Deep Dive
Shownotes Transcript
This Mother's Day, show the moms in your life just how much they mean to you with a stunning bouquet from 1-800-Flowers.com. For almost 50 years, 1-800-Flowers has set the standard for high-quality bouquets.
Right now, order early from 1-800-Flowers and save up to 40% on gorgeous bouquets and one-of-a-kind arrangements guaranteed to make her day. Save up to 40% today at 1-800-Flowers.com slash ACAST. That's 1-800-Flowers.com slash ACAST. The official florist of Mother's Day. BetterHelp Online Therapy bought this 30-second ad to remind you right now, wherever you are,
to unclench your jaw. Relax your shoulders. Take a deep breath in and out. Feels better, right? That's 15 seconds of self-care. Imagine what you could do with more. Visit betterhelp.com slash random podcast for 10% off your first month of therapy. No pressure, just help. But for now, just relax.
What does the term digital workplace mean to you? In today's world of AI, robotics, and AR, it's much more than an office full of people on computers. It can be anywhere from a factory floor to the top of a crane to the cockpit of an F1 car. Wherever your digital workplace may be, TeamViewer's mission is to make work, work better.
Thank you.
So discover how TeamViewer can make work, work better, wherever it happens across your business. Learn more at teamviewer.com slash work better.
Picture this. You're in the garage, hands covered in grease, just finished up tuning your engine with a part you found on eBay. And you realize, you know what? I could also use new brakes. So where do you go next? Back to eBay. And you've got eBay guaranteed fit. You order a part, and if it doesn't fit, send it back. Simple as that.
So when you dive into your next car project, start with eBay. All the parts you need at prices you'll love. Guaranteed to fit every time. eBay. Things people love. This is the Daily Tech News for April 30th, 2025. We tell you what you need to know, follow up on the context of those stories, and just try to help each other understand.
Today, Dr. Nikki tells us how close we are to turning brainwaves into speech. And do we need age verification on the internet? Can we even do it well? I'm Tom Merritt. I'm Jen Cutter. Let's start with what you need to know with The Big Story. The Big Story
Verifying a user's age is heating up as a battleground. More organizations and countries have decided we need to protect the children from all kinds of things. Up until now, most of the attention has been on laws to age-gate porn sites. But there's also social media and dating sites and all kinds of other things that people say, you know what, we need to make sure you're whatever age before you get to use it.
In fact, there's U.S. legislation being considered to require age verification more broadly, and that's gaining support. So Meta, Spotify and Match, the dating service, have launched the Coalition for a Competitive Mobile Experience, a lobbying effort specifically meant to convince the U.S. government that operating system platforms, app store operators should verify age at the app store level.
In fact, the state of Utah already enacted a law in March that requires app stores to verify user age or parental consent in order for kids to use certain apps. A federal law is being drawn up along the same lines.
At the same time, Ars Technica's Ashley Belanger has an extensive write-up on the debate over age verification tools. Belanger notes that the last month a study was found that in areas of the U.S. with age verification for adult sites, 48% of users switched to searching for non-compliant sites and 34% switched to searching for VPNs to get around the restrictions. Belanger also wrote an extensive history of age verification and how users have gotten around it.
The U.S. Supreme Court will rule this summer on the constitutionality of age verification rules in Texas. Yeah, so there's lots of attempts about this. We talked previously about Discord trying out face scans in Australia and the UK. Meta and Google are both testing algorithmic approaches to try to identify patterns that indicate you might be a certain age or not. They're pretty good at telling adult from child. They're not so great at telling like a 17-year-old from an 18-year-old.
But Jen, the Ars Technica article was really good at pointing out that every time there is an age verification scheme, people, especially really smart teens, figure out how to get around it. Whether there used to be one where you used your credit card to verify age. There's a lot of good info on that in the Ars Technica article. And kids would just borrow their credit card or borrow without permission their parents' credit card to get around it. Even these algorithmic things
you know they're not perfectly accurate and there's probably ways kids can figure out how to get around those two it is tricky in my own uh friends discord uh we were really happy when the tests were not in canada because we decided like all right well when that hits us we're going back to irc yeah we're not dealing with this and then
in gaming for my friends with kids who are old enough. What was it? Counter-Strike had it where like you had to have a cell phone number or a credit card number. And, you know, a lot of these 18, 19 year olds don't necessarily have that
these days, because some of them just actually choose to be offline and I respect it. So their dad had to go figure out like, okay, how am I going to get all these numbers? How am I going to use the same credit card for multiple accounts without doing it as a family account, which then locks out the other people? There's a lot of hiccups and
I am the person who, when I was in my early 20s, did not look like I was in my early 20s. Thankfully, it was nice. But it was also really frustrating when I was, say, going to the LCBO or other places. And they're like, oh, is your ID real? It's Canadian. And I'm like, yeah, well, you know, Canadian ID is real. Yeah, those are real. And yes, I am of age to buy this beer. Thank you very much. It's like, I can rent a car, gosh darn it.
Well, and that shows one of the difficulties on the internet, too. It's hard enough to tell how old someone is by looking, as you just illustrated. It's almost impossible to tell how old someone is by looking on the internet, which is why they're trying to figure out the analog for that, whether it's a face scan to be like, okay, we're going to have the computer do what the LVCO clerk was trying to do, look at your face and guess how old you are. Algorithmic approaches are sort of like
analyzing behavior, they're probably more accurate than just a human looking at someone, arguably, but they're not perfect. I'm not trying to argue that there shouldn't be age verification, that there isn't a reason, but I feel like we're investing a lot of effort in the no one's responsible end when we need to also invest in
in the how do we get parents and kids to have enough education to understand what is working and what is not and how to sort of prevent the kids from going to those places in the first place. And I know the objection is like, well, kids are going to do it anyway. But that will work for a lot of kids to educate them like, hey, beware of this stuff. And then you deal with the age verification algorithms and
all the privacy implications those have and try to figure out that end of it as well. And I'm not trying to pretend I have the solution to that end either. Yeah. Like this definitely stops, say the 11 year old from checking out things, but,
I am in the age group where I figured out how to beat net nanny, how to beat all this stuff. Kids in schools with these gates, I know how to beat them. When you start having faces, okay, how does it work? AI generated images seem to work pretty well for a lot of these things. Yeah. Let's assume that it gets better and it can tell a generated image from a real image. Yeah.
Eventually, they'll just get a friend to sit in the seat during the face scan. Exactly. You can borrow an adult. Because then you run into the – There is no end of ways around these sorts of things. That doesn't mean they shouldn't be used, but we need to understand that. That's what the Ars Technica article was arguing too is we need to understand that when we're crafting these things and not pretend like tech can just solve it perfectly. And from the privacy standpoint, okay, like let's say I use my face for my –
That's great. And then let's say one of my cousins who also looks really young also needs my face. Does that become a problem? Because how long are these sites saving my face? Is this all linked to me forever? Is this going to link me across the internet for every site that I visit? Honestly, if we solved online ID problems,
Right. The idea that like I should have my own information under my control and that became pervasive and essential and you had control. So there was not a privacy concern because you could decide who knew what about you. Suddenly it would become a lot harder to pretend you're not the age you are because you'd want accurate age for other reasons.
Kids are going to want an ID that says they're under 18 to get a kid discount or to get into a game that's age-gated the other way where adults aren't allowed to play it or communities where adults that aren't parents aren't supposed to be in there. And if we had a really good identification system that was privacy protective and under your control, suddenly a lot of these problems become easier to solve because they're just sorted out by behavior.
Well, there was a lot of games that I could not play as a teenager, like South Korean RPGs, because you needed a specific country ID to even sign up. So I'm thinking like, okay, like, what if we do this by country? Because I would legitimately feel a little better if I knew that my data was being kept local. Yeah.
Yeah. And you don't even need to keep the data to determine an age. You can just say keep the yes or no. They pass the test. You don't have to keep the facial scans. But if all our identification is under our control, then again, it's a whole better situation for a lot of reasons, not just this one.
DTNS is made possible by you, the listener. Thanks to Adam L., Tony Glass, Philip Less, and our new patron, Alex. Yay, Alex! Thanks, Alex. Picture this. You're in the garage, hands covered in grease, just finished up tuning your engine with a part you found on eBay. And you realize, you know what?
I could also use new brakes. So where do you go next? Back to eBay. You can find anything there. It's unreal. Wipers, headlights, even cold air intakes. It's all there. And you've got eBay guaranteed fit. You order a part and if it doesn't fit, send it back. Simple as that. Look, DIY fixes can be major. Does
Doesn't matter if it's just maintenance or a major mod. You got it, especially when things are guaranteed to fit. So when you dive into your next car project, start with eBay. All the parts you need at prices you'll love. Guaranteed to fit every time. eBay. Things people love.
Every idea starts with a problem. Warby Parker's was simple. Glasses are too expensive. So they set out to change that. By designing glasses in-house and selling directly to customers, they're able to offer prescription eyewear that's expertly crafted and unexpectedly affordable. Warby Parker glasses are made from premium materials like impact-resistant polycarbonate and custom acetate. And they start at just $95, including prescription lenses. Get glasses made from the good stuff.
Stop by a Warby Parker store near you. Hey, it's Mark Maron from WTF here to let you know that this podcast is brought to you by Progressive Insurance. And I'm sure the reason you're listening to this podcast right now is because you chose it.
Well, choose Progressive's Name Your Price tool, and you could find insurance options that fit your budget so you can pick the best one for your situation. Who doesn't like choice? Try it at Progressive.com. And now, some legal info. Progressive Casualty Insurance Company and Affiliates. Price and coverage match limited by state law. Not available in all states.
Work management platforms. Endless onboarding, IT bottlenecks, admin requests. But what if things were different? Monday.com is different.
No lengthy onboarding. Beautiful reports in minutes. Custom workflows you can build on your own. Easy to use, prompt-free AI. Huh. Turns out you can love a work management platform. Monday.com, the first work platform you'll love to use.
Worried about what ingredients are hiding in your groceries? Let us take the guesswork out. We're Thrive Market, the online grocery store with the highest quality standards in the industry. We restrict 1,000 plus ingredients so you can trust that you'll only find the best high quality organic and sustainable brands all free of the junk. With savings up to 30% off and fast carbon neutral shipping, you get top trusted groceries at your door and you can stop worrying about what your kids get their hands on.
Start shopping at thrivemarket.com slash podcast for 30% off your first order and a free gift. There's more we need to know today. Let's get to the briefs. Here's a roundup of some AI service news. Alphabet CEO Satya Nadella told the court Wednesday that he believes Gemini will be added as an option on iOS later this year alongside ChatGPT.
OpenAI has rolled back changes to GPT-4.0 that, in CEO Sam Altman's word, had made the chatbot too sycophant-y. Basically, it was agreeing with people on whatever they said. You know, that's a perfectly valid way of saying that, Jen. Good job. Don't regret any of your decisions. Yeah.
DeepSeek updated its Prover model, which can solve math proofs and theorems. Prove V2 has 671 billion parameters and uses a mixture of experts' architecture. Xiaomi launched its own open-source reasoning model called Mimo. Xiaomi claims it can outperform OpenAI's O1 Mini and Alibaba's Qen models. Alibaba just updated its flagship model on Tuesday.
And Wikipedia said it will not use AI to replace moderators, but will use AI to help them automate tasks, find information and onboard new volunteers. Man, the Wikipedia Foundation does a good job on this stuff. It was a very clear post that avoided a lot of the usual misunderstandings of things.
Go read that post. We'll have a link in the show notes. It's a good way of saying, hey, we don't want to replace humans. We want to make this job easier for humans because it's a hard job. And actually, if we make it easier, more humans might want to volunteer to moderate Wikipedia, which would be a good thing for everybody. Yeah, I hope it's a good framework that I hope other companies pull from because it does make humans feel a little better. Yeah, yeah, exactly.
Here's a roundup of some of the top news that came out of the kickoff to Meta's Llamacon conference. That's the one about their AI tools. Microsoft's Satya Nadella showed up and joined Meta CEO Mark Zuckerberg on stage. During their conversation, Zucker asked him how much of the code inside Microsoft was generated by
And Nadella said he estimated 20 to 30 percent of code inside Microsoft repositories was written by software. Nadella said generated code was a mixed bag, though. He said they've made more progress in Python, less in C++. And then asked Zuckerberg the same question. And Zuckerberg was like, I don't know. I don't know how much code in meta was generated. We do know last week in its earnings call, Google said that more than 30 percent of its company code was generated.
Meta also announced private processing for WhatsApp that can summarize unread messages while keeping the messages private. Sounds like some kind of differential privacy is being used, so it's possible. I didn't see the details of how they're doing it. Meta will let independent security researchers audit this program and will offer bug bounties as well, so that's really good news. No word on when it would actually launch, though. And finally, Meta also changed the privacy policy on its Ray-Ban Meta smart glasses.
Meta AI with camera will now always analyze what your camera can see unless you turn off the virtual assistant altogether. It doesn't mean it'll store it, but it'll be constantly taking a look so that it can answer your questions faster. And you can no longer choose not to have your voice recording stored in the cloud. They will be stored up to a year in most cases.
That one feels like a bit of a bummer. I would want control for my extremely expensive device. Yeah, Amazon did this too. And I'm curious why both these companies felt they could get away with this. It feels like they know people are distracted by so many other things that they won't get as much of a backlash over doing this. And obviously it makes it easier for them to train up their models. Yeah, because people who bought it already no longer have the choice to not use it without ruining their investment. Yeah.
EA announced targeted team adjustments on Tuesday, which Bloomberg estimates means about 300 to 400 eliminated jobs, some of which were at Respawn. It also reportedly canceled an unannounced Titanfall game codenamed R7. EA's Respawn only said it canceled two early-stage incubation projects. Respawn says it will focus on Apex Legends and Star Wars Jedi franchises.
AppFigures estimates that the Android app marketplace now hosts 1.8 million apps worldwide. That's down from 3.4 million at the start of 2024, a 47% drop. If you're thinking like, well, maybe people just don't make as much apps anymore. During the same period, Apple's App Store went from 1.6 million to 1.64 million. So it's not a general decline in apps. It appears to be related to Google raising the minimum wage.
quality requirements for apps, which they did in July. They used to just say the app has to run basically without crashing. The new rules banned apps with limited functionality and content. So if somebody just uploaded a PDF as an app, which has actually happened, that will no longer be allowed in the App Store. Google also says it has improved its threat detection that has helped cut down on apps. And this February, the EU announced
requiring names and addresses in app listings, so that might have had an effect as well because people didn't want to have to tell you who they were. AppFigures does note that new releases on the App Store are up 7.1% as of April. I do like how App Stores are getting a little more involved and also before the governments necessarily make them get involved. Like Nintendo had to clean up their store after
after it was revealed just how much AI slop was on there. And yeah, like a lot of, it was a big problem with Android apps where people were, you know, domain jacking and directing people to malicious apps. So the more clean up there, the better. Yeah. And especially because it's going to make their search work better if you don't have to wade through a bunch of PDFs that are miscarrying as apps, right? It's just better for everybody.
Nintendo updated its Switch software in preparation for the Switch 2 coming June 5th. Existing Switch owners can take advantage of some features right away, like the Virtual Game Cards function that lets you lend digital games to anyone else on the same Nintendo family account that can connect to your Switch locally. GameShare lets others play a multiplayer game you own over local wireless.
And you can now use the system transfer to Nintendo Switch 2 to get your data already in the cloud to download to the new Switch when and if you get it. Now, uploading your data resets your Switch to factory settings, so please do not do this until you are ready. Yeah, I think you can restore it too, but, you know, kind of a pain. So wait, wait until you're done with the Switch.
before you do that. But the virtual game cards function is really cool. It replicates sharing a game because you have to be in the same room with somebody the way you would if you handed them a game card. But 3DS used to do this too. Like, hey, do you all want to play this multiplayer game? Just definitely be very close in the same room. Yeah.
Here's an example of why you need to keep your devices updated and patched. Researchers at Oligo Securities found 23 vulnerabilities in Apple's AirPlay protocol and AirPlay SDK. They reported the vulnerabilities to Apple. So this is a good news story. I always try to emphasize when somebody finds 23 vulnerabilities, it's tempting to be like, oh, no. It's like, yeah, there's always vulnerabilities. The question is, did you find them? Good news. They found them. They reported them to Apple. Apple patched them in March on March 31st.
The vulnerabilities can be exploited by attackers if they are on the same network as you. So this is something they have to be pretty close. However, you don't want that to happen. So you need to patch because some of the vulnerabilities included no click, meaning you wouldn't get the AirPlay confirmation dialog box that is meant to let you know someone is connecting via AirPlay and you might not know you've been compromised.
I finally got my new phone. And then as I left the store, I had my old phone and my new phone in the pocket. And then my notifications were going crazy. So like, you're now connected. You can do this. And I'm like, I had no idea that was a thing and immediately disabled all my airplane. Yeah. That's the other way to mitigate against this, right? Is you can restrict airplane so that it just doesn't connect to anything. And that works too. Waymo announced a partnership with Toyota to explore bringing self-driving tech to personally owned vehicles.
It could also mean Waymo might start using Toyota vehicles in the Waymo ride-hailing fleet. Yeah, so if you heard the story earlier that Waymo might start exploring personally-owned vehicles, well, this is the next shoe-dropping. I don't know how many feet this story has, but this is the next one. They're talking about it with Toyota. So we're years down the road from this being something you could buy, but they're looking at how to be able to do it.
Online graphic design platform FreePik, that's pick with a K, P-I-K, launched an image generation model on Tuesday that it says was trained on commercially licensed safe-for-work images. So if you're worried about the provenance of training data, this is good for you. F-Lite has a standard version that they say is pretty predictable and a texture version that's better at texture and creative compositions, but they say it is a little more error-prone.
F-Lite is available for free. If you have a GPU with 24 gigabytes of VRAM to run it on, you can run it yourself. It's available under a CreativeML OpenRailM license. You can read that on their GitHub. FreePIC joins Adobe, Bria, Getty Images, Moon Valley, Shutterstock, and a few others in offering image generation based on licensed data. So this is becoming more common.
Finally, in news related to how much tech you'll see on shelves in the U.S., the executive director of the Port of Los Angeles says the port expects imports to drop next week to two-thirds of its normal level at this time of year. He expects retailers now have about five to seven weeks of inventory left before you see your choices reduced on shelves. Yeah, this isn't exactly a tech story, but we included it because I think a lot of people have been saying like, oh, I better hurry. Yeah.
And for a long time, I've been saying, well, we don't know what's going to happen and when it'll have an effect. And this is the first person in a position to know because it's the person that runs one of the major ports in North America saying like, all right, you've probably got, you know, if something doesn't change in the next five to seven weeks, you're going to have an effect.
you probably start seeing effects on availability at that time. Now, this is for all goods, so it may affect tech differently because there are more exemptions for certain kinds of tech like smartphones and stuff like that. But, you know, cables and accessories and things like that, you might see things change. And it might even spill over to Canada, Jen, because even though you can do imports, a lot of stuff comes through the port of Los Angeles. So there might be some halo effects to Canada and Mexico as well.
Yeah, because a lot of ships in BC then drive stuff to Alaska and down south and the other way. So, yeah, so Canadians are keeping an eye on this. But there's also the companies like, for example, Nintendo, who front-loaded as many shipments as possible just to try to stave this off. Yeah, yeah. And that's what he's talking about is pretty much all companies have done that. And now they've got –
And, you know, they've got a certain amount of inventory depending on who they are before they run out. Those are the essentials for today. Let's dive a little deeper. The ability to turn your brainwaves into speech is getting closer and closer. How close? Tom talked to Dr. Nikki to find out. There are lots of buzzy products and lots of less buzzy products with very good applications that
couple together generative models and voice synthesis to create the ability for people who can't speak to speak. Multiple universities and companies are working together to make brain to synthetic voice speech as fluent as any other chat between two people. So we thought we got to get Dr. Nikki on to talk to us about this. Dr. Nikki, thank you for joining us.
Beep boop, beep boop. Yes, it's me. Yours is on the fritz, I see. But how does this work? Let's give the basics to people who haven't heard about this before. How does this translate brainwaves to speech?
Sure. I'll say it in more of a generalized way. So first, you can measure electrical activity from your brain. That's how an EEG cap works, if you've ever seen those in a medical show or something. Yeah, you put the little things on the head. Yeah, electrodes, and then your neurons make electricity, basically, and that can pick it up.
And then you can process this into neural activity. This we can do very easily already since the 80s. Then comes in the AI stuff. So after that, neural networks and machine learning will work together to convert these activity signals into sounds based on training data sets. And the really interesting part here, and I think the innovation, is that they can do this every 80 milliseconds. Wow.
And after that, large language models will generate the top 100 most probable word sequences that are coming from this electrical activity. And it'll send the most likely output to the person's audio device, resulting in a word, hopefully. And this is called speech neuroprothesis. This makes so much sense. So there's a model that is trained to figure out what sound would come out
If this brainwave happened in most situations, and then there's a large language model going, well, those sounds together sounds like they're trying to say the word cow or whatever. Yeah. Cause we can do this without AI. It's just really slow. Yeah. Yeah. So the, so the, the, the models here speed things up. Um, and we are doing this, like this is working for some people, right? Yeah.
Yeah, so we are successfully doing this in humans. In fact, last month, a paper was published in Nature Neuroscience on specifically this. So a clinical trial patient named Anne had severe paralysis. And in the use case here with this prosthesis, she was able to drive a continuously streaming naturalistic speech synthesizer. Say that 10 times with that normal voice. Yeah.
And it was trained actually on her voice that was from recordings from before she was paralyzed by a stroke. So we have some audio from that. And let's have a listen. So today you will live.
Okay, so that's a little halting, more halting than I would have thought, but you can understand it. Yeah, it says, so did you do it? Where did you get this? Okay, so it's clearly not being faked. And I feel like there's another step that could be done to like clean it up even more, but that's impressive.
Yeah. And, and the researchers who did this work said that we're, instead of being decades away of cleaning that up, we're maybe a few years away, which is really promising. Um, and, and like we mentioned, one of the biggest problems with the current neuroprothesis, even before this one is latency. If you've ever heard audio from Stephen Hawking, you're familiar with this. Um, but with this new method, uh,
And this patient did not need to wait to finish her sentence to start a new one. She can have kind of slower but still conversations, which was really not an option before.
And it reduced the lag between her brain signals and the audio output from eight seconds down to one. So it's a pretty big improvement. But as you can hear in that clip, she's still about at a third of a rate of normal speed of conversation. Yeah. And the first thing that occurred to me when I looked at it was like, well, you could just take the text of what she said and run it through any text to speech and make it sound better. But that's going to introduce lag, right? And you have to wait for the sentence to finish to make sure that it's...
The text was visible in this video because that was like showing her what to say. But the goal would be that she's just using her own words. So you wouldn't have text. So what makes it so difficult to synthesize it, quote unquote, on the fly?
Yeah. So one of the major obstacles right now in the development of brain-to-voice technology is the amount of time that it actually takes participants to learn how to use the system. Because it's not just straight up, we plug something in your brain and then you just talk in your brain and it comes out. It's actually mapping onto the area in your brain that's responsible for movement of the muscles of the face that are related to speaking. So you have to kind of learn
kind of brain impulses to make that would maybe move your lips a certain way that would make a certain
a certain sound and that sounds complicated because i think it would be in real life to have to do this yeah um so this is happening if you're out of practice like these are people yeah you've never done this before they're not doing this every day yeah exactly um potentially you could generate speech by going into other brain areas we know that there's other areas in the brain that are responsible for different things related to speech not just the motor properties but um
it's complicated but lots of other things that are related to speech but some scientists note that this might lead to you saying your inner thoughts out loud which i thought that's what everybody's worried about with this kind of thing so yeah you want to be careful like that would be an advance to be able to read your inner thoughts out loud but that would bring up a whole other that's some evil villain stuff right yeah the whole other list of ethical such uh questions uh once you start doing that exactly uh
It's good to remind people that that is not what is going on with this. How accurate is this right now? You know, I was pretty surprised. The neuroprosthetics lab at the University of California, Davis, claimed that they were able to decode speech at a 98% accuracy rate. Considering that it's an LLM that's putting the output out, you're going to have some errors. But for someone who may not have had access to something like this before, that's pretty awesome.
However, these prostheses, I need one, they can't capture important speech qualities like tone and timing. So they're going to keep working on that. But overall, I think this is for once a really cool use of AI. Yeah.
Yeah. And I think it has all the hallmarks of something that could become really common really fast. Yeah. So this is the kind of thing that models are very good at is being trained, machine learning models in particular, being trained to recognize this means that, this means that kind of sound.
Large language models, very good at predicting like, well, if these kinds of sounds are coming together, it's probably this word, right? So we've got technology that can do this. And if the barrier to entry is training people, then
That feels like something that the more they do it, the better they're going to get at understanding how to train people and get them proficient at it faster. And then it's a matter of just smoothing it out, like you said, and creating tone and timing. I don't know about you. I wouldn't be shocked if...
This just becomes natural when people lose the ability to speak or are born without the ability to speak that the doctor just says, well, let's sign you up for the training and get you on your way to talking this way. It might become something that we're just used to seeing in the real world.
We should put this in our next predictions episode for 2026. But I think this is one of those things that is like such a perfect use case for AI and it's for good. And it, it, even if it works 40% of the time, that's still better than not being able to speak at all. If that's what you're going for, maybe some people don't want to use it, but if you do want to be able to speak and you're able to speak at a one third of the speed, that's a huge improvement on zero. So that's,
So maybe I'm sure there's ways that this can go wrong, but it almost feels like you can only win with this kind of technology and the rate at which these kinds of things progress with, with how fast models are able to learn. Looks like it's only looking up from here. So good news. Yes. Good news again, Dr. Nikki, thank you so much for bringing it to us and helping us understand it. If folks want to follow what you do, where should they go?
Absolutely. So I will post this over on my website, NicoleAckermans.com. And you can also find me at my name on blue sky and I'll see you there. Thanks, Nikki. Of course.
We listen to our supporters because we make the show for you. I mean, we love doing it, but if you weren't listening, it really might not get done. So if you have a few minutes, please fill out our annual DTNS survey. Let us know what you think about the DTNS family of shows, what's working, what's not, what thoughts you have. It really helps us improve the show, and it's fun to do. There's a lot of questions about Pi, so go do it. DailyTechNewsShow.com slash survey. ♪
Hey, this is Farnoosh Tarabi from the So Money Podcast. Running a business means wearing a lot of hats, but ordering supplies shouldn't be one of the ones you don't like. Walmart Business helps organizations like yours save time, money, and the headache of managing purchases. From office essentials to bulk break room snacks, it's all in one place, online, in-store, or right in their app. Sign up for free at business.walmart.com and get back to what really matters, running your business.
And use code ACAST at checkout.
We end every episode of DTNS with some shared wisdom. Today, David is helping us understand. Yeah. So in the episode yesterday, we chatted with Bodie about some coming EVs. And David said, I always enjoy listening to him when he's on the show. I especially enjoyed this conversation because one of the cars he talked about was the MG4 and I just bought one.
I think maybe he was confusing the MG5S, which is an SUV with the hatchback, which is the SE. But other than that, he was spot on. People, and that includes me, love these cars. I watched a lot of reviews before buying my car and noticed a split. Professional reviewers who routinely drive cars like Rolls Royces and Jaguars highlighted the car's shortfall. Zero to 60 is slower than, say, a Tesla Plaid, whereas normies loved Jaguars.
Everything about the MG4. Well, almost everything. The way it tries to comply with the EU's driver assist laws makes it nag the driver a bit too often. Thank you for your coverage of the EV sector. A few years ago, DTNS hosted an EV roundtable with speakers like Alison Sheridan, and that was a large reason why I decided to make the switch. Kind regards, David.
Very nice. Yeah, no, I love hearing from people who actually have and actually use this tech on a daily basis because it's a huge investment. So you know they're getting good information when they feel confident to drop that kind of investment. Yeah, and trust me, if someone doesn't like their car, I have found they're generally telling you about it. So it's good to hear that David likes the MG4.
What are you thinking about? Got some insight into a story. We, again, really love hearing with you. So please share it with us. Feedback at DailyTechNewsShow.com. Big thanks to Dr. Nikki and to David for contributing to today's show. Thank you for being along for Daily Tech News Show. It is made possible by our patrons at Patreon.com slash DTNS. Don't forget, you can get it ad free either through Patreon or through our sub stack. Head to DailyTechNewsShow.com slash subscribe for that. Talk to you tomorrow.
The DTNS family of podcasts. Helping each other understand. Diamond Club hopes you have enjoyed this program.
Hi friends, Nikayla from SideHustle Pro here. Whether you're running a non-profit, a school, or a small business, Walmart Business is here to support your mission. They make it easy to order what you need from tech and cleaning supplies to everyday essentials, all at low prices and with helpful tools like spend tracking and tax-exempt purchasing for eligible organizations. Because when your operations are smooth, your impact can be bigger. Visit business.walmart.com to get started.
What makes a great pair of glasses? At Warby Parker, it's all the invisible extras without the extra cost. Their designer quality frames start at $95, including prescription lenses, plus scratch-resistant, smudge-resistant, and anti-reflective coatings, and UV protection, and free adjustments for life.
To find your next pair of glasses, sunglasses, or contact lenses, or to find the Warby Parker store nearest you, head over to warbyparker.com. That's warbyparker.com.