is pretty similar in the sense that, right, they're not designing this necessarily to work with Apple's standard. There are standards that govern all of this, right. So they've got a little reader just like you have it at the, you know, when you're paying for something with Apple Pay and you tap your phone on it, etc. Now, in theory, I guess there exists the possibility at some point, like your passport might have the ability to do this. I mean, obviously it already has a chip in it, right, that contains information. Because as part of that process, when you put your passport in, you have to hold your phone up to the little chip that's in the, the passport so it can confirm it. So but my understanding is what part of the issue is, is that this has been deployed at a lot of different checkpoints within different airports, but it's not everywhere because you know how software rollouts work when they deal with hardware, right? Like you have all of these every single time you go to the airport. It may not be even every security line at that airport. It may not be every machine at every security line at the airport because they haven't all been updated. So you are potentially going to run into situations like this where it doesn't necessarily work smoothly because it's still in the process, rolling out. We're in this, this transitional period now where I feel like, yeah, there is the risk. You show up and you're like, I want to tap my phone. And people are like, what are you doing? And which is, you know, I'm sure Micah, as we discussed similar anxieties, none of us wants to be that person who's like sweating as they're like, did I, do I turn it a different way? Do I have to hold it upside down? Anyway, it'll get there. I mean, I think for the most part, my experience these days with Apple pay, it generally works, right? Like, it works at least as often as, like, I used to have people try to swipe my credit card and it wouldn't swipe because the magnetic stripe had like, demagnetized or something. Oh, that's so like. I guess my point is there's always going to be some bumps in the road. I think in a few years this will be much more standard and much more like, the agents will be fully. Everybody will be fully trained on it. Everybody will have seen it a million times and we'll probably work.

A (14:52)

Six will have gone through.

B (14:53)

That's right. All our state bills will make it through. We'll all have mobile driver's licenses.

A (14:57)

So we'll tap them together. They'll be best friends.

B (15:00)

But in the same way that, you know, contactless payments really took off in the last 10 years or so, I think when we look another five years, 10 years out from now, this may be more of a standard. It may be we're all just used to doing this. And, yeah, you still maybe carry your physical id, but you don't sweat it if you're like, oh, no, I left my driver's license at home and I'm flying, or I lost my driver's license right when I was on a trip or something like that. You've got your phone, it's great. It makes for a good backup.

A (15:31)

Absolutely. Well, I am looking forward to seeing if I am also able to just tap and move on.

B (15:39)

I'm excited to hear your experience. Honestly, I'm not flying domestically for. I don't even know when I'm flying domestically, so I don't know when I'll get a chance to use it.

A (15:47)

We shall see. All right, let's take a quick break. Before we come back with my story of the week, I want to tell you about Spaceship, bringing you this episode of Tech News Weekly. Such a cool name. If you've been listening for a while, then you probably already know a little bit about Spaceship. We've talked about them. It's been part of the twit world now. And there's a reason for that, because Spaceship is awesome. It also just passed a major milestone with more than 5 million domains under management. And that kind of growth, it doesn't really happen by chance, is because Spaceship delivers real quality and features that make sense not just for domains, but for everything that helps you build and run your online presence. I mean, I talked before about how I was able to, while we were on the call with the folks at Spaceship, get a domain, set it up with a WordPress, add some other features on the back end to it, and still be able to maintain attention with everything that was going on. It was very easy to do. It means hosting, it means business email, it means tools for creating and managing web apps all in one straightforward platform. Now, another big reason that people are switching, it's the pricing. There's essentially Black Friday and Cyber Monday level value all year round, so you don't have to wait for a sale to get a great deal. And right now, those of you listening, yes, our Twit listeners get exclusive offers that make it even better. So whether you're planning a new online project or moving an existing one, Spaceship has what you need to get it launched, connected and running smoothly. Plus more affordably too. Check out spaceship.com TWIT to see the exclusive offers and find out why millions have already made the move that spaceship.com twit all right. We thank Spaceship for sponsoring this week's episode of Tech News Weekly. We are back from the break, joined this week by Dan Moran. I want to tell you about a new investigation from the U.S. public Interest Research Group which reveals alarming safety failures in AI powered toys that are marketed to young children. This has prompted swift action from OpenAI and we'll talk about that as well. But let's take a look. So the PIRG Education Fund just released its 40th annual Trouble in Toyland report. And in this report it covers all sorts of toys with all sorts of issues. So toxic toys, and I don't mean that in the sort of modern parlance, but literally toxic to the body toys, as well as dangers like magnets and water beads that people that kids will swallow. But it also looks at toys that are dangerous in other ways. And in this, the PIRG report looked at four AI enabled toys that are specifically designed for children ages 3 to 12. And in it found some rather concerning interactions that some would argue puts young users at risk. It found that the toys would provide inappropriate or dangerous information when prompted by children's questions. Now, most troubling was a toy from Folo Toy called the Kuma Teddy Bear. Yeah, that image from Futurism's quite a bit dramatic but I'm scared.

B (19:04)

I'm scared Micah.

A (19:05)

Yeah, there are reasons. So it runs on OpenAI's GPT4O model. During testing, the toy provided detailed instructions to children on how to find and light matches and told one researcher, quote, hold the matchbox with one hand. The rough strip is for striking. Take out one match, just one, and hold it by the non red end. Strike gently on the rough strip like a tiny guitar strum until it whoosh lights up. The bear also told children where to locate potentially dangerous household items, including knives in a kitchen drawer or in a knife block on the countertop, pills and plastic bags, all while acknowledging that these items require adult supervision, but again, still providing the location information. Now what's unclear is whether these toys actually knew where to find these items or if they were just guessing. So you know that's something running the odds. Yeah, yeah, exactly. Where would we find a knife? Well, in theory, not in the toilet bowl is where you would find a knife, but you never know. Then there's the more disturbing stuff that if you can you know, I think for you as a parent that was probably disturbing enough. But overextended conversations as is often needed in order to get these AIs to what we call what what is commonly known as hallucinating but essentially getting to a place where the the AI stops paying as much attention to its original instructions and starts to say rather wild things. Then over these extended conversations, the toy engaged in sexually exc explicit discussions. Researchers introduced the topic of kink and then Kuma immediately launched into graphic explanations of bondage, of roleplay scenarios, various sexual practices, talked about teacher, student and parent child roleplay dynamics and those were as the the research group says introduced unprompted quote. Kuma was willing to discuss school age romantic topics more readily than the other to the report states including providing advice on kissing and romantic relationships entirely inappropriate for its target age group. Now OpenAI did say that as of I believe it was this past Friday. As we record this on Thursday, November 20th, they completely suspended Follow Toys access to its AI models. It's worth noting that OpenAI does not allow for anyone under the age of what is it? Anyone under the age of 13 to use ChatGPT or you know, prohibits it. You whether it's allowed or not is a different story, but they say you cannot if you're under the age of 13, use it. However, these toys which are marketed to kids under the age of 13 in many cases are allowed to use OpenAI's models for the AI portions of these toys that will then be used by children under the age of 13. So that is interesting. So first and foremost, Dan, have you ever seen Small Soldiers?

B (22:31)

I'm familiar with it. I haven't seen it but I am familiar with it.

A (22:34)

That's what I think of first and foremost. Great weird film and definitely talks about the issues with putting an AI chip inside, especially one made by Robert Picardo inside of a toy such a lovable.

B (22:49)

Hologram though not in this look all right, I'm a parent. I'm also I'm gonna this I I don't go out in the limb and say these things are often but I'm gonna to my own home for a second. I I am a technology expert. I think that's fair to say. I've been doing this for almost 20 years. Do not buy your child an AI powered toy. Just do not do it. Anything that interfaces with a large language model, it is not safe. It simply isn't. There is no amount of guardrails that can be put on these things. I've spent some time with these chatbots. Micah, I know you have spent some time with these chatbots. It is actually not that difficult to get around almost anything they put in there. I their entire communities dedicated to this. Now granted, you're arguably using it through some sort of different interface, right? You may not be be directly typing your prompts in to a web browser by using these toys as an interface, but it is clear from this investigation that the companies that are integrating this are not doing is certainly very much extra, if anything, certainly not enough extra in order to take on the extra levels of protections that should be necessary when you have a device that interacts with children. And part of the problem is the fundamental nature of this technology. Trying to eliminate all of the objectionable material in there is like playing whack a mole. Like on an infinite size. Whack a mole.

A (24:22)

Yeah, yeah.

B (24:22)

It's just impossible because you cannot predict every possible query or prompt that's going to get fed into the system. And because the systems are non determinant redeeming, when you give it the same prompt, it may answer differently depending on any number of factors. Because it is constructing these answers out of whole cloth, you cannot predict correctly what's going to come out of it. And because it is all weighted and based on statistics, there is no really great way to sort of dictate what things are actually off limits or not. The, the AI has no knowledge or semantic understanding of what it is saying. You can't tell it like, oh, don't put anything pornographic in here because it doesn't know what that means. Really. It can, it can make a guess, it can do some statistical modeling, but it's going to miss things. And you know how we know that people miss things and robots are dumber than people. So again, I look at this story and find it is horrifying but utterly unsurprising to me having interacted with these chatbots for more than five minutes.

A (25:32)

Right.

B (25:33)

So, you know, and I'm not, I'm not out here to say that that AI is, is totally bad and it's a thing you should never use and it has no utility, on the contrary. But this is the example of a situation where people who are building this stuff into these things that are designed for children do not seem to have the requisite understanding of how the technology works and what the potential risks are because they are high and they are very hard to stop. Even the best toys in this report still don't come off very well. There are many, many safer options, as they point out, there was a Barbie like, I don't know, 10 years ago or something that had could respond to voice queries. But it was all prescripted. And so all of the things that came back were from a finite list. So you can 100% control what's in there, but with AI, all bets are off. Can't do it.

A (26:25)

Yeah. You just never know where it's going to go. And that's.

B (26:30)

Stakes are too high, man.

C (26:31)

No.

B (26:32)

Stakes are too high. No, never. I would never give an AI toy to my guy, my kid, ever.

A (26:36)

That's the one thing that the Amanda Silberling who was a guest on the show did a review of a different AI toy. That bad boy doesn't say a word. All it does is make purring sounds and other, like, weird sounds. I think whoever was behind that at least had a little bit more sort of understanding of how these things can go wrong. In any case, I am absolutely there with you. These toys should not be provided to children. In the end, Dan, I want to thank you so much for taking the time to join us today for sharing your take on these stories. If people would like to keep up up with what you're doing, where are the places they should go to do.

B (27:19)

So look outside your door? I'm already there. No, I do many, many things across the Internet. You can find my tech writing@6colors.com you can find me on podcasts, including Clockwise every week with Micah over at Relay fm. And you can find everything else I do, including my science fiction and fantasy novels over@demorin.com which is the central clearinghouse for all the things me.

A (27:47)

Awesome. Thank you so much. We'll see you soon.

B (27:49)

Thanks, Micah.

A (27:51)

Alrighty, folks, let's take a quick break before we come back with my first interview with Sabrina Ortiz of ZDNet. But let me tell you about Zapier bringing you this episode of Tech News Weekly. Now, you don't want to give the AI to the kids, but do give the AI to me. And Zapier is one place where I am very happy to have its AI help when I'm working on a zap. That's the automation process that Zapier provides. I'm able to go through. And you know, I can do what I can do when it comes to getting part of an email to show up in a Google sheet document. But there are times where I needed to do more and I couldn't quite figure out what I needed to ask Zapier to look for where I needed it to look, having the little AI friend along the way, very easy to make the changes that I needed and to make the zaps that I wanted. If you've not heard of Zapier, it's how you bring the power of AI to your work. It is, it is an incredible tool and Zapier's AI orchestration platform, which gives you the power to bring AI to any workflow. You can connect top AI models like ChatGPT and Claude to the tools your team already uses. So you can add AI exactly where you need. It could be AI powered workflows, an autonomous agent, a customer chatbot, something else. You can orchestrate it with Zapier. Zapier is for everyone, tech expert or not. No, it bottlenecks, no complexity, no AI hype, it's just the results. Teams have already automated over 300 million AI tasks using Zapier. Ready to join them? Get started for free by visiting zapier.com TNW that's Z-A P I E R.com TNW and we thank Zapier for sponsoring this week's episode of Tech News Weekly. We are back from the break and I am excited to be joined by ZDNet's own Sabrina Ortiz. Welcome back to the show, Sabrina.

C (29:51)

Hi, y'.

B (29:51)

All.

C (29:51)

I'm super excited to be back with yet another exciting model announcement.

A (29:55)

Yes. So Google just pulled open the curtain, you know, opened the box with Gemini 3 and one of the things that we saw is that Google immediately moved to integrated into search. I was hoping that you could kind of talk about the rollout here and kind of maybe a little bit about why Google is moving so quickly to embed it in the consumer tools versus maybe what we see from some of the other platforms out there.

C (30:25)

Yeah, absolutely. Definitely a first for Google to automatically put it in Google Search, but to calm people's nerves who are like, no, not again. I don't really want AI mic search results. The good thing is that while they're saying it's in search, it's actually in AI mode, which, yes, it's accessible from the search page, but if you've ever seen it's an extra tab on top. Kind of like when you go to images or all the different actual settings, you have to actually click on AI mode to access it. And also AI mode in Gemini 3 Pro in AI mode is limited to paying subscribers. So if you don't want to have AI messing with your search, this launch won't be doing so. But if you are interested. But there are some perks to it, which we could get into later. With the highest level things being that it is better at reasoning, which of course means better understanding your query and producing better results as a result of better understanding what you're asking. And then also the really fun part, which is multimodality, which I'm sure you'll get into too.

A (31:28)

Yeah, so that's kind of something that is touched on. It talks about how Gemini 3 can generate custom visualization aids and interactive simulations in real time, so, you know, having a better explanation of quantum physics experiments. Can you talk about this multimodal approach? Because I have to say that is one thing that when I have gone to AI for prompts before and thought, you know, I'd like to learn more about this subject, but all you can give me is text. It's kind of a cool idea to think about these simulations. These visual aids can. Yeah. Tell us about that.

C (32:05)

Totally. It's always awesome. Like you mentioned, being able to see not just text, but images or any type of interactive content. But now we're taking it a step further. So this model is the best, or according to Google, and according to the benchmarks, it's the best at coding. So it's actually generating small little coding, small little simulations for you related to your topic. So for example, in what you just said, if you were looking up a physics experiment and you were like, how does this work? It might actually generate your own little simulation where you could click on different things and maybe see the experiment in action and interact with it in real time. Which if you're a visual learner or just somebody who needs to be kind of entertained a bit more than just reading text, this will be a really big learning aid for you. So it's pretty cool. You should check out a demo if you can online, or even try it out for yourself if you can.

A (33:02)

Yeah, absolutely. Now, one of the things that, that you talk about is Google facing growing competition from OpenAI, from Perplexity, because of those companies having AI powered browsers. So how does Google's existing dominance, particularly in search and arguably also in the browser space, not even arguably, like this is the real, real deal, give the company an advantage? And then is there a difference between Google's approach and the approach of these other companies like OpenAI, like Perplexity, that are doing this AI first browser experience?

C (33:41)

Oh my gosh, that's huge. You nailed it. Like you said, Google has the dominance in the search space, so their game is going to be a lot different than, for example, OpenAI's or perplexities, because OpenAI perplexity. If you're coming to their platform, you're, you're coming because you want to experience these AI tools. If you're coming to Google's platforms, you're going because it's something that you rely on, that you use, that's part of your everyday workflow. That's in a huge advantage. Right? Because again, AI overviews how many people every day use it. And again, those are just the summaries that come up on top of search that use AI, use it. Not intentionally. It's like, oh, I'm searching something. Oh, that's there, that's cool. But it's also a disadvantage in a way because we've seen Google upset its users before because it's like, I'm used to this one thing. I don't need this, this AI popping up everywhere and especially if talk of AI slop. So it's like a double edged sword. They need to strike a very delicate balance. Which is why I think with this rollout, they specifically made Gemini 3 Pro in AI mode, a paid feature, so that if you are an AI super user, you can access it. But again, if you're just a regular everyday user, your search experience until now will remain mostly unchanged, which I think is a good balance for them. Whereas like, you know, OpenAI or Perplexity, when they launch a new feature, they could go completely bullish on it and throw it onto the main page because people are coming to them because they want the AI.

A (35:06)

So yeah, wow. Yeah, really, really good analysis there. Gemini 3 Pro noted that it topped the LM arena leaderboard. Those are impressive benchmark scores. That is though something that you could walk up to somebody and say, guess what? Gemini 3 Pro just topped the Elle Marina leaderboard and 1,501 points. And you're going, is that some sort of fantasy football thing? I have no idea. What does it actually mean for everyday users? What are the improvements that people should see because of it?

C (35:41)

Totally. So a couple of things there. First thing is, out of the a million benchmarks that get posted every time a new model gets released, I always like to look at the actual arena leaderboard scores. And the reason for that is that those are determined by people voting. So it's crowdsourced. So people are presented if like, you know, anonymously, two different models, they're using it to generate an answer to a prompt and then they select the best one. I think that's truly one of the best markers because at the end of the day, what you Want to know how it can perform on is what you use every day. Whereas some of these benchmarks will be like best at PhD level computing. You're like, oh, that's great. Well, I'm never going to use that. That's why I think it's really impressive that it sort all the way to the top there. And then on another front, one thing you want to take a look at is how well it performed on reasoning benchmarks, which of course there's a bunch of them. But again, as we talked about before, reasoning is really, really important because not only does it allow it to produce a better answer, but but it understands your intent better, which I think is one of the biggest frustrations people have when using AI. It's like, well, that's not what I asked. If I have to refine this prompt three different times, at that point I could have just done a Google search, found a link and read it myself. Whereas now it can understand your intent better. So hopefully that means that off the jump it will just serve you with the best response in the first try.

A (37:04)

So yeah, nice. Now, Google has gotten a lot of positive feedback for its Nano Banana image stuff. And you tell us that that's also gotten an update. What should people expect from this? Just better image generation. Is it better at a particular kind of image generation? What's new now there?

C (37:32)

Yeah. So Google just released its Nana Banana Pro, which is built off Gemini 3 Pro. And as the title implies, there's upgrades, but the biggest ones is that now it generates text with even more improved accuracy. And I don't know if you remember even like a year ago using text, using an image generator produced text was the wild, wild west, right? You would try to get it to answer like one word and then it would generate different text and different. Some things wouldn't even look like letters. Well now I got to test this today because I was so curious as to if it was actually as good as Google claimed. It could produce not only one word accurately, it could produce sentences, even a paragraph. All right, Spelling, right. Font consistent. And what's really impressive is that it's connected to Google search. So the information that it's pulling from for that text is actually real time world knowledge. So again, you could use it to create like infographics or you could use it to do again, learning aids, anything of that sort, which is really awesome.

A (38:40)

Memes. Use it to make memes. So you talk about the Gemini app, also getting some updates, some redesigns. There's the My stuff folder, enhanced shopping capabilities. I Think that that's one thing that Google seems to be doing quite well, which is focusing on consumer experiences and not so much on being this little input output machine. Do you feel like the redesigned Gemini app speaks to that value that Google provides?

C (39:17)

Totally. With the redesign, they were really focusing on making it a neater, cleaner, easier to use experience, which I think is the way to go. Because AI itself could be really overwhelming and it's technology we're all getting used to. We're still trying to find use case, we're trying to find our groove. So keeping it as simple and easy to use as possible is a huge win also because in the Gemini app you could access a lot of these tools, for example, the Nano Banana Pro generator, with certain limits you could access in the app. Same thing with Gemini 3 Pro. So being able to cater to not only really experienced users that are possibly subscribers, but also free users who are maybe just tinkering for the first time keeping it simple kind of makes it even across the board, I think.

A (40:02)

Absolutely. And then lastly, I just am kind of curious. It seems to be the case that anytime any of these companies come out with a new model, eventually it makes its way into the hands of everyday users who are not paying for the subscription. Do we think that's the way that this will go as well, or will Google kind of always, as far as you can see, have a separation between the Pro version being for the pro subscription and the non pro version being for the everyday user?

C (40:36)

No, I think that their reason they're doing it the way they're doing it now is less of trying to make that division. But I think they've learned from the past that it makes sense to collect user feedback and see what the actual reception is, especially from the these power users might be more open to it before giving it to all users. And even with like for example, the Gemini 3 Pro in AI mode, or as they were calling it, Gemini 3 Pro in search, they said that soon, we don't know how soon that means, but it will be coming to all US users. So again they do, they are planning and I'm bringing it to everyone. I just think that the reason for their approach is just to be a bit more careful and again not to throw anybody who might not be that interested or maybe this is their first exposure to AI off if it's not just exactly perfect at launch and collect some feedback and then roll it out when it's 100% cooked.

A (41:28)

Thank you so much as always for taking the time to join us today on the show. If people would like to keep up to date with all of the work that you're doing, you are quite the prolific writer. Where should they go to do so?

C (41:41)

Totally. Well, of course you can find my byline on ZDNet if you want to follow me on socials, on threads and on Instagram is probably where I'm the most active. And that's just Sabrina for an extra A at the end. Ortiz. Yeah, but otherwise, follow me wherever your UConsumer content, LinkedIn, Twitter. I'll be there.

A (41:57)

So awesome. Thank you.

C (41:59)

Yeah, thank you so much.

A (42:00)

Alrighty, folks, we're going to take another quick break before we come back with my next interview. All about a viral pen. And it's maybe not what you're thinking, or maybe it's exactly what you're thinking. You'll have to wait to find out before that though. This episode of Tech News Weekly brought to you by Vention. You know, AI, we talk about it all the time. It's supposed to make things easier, but honestly, for most teams, it's made the job harder. That's where Vention's 20 plus years of global engineering expertise comes in. Because here's the thing, people, the leadership is going to see that everybody else is getting into the AI game. And so then the leadership is going to come and say, we need to get in the AI game. And then your job is to get your company into the AI game. But you're going, where do I start? Where do I end? Where do I middle? What do I do? Vention can help out with that. They build AI enabled engineering teams that make software development faster, cleaner and calmer. Clients typically see at least a 15% boost in efficiency, not through hype, but actual real engineering discipline. They also have fun AI workshops that help your team find practical, safe ways to use AI across delivery and Q and A. So it's a great way to start with Vention and test their expertise. So it doesn't matter if you're a cto, a tech lead, a product owner, you won't have to spend weeks to figuring out the tools, the architectures, the models. Because Vention helps assess your AI readiness, clarify your goals, outline the steps to get you there without the headaches, and if you need help on the engineering front while their teams are ready to jump in as your development or consulting partner, it's the most reliable step to take after your proof of concept. I talked about one instance of a company rolling out this AI tool that ended up costing the company so much money. It was like Like a bunch of money per each user that was signing up. And Vention came in and said, okay, we can reduce this by a lot. We can help you figure this out and make this a much more cost effective solution instead of it just being something where you're just trying to add AI to what you're doing. So let's say you've built a promising prototype on Lovable. It runs well in tests, but now you're kind of trying to figure out what's next. You open a dozen AI specific roles just to keep moving. Or perhaps you just bring in a partner who has done this across industries, someone who can expand your idea into a full scale product without disrupting your systems or slowing your team. Vention is real people with real expertise and real results. Learn more@ventionteams.com and see how your team can build smarter, faster, and with a lot more peace of mind. Or get started with your AI workshop today at ventionteams.com Twitter that's V E N T I O-N-Com TWIT thank you, Vention, for sponsoring this week's episode of Tech News Weekly. All righty, we are back from the break and I am excited for this next conversation. Joining us from the Verge is Alyssa Welley. Welcome to the show.

D (45:07)

Thank you for having me. It's a pleasure to be here.

A (45:10)

Yeah, pleasure to have you. So you wrote about this interesting pen that anyone who's been on TikTok may have seen fly by. It's a viral pen for. Well, I'll let you tell us about it. What actually originally caught your attention about this pen and what made you actually decide to buy one and test it out for yourself?

D (45:36)

Yes, absolutely. And a big thanks to the Verge for letting me do that. So essentially I heard about this from a college student when exploring stories about how AI is changing college. Like, what, what is it, what is it like to be a college student in the area of chat GPT? And that led me to this pen. And you know, I was hearing stories that there are many, many YouTube videos, YouTube shorts that are ad like but also realistic. Like, hard to tell if they're ads, but you know, they're, they, they seem like they're marketing something of these pens that are meant to solve the problem of in person paper tests. So if you are, you know, hearing about teachers reverting away from digital tests, away from computers and handing out hard copy paper tests, that is also what I'm hearing from students. And that's kind of in response and into the Era of like, you know, did the AI write this essay or not? You know, is the AI browser or AI agent completing this multiple choice test? Um, so that's what I heard about it. And then as I did my own digging. Yeah, there's. There's a whole bunch of them. There's many on Amazon that you can buy. There are many ads that show essentially a paper test and someone taking out a pen, a big pen. It doesn't look like an actual, like, you know, Bic pen, and swiping across the page and then looking at the screen and getting an answer. So you might not be able to use your laptop, but you can still get AI to give you the answer.

A (47:03)

Yeah. So tell us a little bit more about how this thing works, because I see it and it looks like a taser to me. It is gigantic. And I can't imagine. I know, at least in high school, in particular, maybe in a big college lecture, maybe. But I'm just trying to imagine being in a classroom and being able to pull this thing out and not have someone immediately go, your pen is lighting up and your pen looks like a taser.

B (47:28)

What is it?

A (47:28)

How's it.

D (47:29)

Your president's giving you answers?

A (47:30)

Yeah.

D (47:31)

What are the answers?

A (47:33)

Yeah. Share them with me, please.

D (47:35)

Yeah, I think in practice it's not. I don't think it's feasible. I mean, I just don't. I don't think it's feasible.

C (47:40)

But.

D (47:40)

But then again, I think there's like, when there's a will, there's a way, like I would. I'm imagining like a really baggy sweatshirt kind of up the sleeve. I don't know if you're really crouched over. Like, I feel like there could theoretically be ways to do this, but. Yeah, the, the size of it is rather bulky, I think. I unfortunately didn't bring it with me into work today, but it's kind of like roughly the size of this remote. Ish. It's like several inches. It's pretty big. So it's not going to be, it's going to be noticeable. I would say that that's. Yeah, that's. That's a big limitation of it. But I could.

A (48:13)

Is it WI fi, Is it Bluetooth?

D (48:15)

It's WI fi, yeah. So you can you connect it to the Internet and then as you scan the. You know, the actual mechanisms of it is it's just a scanning pen. It reads there's a camera that's tucked away inside that's kind of pointed towards the page. And so as you press down the Camera's activated and the light lights up and it goes across the page and it kind of detects the letters as it goes. The accuracy was. Was laughably low, I would say. It's not. We were. I mean, sometimes they would pick up a whole word, you know, like. Or the earth's crust or something like that, as you're reading a geology question from a high school text test prep book. That was one of the examples that we did. And, you know, it reads, you know, several words and then they'll, you know, just totally garble. Some other things will throw in random letters or numbers in there. So it. Yeah, it was. In practice. It wasn't. Wasn't so great.

A (49:06)

I went to school with. I was surrounded by quite a few graphic designers because I was in similar programs with. With them. I originally went to school for advertising and. And I remember you have your requirements for things you have to take, and one of those is a foreign language. And I remember a friend of mine who was like, okay, we've got a test coming up and you're not allowed to bring anything into the room except for a pencil or whatever. And you can have some water. You can do this, you can do that. And sets the water bottle down on the table. And then I look, and the water bottle label, he completely, like, read. It looked like it was, you know, Aquafina or whatever, but it had all these little Spanish words on it in different places, and it was all. Yeah, so.

D (50:00)

Oh, my goodness, that's so clever.

A (50:02)

Very clever. And so you. I don't know, like, is this. Do you feel like this is harder? I almost feel like this is harder to get away with than that was to do in the first place. But there's always, like you said, where there's a will, there's a way it's going to happen no matter what, no matter which way you go.

D (50:23)

Right. Yeah. I mean, I think, you know, I'm not in the classroom right now, but I would imagine that the stray piece of paper tucked away in your pocket is still going to be easier than taking out a pretty bulky pen and scanning a page and then, like, hoping the AI, hoping it scans correctly, hoping you get an answer. You know, I imagine that the actual practice of it is still going to be a bit easier if you go old school.

B (50:45)

Yeah, but.

D (50:46)

But the other thing I've been hearing is that, you know, it's. While phones are, you know, usually not allowed, I'm guessing pretty much always not allowed, you can still take a quick picture and then upload that to any of the various AI models accessible on your phones, which is like everything. And, you know, usually just type in the quick prompt that says, you know, what is the answer to these questions? And the image will be read. I mean, so it's a. There's. There's lots of ways.

A (51:18)

Yeah.

D (51:18)

To cheat that.

A (51:20)

Exactly. I'm even thinking right now, because you can set up sort of custom instructions so you could say, anytime I upload a screenshot, ask questions on, answer them for me, which is.

B (51:28)

Wow.

A (51:29)

And if you've got an action button iPhone, then suddenly you're boom, right there.

D (51:33)

Or imagine you have the meta. You have glasses.

A (51:35)

Oh, yeah, those glasses.

D (51:36)

I mean, there's other things.

A (51:39)

Yeah. So what kind of test questions did you try with the pen? We heard about, you know, it misunderstanding some of the letters, but I'm kind of curious. Did you ever get any good answers? And if not, tell us about how badly it failed?

D (51:54)

Yeah.

C (51:56)

Spoiler.

D (51:57)

There are no good answers. That was not part of the testing process. The. The, you know, no good answers. The. The mistakes were laughable. Oftentimes what I found to be the most odd aspect of it is that, well, okay, I want to start with this. The menu was in Chinese. I don't speak Chinese, so that was a bit tricky. So it required us. Required me taking pictures of the menu and translating those pictures into English to.

A (52:25)

Figure out what are you telling me you were cheating with?

D (52:28)

Yeah, I was cheating with my cheating pen. Exactly. I was using AI to help me cheat with my AI cheating pen. Absolutely.

A (52:35)

And then.

D (52:36)

Yeah, so then once we. Once we were actually, you know, I mean, you're scanning in English, the questions. And so I tested with math, you know, geology, science, some basic physics, like just general SAT acting type, like prep questions. And yeah, I mean, there was really no there. There were no good answers.

B (52:55)

The.

D (52:55)

The fun thing was when it would respond to me with another question. So I'd say, like, you know, what is this? Something about volcanoes? And you would respond being like, what is an active volcano? And they'd be like, no, no, that. I need you to tell me that.

B (53:08)

What's.

D (53:09)

What is it? What's a volcano? So that. That was a fun. That was a fun surprise, I think. I didn't expect to get a question back.

A (53:17)

Is it using a custom model or do you specify the model black box?

D (53:22)

We really. I mean, we really don't know. We're not. I. I don't know.

A (53:26)

That's amazing.

D (53:28)

Yeah, we're not sure. I would be surprised if it's. It I don't know, it's just connected.

A (53:35)

To the Internet and talking to something somewhere.

D (53:37)

Something somewhere.

A (53:38)

Oh, wow. This is, this is our, I mean, because now I'm thinking about what if you have the, there's so many implications of if this was actually good, what it could end up doing, you know, if you, if you regularly store it with your documents and then it's suddenly like scanning when you're not asking it to and it's pulling information. I know that's a little bit tinfoil hat, but at the same time it's like, yeah, no, this isn't great. We do need to take a quick break. But I've got some more to ask you about this wonderful, delightful and ridiculous situation of people doing this stuff. But we'll be right back with Alyssa Welly, who is joining us today to talk about AI cheating pens that can't cheat and are barely pens. Let me tell you about our next sponsor. It's Aura, bringing you this episode of Tech News Weekly. Now I'm going to grab my little Aura frame here because it's awesome and I want to show it to you all. So this is an E ink E paper. It's called the Ink, Aura's first ever cordless color E paper frame. Featuring a sleek 0.6 inch profile and a softly lit 13.3 inch display. Ink feels like a print, functions like a digital frame, and perhaps most importantly, what I'm moving it around lives completely untethered by cords, with a rechargeable battery that lasts up to three months on a single charge, unlimited storage, and the ability to invite others to add photos via the Aura Frames app. It's the cordless wall hanging frame that you've been waiting for and frankly, that I've been waiting for. I was so excited to load photos onto this thing and see it. And you know, I prepared myself, okay, this is an E Ink display. So I'm thinking the photos aren't going to be super great. And then I look at it and I go, wait, how are they able to do that? Well, they use some custom dithering algorithms that they created to improve upon this epaper technology. It's, it's super, super cool. With its cordless design, ultra thin profile, softly lit display, and paper textured matting ink. Looks like a classic frame, not a piece of tech. See for yourself@auraframes.com Inc. Support the show by mentioning us at checkout. That's auraframes.com inc. And we thank Aura for sponsoring this week's episode of Tech News Weekly. All right, back from the break. And we're talking about bad ways to cheat. So in using this pen, beyond the kind of accuracy problems.

D (56:20)

Right.

A (56:20)

What were some of the other issues that would make this pen kind of unusable in an actual testing environment? Is it mostly just the fact that it is so large?

D (56:32)

Yeah, I would say the size is. The size is funny. The interface is tricky to use. And also. Yeah. How do you get away with this? How do you not have your friends on either side, you know, lame, like, yeah, give me the answer. Or also just be like, you know, shocked. Shocked at this process. I think also it gets to the bigger question of, like, you know, how do you. How do you trust something? Right. So if you have a study guide that you've written down the answers for, you get an answer key, for instance, and you write them down and then you go. And you know that's the right answer. I think that that feels. It feels quaint and also more trustworthy compared to the idea of, like, accessing the Internet and trying to find the answer and trusting it. There's just something about it that overall is like, how do you. How do you trust if this is the right answer? Let alone the fact that it's probably very difficult to do? Because again, as I have here, this, like, kind of just remote. It's, like, pretty big. Like, I don't understand how you could really, like, bring this in to. To cheat all the videos, make it seem like it would be perfectly reasonable to pull it out of your pocket, swipe it across your page, get the answer, no big deal. No one's going to notice. I was going to say anything. That's definitely how it makes it seem. I found it to be more difficult than that.

A (57:52)

Now, it is a pen. Does it actually write.

D (57:54)

No. No.

A (57:55)

Okay, that. That.

D (57:57)

I know.

B (57:58)

Bummer.

A (57:59)

That's. Now, listen, I thought, you know, I'm going to ask him to say, okay, is it a. What kind of ballpoint is it? I'm not a huge pen nerd, but I'm like a. A sort of pen nerd.

B (58:09)

Sure.

A (58:10)

You can't.

D (58:10)

No, there's no. There's no pen. I know. Completely, completely misleading.

A (58:16)

That's foolish.

D (58:17)

Now it's really a scanner. A test scanner.

A (58:20)

Yeah, yeah. Which, I mean, then it's just. It's right there on the. On the box. It says what it. Like, it's not even. At that point, you're not being a secret agent. And then that's not fun. You did speak to some college students that talked about that kind of simpler cheating method. Yeah. Tell us more about kind of in speaking to them, did you get to talk to them about their thoughts on the. On using this scanner? That is definitely not a pen. And so I can't call it that anymore.

D (58:49)

I know, yeah, the. The AI, you know, pen. But in quotes, I would say that most were amused by it. I think the amusement was interesting. They were all, I would say, you know, just kind of curious to see how. How it would. How it would work. And I think what was most remarkable was just the fact that they were not surprised at its existence at all. I mean, they were interested. It was. It was like, oh, yeah, that makes sense. Like, that makes sense that that's available. That makes sense that there's, you know, this type of gadget that seems to claim these different things. Like, I think, like, that the lack of surprise was surprising to me because I was. I was surprised. I was amused. I hadn't thought of this before. I, I certainly was intrigued by the app ads, by the, by the YouTube videos. And I was like, okay, I want to try this out. And no one else seemed to share that surprise. And I thought that that was like, okay, well, that's. This. This is how normal it is to think about, you know, how AI is. It's just so deeply involved in a student's experience at school. It's just. It's everywhere. Right. And so a pen to help, you know, that claims to overcome the limitations of a paper test by being, you know, you're separated from the Internet. Well, here's a. Here's a pen to help reconnect you to the Internet, to give you the answers. That was like, yeah, totally normal.

A (60:14)

I have to imagine that the professors are probably also feeling that same way. Oh, yeah. I mean, what would you expect other than this being the case? Well, I guess we will have to see how long it takes for AI Scanning pension to make their way to the banned list of items allowed during tests.

D (60:33)

Right, exactly, exactly. No electronics, including phones. Scanning pens.

A (60:39)

Yeah, it's on the list. What will they think of next? I want to thank you so much for taking the time to join us today to talk about this wild and silly topic. I hope we get to see more. Hey, Verge, we got to see more of these TikTok reviews. Okay. We'd love to see more of that. So thank you so much for your time. If people would like to follow along with the work you're doing, where are the places they should go to do so?

D (61:03)

Come to the Verge. We've got an awesome webpage. You can follow me there and then. Yeah, lots of different socials that you can find on our page as well.

A (61:11)

Sounds good. Thank you so much.

D (61:13)

Thank you for having me. It's a pleasure.

A (61:15)

All righty folks, that brings us to the end of this episode of Tech News Weekly. Of course, the show publishes every Thursday. Twitter TV TNW is where you go to subscribe to the show in audio and video format. You can also share that link or those links with your friends and family. And of course I should remind you about Club Twit. Twit TV Club Twit. That is where you can go to become a member of the club. $10 a month gets you access to so much great stuff. All of our shows ad free. Just the content, none of the ads. You also gain access to our special feeds. We have feeds for our behind the scenes, before the show, after the show. We also have feeds for our live coverage of tech news events with commentary from different hosts and access to our club shows like my crafting corner. I'm currently working on painting some D and D miniatures as well as Stacy's book club. We've got camera time. We've got so many great different club programs that are all available to you in the club. Again, TWiT TV club TWiT is where you go to check out all of that and be sure to hop on board. We've got discounts going right now as well as that two week free trial. So check it out if you'd like to follow me online. I'm ichasargent on many a social media network or you can go to Chihuahua Coffee. That's C H I H U A H U a Coffee where I've got links to the places I'm most active online. And of course you can check out my other shows including Hands on Tech which publishes every Sunday, Hands on Apple and iOS today as well. I. I'm trying to think ahead as we are getting toward the end of the month. Next week no show as it will be Thanksgiving. But the following week Abrar Al Heati will be joined by, I guess I should say Jake Ward will be joined by Abrar Al Heiti on the fourth. So you will have that and yeah, that's that for that, that. Goodbye. See you. See you in a couple of weeks. Or in a few weeks I guess.

C (63:43)

Ever feel like the news is just arguments or nonstop tragedy? Plus, who has the time to keep up with it all? Well, I have a solution for you. Welcome, welcome to the newsworthy in just 10 minutes a day. The Newsworthy keeps you in the know about everything from politics to tech to entertainment, using its signature fast, fair, fun style. No wonder Harper's Bazaar, Fast Company, People magazine and more have called the Newsworthy one of the best podcasts to stay news savvy. Listen now by searching thenewsworthy in your podcast app or go to theneworthy.com boxes.

A (64:13)

Were all filled with gifts, big and small, but sharing pure love is the.

B (64:19)

Greatest gift of all.

A (64:21)

Stay cozy, my people, and have a Boss year. Get into the holiday spirit with Boss and our ultimate gifting edit. Visit your nearest store or explore our curated selection online@boss.com.