This Week in Tech 1035 Transcript

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

00:00 - Leo Laporte (Host)
It's time for Twit this Week in Tech Great panel for you. My car guy, sam Abulsamit, is here, my internet lawyer, kathy Gellis, is here and our eighth most popular guest, larry Magid, is here. We're going to talk about the passing of a computing legend with some great stories. Tiktok is still alive and, good news, freefile goes open source. All of that and more coming up next on TWIT Podcasts you love. From people you trust.

00:33 - Cathy Gellis (Guest)
This is TWIT.

00:40 - Leo Laporte (Host)
This is TWIT this Week in Tech, episode 1035, recorded Sunday, june 8th 2025. The droids are in the escape pod. It's time for TWIT this Week in Tech, the show where we cover the week's tech news. We've got an all-star panel assembled for you. As usual, larry Magid is here from connect safelyorg. Good to see you, larry, welcome. Good to be here. Thank you. When was the first time you were on Twit? You've been on for a long time.

01:11 - Larry Magid (Guest)
A long time ago and you know I did something kind of fun. I took your spreadsheet which shows all of your guests for the last umpteen years, yeah, and I ran it through ChatGPT and I'm the eighth most popular person on your list. It listed, you know. I asked it to rank me because I'm very vain, Chat.

01:26 - Leo Laporte (Host)
GPT will tell you how popular you are.

01:29 - Larry Magid (Guest)
Well, it'll tell me how many times I've been on relative to other guests. So, and it actually went back and looked at the entire oh yeah, well, that makes sense. Yeah, we'd like to have you on. Yeah, well, it was fun to be on. I love being on this show. In fact, I have to tell you, when I travel, I am more likely to get people to say I heard you on Twit than I am to say I heard you on CBS News. Well, that says something. A lot about.

01:56 - Cathy Gellis (Guest)
CBS News, I'm not sure which. I'm finding this a little bit confusing, Like I've heard that voice before, but not here. I haven't been on with you before.

02:02 - Leo Laporte (Host)
That'sathy gellis. She is, of course, a contributor at tech dirt and she is an officially an attorney at law cgcouncilcom and we love having kathy on all the time, but especially when the supreme court's been active, which has been a lot lately, because she is admitted to uh to uh. What do you call it? Lawyer in front of the supreme court.

02:22 - Cathy Gellis (Guest)
I'm technically a member of the Supreme Court bar. This means I get to do stuff at the Supreme Court like file briefs and, in theory, argue in front of them. But that's kind of hard to arrange.

02:32 - Leo Laporte (Host)
Yeah Well, you never know, if I ever have a case in the Supreme Court, you're hired. Thank you, ok, actually, you would definitely. I'm not even joking. You would definitely, I'm not even joking, you would definitely be hired. Also here, our car guy Sam Abulsamed. He's VP of research at Telemetry and, of course, hosts the wonderful Wheelbearings podcast and is probably the number sixth most popular host.

02:55 - Sam Abuelsamid (Guest)
I don't know about that, because I think the first time I was on Twit was probably about 10 years ago.

03:00 - Leo Laporte (Host)
Yeah, larry goes back to the Dvorak era. That's how long Larry's been on.

03:05 - Larry Magid (Guest)
I'm actually chat bt because it might remember the spreadsheet I uploaded most popular ask it again. Sam is number 39. I have the list here.

03:11 - Sam Abuelsamid (Guest)
I have it all okay, sam is 39, oh, okay if you counted all my appearances on the tech guy and other stuff, the tech guy you used to be on every month on the tech guy, so actually it was like every week was it every week, oh it was.

03:24 - Leo Laporte (Host)
It was every week. I was desperate to fill my sergeant number six according to chat gpt yeah, because uh martin sergeant like micah oh yeah micah?

03:35 - Larry Magid (Guest)
yeah, of course, because he's a staff came in hot team on top I should be number one, but you never know. Number two I'm jealous. A lot of vacation time. You know you love jeff more than you love me. What? No, I don't.

03:45 - Cathy Gellis (Guest)
No, I love jeff I'm telling jeff don't tell jeff, I said that uh it's the mean girls episode normally.

03:53 - Leo Laporte (Host)
I save, nor I don't want to. This is a lively, fun bunch and I want to sadden everybody but and normally I save the uh, the obituaries to the end of the show, but this one is too big a deal to not mention. Uh, bill atkinson, who was a absolute pioneer in our industry. Um, the 51st employee at apple. Uh, he created quickdraw, which is the apple graphics. Um, primitives did it for lisa. Then Apple created MacPaint and HyperCard, which was in many ways a precursor to the World Wide Web. Passed away June 5th, a Pancreatic Cancer which sadly also took Steve Jobs and Bill Atkinson's mentor I spent a lot of time and Jeff Raskin as well. Did I say you?

04:43 - Larry Magid (Guest)
mentioned Steve and then Jeff as well. Yeah, yeah, I said steve and then jeff as well. Yeah, yeah, yeah, jeff was kind of one of the the original creator of the macintosh, actually yeah, well, it was uh actually raskin who brought bill atkinson in.

04:54 - Leo Laporte (Host)
he had been a mentor of bill's. He's in the art department at the uc, san diego and uh and bill is a bill is an intro was an interesting fellow. I was very lucky to spend quite a few hours with him. I think one day we spent five hours talking and turned it into a couple of triangulation episodes, but he was a very creative guy. His original work was in brain science and, in fact, he was working on his PhD in brain science when Jeff Raskin said you've got to come to Apple. In fact, I'm going to let Bill complete this story. This is from one of our conversations back in 2018.

05:33 - Bill Atkinson (Interviewee)
Right, and he didn't take no for an answer. He sent me a round-trip airfare with a little note saying no strings attached, just visit for a weekend. And my dad lived in Los Gatos, not too far from apple, so I figured okay, I come down for a weekend.

05:48 - Leo Laporte (Host)
Apple's pretty small this time about 30 employees, yeah and steve jobs.

05:52 - Bill Atkinson (Interviewee)
I don't know what jeff said to him, but steve dedicated one entire day to recruiting me. He took me around to everybody that worked at apple. There were 30 people. They seemed like they were having fun, they were working on interesting things, uh, and their their heart was in the right place. They wanted to do something good in the world, but that wasn't enough. I had to finish. I wanted I really consciousness is pretty amazing frontier and I wanted to work on that.

06:23
But toward the end of the day, steve took me aside and he said when you read about some hot new technology that really was created two years prior, there's a lag between the time when something's created and the time when it's available to the public and if you want to affect the world, you have to get ahead of that lag and come to Apple, where we're inventing the future. You can help to affect the world. You have to get ahead of that lag and come to Apple, where we're inventing the future. You can help to invent the future and affect millions of people's lives. And then, just as I was leaving, he gave me this visual. He said think how fun it is to surf on the front edge of a wave. And how not fun to dog paddle on the tail end of the same wave.

07:03
And I kind of stuck my craw.

07:05 - Leo Laporte (Host)
He was a good salesman man, I tell you two weeks later I had quit my neuroscience program.

07:11 - Bill Atkinson (Interviewee)
Wow, never finished my degree. Wow, I had moved from uh up in seattle down to silicon valley, which when I left was santa clara valley, right it was prune trees everywhere yeah, and I was working at apple two weeks after.

07:26 - Leo Laporte (Host)
It's a brave thing how did you judge amazing stories from bill atkinson? He wrote hypercard after an acid trip. A lot of those stories are, uh well, on our page. If you search for bill atkinson, there's, I think, five shows that we did with bill. Another place to go is um andy herzfeld's folkloreorg, where many of the stories about the original macintosh team are.

07:53
In fact he tellsa great story about uh, bill and quick draw, which was, as I said, the graphics primitives very important to both the lisa and the max predecessor and the macintosh. And to this day, uh, brilliant. Uh, he wanted to draw ovals and circles very quickly, which is hard to do on the mac at that time because it was its processor, didn't have floating point math, it only did integers. Uh, and and circles need square roots. You know, pi r squared. Bill came up with a way to do the circle calculation with integers just addition and subtraction, not even multiplication or division. Uh, and very proudly came in to show the team because he worked at home, he programmed at home and uh and he showed them drawings, showed the lease of drawing circles, hertzfeld writes, but something was bothering Steve Jobs. Well, circles and ovals are good, but how about drawing rectangles with rounded corners? Can we do that now too? To which Bill replied no, there's no way to do that. In fact it'd be really hard to do. I don't think we really need it, hertzfeld says.

09:03
I think Bill was a little miffed that Steve wasn't raving over the fast ovals and still wanted more. Steve got more intense. Rectangles with rounded corners are everywhere. Just look at this room. Sure enough, there were lots of them the whiteboard, the desks, the tables. Then he pointed out the window. Look outside. There's even more, practically everywhere you look. He even persuaded Bill to take a walk around the block with him pointing at every rectangle with rounded corners he could find. When Steve and Bill passed a no parking sign and you know how they look with rounded corners it did the trick. Okay, I give up. Bill pleaded, I'll see if it's as hard as I thought. He went back home to work on it. The next afternoon he came back with a big smile on his face.

09:39
The demo was now drawing rectangles with beautifully rounded corners. Bl now drawing rectangles with beautifully rounded corners blisteringly fast, almost at the speed of plane rectangles. And if you think about it to this day, computer interfaces, if they were just square rectangles, would be ugly. They would be blocky, rounded rectangles, did transform the interface and Bill did so many things like that. I don't know why this one hit me really hard. Maybe because I felt a real kinship with Bill, but I think also because it is in some ways the end of an era. It's been more than 40 years since Bill did that seminal work that made the Macintosh what it is today, what computing is today. That's a long time. It feels like yesterday the first time I saw a Mac.

10:23 - Larry Magid (Guest)
It also reminds us we're getting older.

10:29 - Leo Laporte (Host)
Yeah, he was only five years older than me. When he's younger, than me.

10:32 - Cathy Gellis (Guest)
Yeah, but uh, what a loss window around a rectangle yeah, everywhere you see it around the window just to see my own uh, yeah, he, he became a photographer later in life.

10:44 - Leo Laporte (Host)
Beautiful, beautiful work, amazing work, worth going to that website. It's just, yeah, oh, go see it. Yeah, billackinsoncom. He had, he did rounded, uh, rounded, rectangles. He did thin sections of rocks and minerals and stuff I have. He gave me some and I'm, which I treasure um, a really uh, beautiful, sweet, brilliant man. John Gruber said, perhaps the best programmer of all time. Certainly we're very lucky to have had him.

11:13 - Larry Magid (Guest)
So I thought I'd start the show just to build. I knew him then. I have not seen him in probably in 30 years, but I remember him as just being a very charming energetic guy.

11:21 - Leo Laporte (Host)
You see how different he looked. He had long hair and a big, big, viva zapata mustache and bright blue eyes and apparently he was I didn't know him in in that day. I did, I didn't know andy a little bit, but he didn't know akinson, apparently. Yeah, very energetic, very passionate, yeah, um, he, uh, when we were interviewing him, his at the beginning his wife Sue was very ill and passing and he was caring for her and it was very hard. And then later, when I interviewed him some years later, after she passed, he said he did an ayahuasca ceremony and talked with her and she gave him permission to move on. He did in fact find another wife who he loved deeply and he was surrounded by his family when he passed um after a somewhat long illness. So, bill atkinson, rip and thank you for all the round wrecks we had and hypercard boy.

12:15 - Larry Magid (Guest)
That was a breakthrough. Hypercard amazing, amazing, I. I remember having it and I actually helped write a book about it. It was. It was oh, did you? Yeah, I've got it on my shelf here, but it was just an amazing the fact that you could do that. And you're absolutely right. I mean the World Wide Web. I mean I give a lot of credit to Sir Bernard Lee, but I really think that Atkinson gets the real credit for it.

12:39 - Leo Laporte (Host)
If, at the time, we'd had the Internet, it would have been the World Wide Web, but we didn't. And Bernard Lee obviously was inspired, uh, by what hypercard was. Uh, if you'd never used it it maybe it wouldn't come clear to you, but it was a. Really it was amazing. I remember I used it to plan my radio show in the mid 80s. Uh, my dad used it to do a history of dar and sailing in the Beagle and discovering evolution and so forth. So many HyperCard stacks, and this was before the internet.

13:12 - Larry Magid (Guest)
And also when the CD-ROM came around. I can't remember what they called it then, but CDs all had this kind of interactive clicking and moving from one spot to another. It wasn't called HyperCard, but it was very similar.

13:25 - Leo Laporte (Host)
Yeah, and moving from one spot to another. It wasn't called hypercar, but it was very similar. Yeah, oh yeah, I think this whole hyperlinking thing um really dates back.

13:31 - Larry Magid (Guest)
Actually I meant the dvd more than the cd, but the point of it right, right um, anyway, I've been, I've been sad for the last 24 hours.

13:40 - Leo Laporte (Host)
It was I got an email from uh john slanina, our studio engineer, with a bunch of pictures of bill at the studio, and I said oh. I said, oh, that's nice, I don't know why he sent me that. Then I saw it was. I got an email from uh john slanina, our studio engineer, with a bunch of pictures of bill at the studio and I said oh. I said oh, that's nice, I don't know why he sent me that. Then I saw the news and I thought, oh, I'm very sad, uh, very, very sad, uh, all right, now we can move on to some jolly things. Well, not really. Uh the uh, and I'm glad you here, kathy.

14:05
There's a lot of court news. Openai has been ordered to preserve all the chat GPT logs, including deleted chats. This is in the New York Times versus OpenAI lawsuit versus uh open ai lawsuit. Uh, frankly, I'm stunned. Um, open ai argued the order really, uh invaded the privacy. They warned the privacy of hundreds of millions of chat gpt users globally is at risk. And it's true because, uh, you know it does. It will by default remember your chats, but there's a button to delete them. Yes, um, apparently openai has everything that was intentionally deleted.

14:57 - Larry Magid (Guest)
I don't know, that's speculation with me because I use it for medical research and I do delete some data that I, that I give it, yeah and it, and I always assumed that it was gone.

15:09 - Cathy Gellis (Guest)
I think they start to risk issues with compliance with anybody's privacy laws if they say it is and it isn't, or if they're required to delete it and it's not. I think this is an order that there's clearly tension here. I don't know the details enough to know exactly where the specific legal pain points are, but conceptually this is a problem because you have some regulatory pressure saying get rid of this stuff. The user has just told you to get rid of it.

15:39 - Leo Laporte (Host)
And then you have another form of regulatory pressure saying you must keep it and you can't please both masters at the same time and well, I think this is the tip of the iceberg of this type of tension now open ai says we are retaining all content indefinitely going forward based on speculation that the news plaintiffs might find something that supports their case. They say this is rushed. This has really nothing to do with open AI. New York Times you know claims open AI ingested the New York Times content illegally and uh, in fact you may remember, in the suit the New York Times jumped through crazy hoops to make chat GPT spit out actual paragraphs from the New York Times, but it only did that by providing the preceding three or four paragraphs and saying what comes next.

16:29
Openai is going to challenge the order. They're pushing for oral arguments, hoping the user testimony will sway the court to set it aside. Maybe, kathy, you can get involved in this. This seems to be a real overreach. To be honest with you. You, first of all, I'm not a fan of the new york times lawsuit in the first place, uh, but for then the court to order this huge invasion of privacy yeah, there there's I.

16:57 - Cathy Gellis (Guest)
I understand what the court's thinking, because normally you do have to put litigation holds and not delete any evidence that's going to be relevant to a case. So there's, that's not completely out of the blue that's the kind of thing that would happen.

17:11 - Leo Laporte (Host)
If it's the company's email, the company's conversations, right, but this is their users conversation and who is it?

17:17 - Cathy Gellis (Guest)
compromise, and first off, it's getting retained and second off. The reason it's getting retained is because, in theory, somebody else is going to see it and you know there might be an adequate protective order about who gets to actually look at it. But the whole point is that somebody is going to be pulling through this data. In fact, it's going to be multiple. Somebody is pulling through this data because, at minimum, you're going to have an expert on the New York Times side and somebody on OpenAI's site so they can do some analysis on it. So the whole point of keeping it is so somebody can look at it. But that's where the problem is.

17:53 - Larry Magid (Guest)
And don't the New York Times lawyers fight to make sure that their reporters don't have to reveal their sources? I mean, isn't this almost in a similar kind of vein?

18:01 - Leo Laporte (Host)
Isn't that ironic Because there would be reporters sources in there potentially? Yeah, a lot of reporters use whisper and other open ai technologies to transcribe conversations with sources. I mean, a good reporter probably shouldn't uh, they should be doing something locally, but nevertheless, I think it happens so what?

18:20 - Sam Abuelsamid (Guest)
what would actually be in these logs? Would it be the prompts and then the responses?

18:25 - Leo Laporte (Host)
Yeah. And presumably any documents provided in the prompt right? So if I say here's a transcript of my interview with a top secret source in the federal government, can you transcribe it and summarize it? I don't know, but I presume the conversation would be saved as well as the results. That's part of the prompt. That's the. That's the rag.

18:51 - Cathy Gellis (Guest)
Yeah, and especially because I think some of the litigation is really not just talking about what, what went into it, but what was it spitting out, because it's going to look and see where the replications of copyrightable material, but this just seems really exhaustive. It's not at the moment, it's not even a finite universe of. So actually, one of the things to bear in mind in a lot of these discovery disputes is that it's broadly construed but you really only have to produce relevant information and I think there's a pretty strong argument expedition in a way well, fishing, but also that it's overbroad, um.

19:29
And then you know some of the defenses against um producing things are it's not relevant, uh, it's unduly burdensome because this is also going to keep a massive, massive, massive amount of data. Like physically can it actually be archived? They may, you know, open AI may want stuff to get deleted so they can get space back. So there are some defenses to it because even in our litigation system, which has very, very, very broad rules about what is producible, there are some limits, and this one seems to probably cross some of those limits.

20:11
But conceptually, also for what Larry was saying, it's not exactly the same thing, because these are different little vectors of law that are causing something to either be withheld or shared. But one of the problems we have is, like law is kind of splintered, where you get like a vector of law saying about shall store or there that shall save or delete or something, and we don't really have a coherent system that deals with all these different avenues and sometimes when reporters are getting told to produce something, somebody will dig up a vector that came from like another area of law and said, well, given these general rules about what is shareable, you should be subject to it. So it's not exactly shooting themselves in the foot, but it's. The ethos is poor and, yeah, they could actually be making some law that will show up and hurt them later.

20:59 - Larry Magid (Guest)
This comes up in the whole child pornography issue, because prosecutors want evidence preserved, but companies often have privacy policies where they automatically delete data after a certain time. And so you have these two different legal principles both of which have some merit essentially clashing with each other.

21:19 - Leo Laporte (Host)
Well, I don't want to belabor this any longer. Maybe people should use a different ai, which, of course, is exactly what chat ppt is afraid of and probably the new york times is thrilled to hear about if the times you know won this lawsuit, then they would be probably going after the other ai companies as well.

21:37 - Sam Abuelsamid (Guest)
So, you know, using using any ai that's running in the cloud would oh, that's at some point, you know leave you susceptible to the same problem, the same privacy issue.

21:48 - Larry Magid (Guest)
Plus, a lot of companies are actually using the open AI models. I don't know if that would be impacted or not. If someone else is using, you know, open AI LLM, would that also be subject to this court order?

22:01 - Sam Abuelsamid (Guest)
Yeah, I mean if you're using Siri and it, you know, if you're using Siri and it sends a query to ChatGPT or Bing, you know, or?

22:10 - Leo Laporte (Host)
That's why Apple warns you. By the way, when you do that, You're sending data to OpenAI. Your privacy is now moot. All right, we're going to take a break. There's more AI news. We've got a great panel. Larry Magid is here, kathy Gellis, sam abul samad. There's some car stories for you, sam. We got some car stuff. I want to talk about that slate truck too. I mean, that's not, that's old news a little bit, but I'm very curious what you think about it. Uh, but first a word from our sponsor, and a good sponsor. It is when you want to protect yourself.

22:44
Bitwarden, the password manager I use and recommend. Steve Gibson uses and recommends. It's the trusted leader in passwords, secrets and passkey management. Yeah, it's great for passkeys. With more than 10 million users across 180 countries, over 50,000 businesses, uh, worldwide, bitwarden continues to protect individuals and businesses everywhere. G2 consistently ranks at number one in user satisfaction. Every year they do a world password day survey, a survey to. In fact, you can read the whole, uh, the whole survey results on the bitwarden website.

23:23
But one of the things I thought was interesting you often think of the Gen Z, the internet generation, the generation that was born during the time of the internet and grew up with the internet, you'd think they would be the most internet savvy. Right, they may be digitally native, but they are actually, according to this survey, guilty of the highest incidence of one of the most dangerous things you can do with passwords reusing the same password on multiple sites. Survey says a 72 percent of gen z reuses the same password across accounts more almost three quarters even though they know it's you shouldn't. 79 of them say yeah, way no, it's risky. Here's the other thing, the worst thing. So when you know how you get that email from a company that says we've been breached, we're going to reset your password because we don't know, some bad guys got all the passwords first of all. That's why you don't want to reuse a password, because if you use that password on that site and many other sites, bad guys are going to now try that password and log in everywhere, right, and and may get into something you don't want them to get into. So when you get that email, you should change your password, right? 59% of Gen Zers just reuse the existing password when they're updating their accounts. So everybody needs it.

24:40
And if you're a business and you've got employees who are Gen Z and I bet you do you need to know about Bitwarden and their new access intelligence. It's a new capability that helps businesses, help employees, proactively defend against internal credential risks and external phishing threats. Two core functionality here. Two core functionality here risk insights reduces alert fatigue and allows it teams to identify, prioritize and remediate at-risk credentials things like passwords that are showing up in breach sites and so forth. They also have an advanced phishing blocker which will alert and redirect users from known phishing sites in real time. When they go to a site and try to enter their password, it goes whoa, using a, a continuously updated open source block list of malicious domains. The thing about Bitwarden you should know and you'll want because, well, we can see your employees may not be the most sophisticated Bitwarden is easy.

25:36
It's simple. They prioritize simplicity For you. It's easy to set up. It only takes a few minutes. It'll import from most password management solutions, so you know you're ready to go quickly.

25:48
Of course, one of the things I love about bit warden In fact, it was the very first thing that brought me to bit warden it's open source GPL license, so anyone can expect the code on github. It's regularly audited by third-party experts and it meets stringent security and compliance requirements SOC 2, type Type 2, gdpr, hipaa, ccpa, iso 27001-2002. You and your businesses deserve an effective solution for enhanced online security. So get started today with Bitwarden's free trial of a Teams or Enterprise plan. And I know all of you listening are good with your password hygiene. You don't reuse passwords. You already use Bitwarden or another password manager. But if you don't or, more likely, if you know somebody who don't we all have friends and family who have the worst password hygiene Tell them about Bitwarden. It's free forever across all devices. When you're an individual user, it's open source across all platforms. Unlimited passwords, pass keys to bitwardencom slash twit. I love Bitwarden. You will, too. Bitwardencom slash twit.

27:01
Back to the news and the house are trying to pass does have a provision in it that would force states to not regulate ai in any way for 10 years. Right now, I'm not a big fan of ai regulation, but that does seem like a bit of a first of all. It's a long time, uh, and it does seem like the state should get to do that. Uh, it is not likely that that provision will be removed in reconciliation. That's. That's not what the sticking point is for this bill.

27:37
So um rokhana uh, from whose district includes silicon valley, he's in the house, says what this moratorium does is prevent every state in the country from having basic regulations to protect workers and protect consumers. Uh could restrict state laws that attempt to regulate social media companies. Prevent algorithmic rent discrimination. That's probably part of the problem is this AI is very broadly defined in this and anything algorithmic could be considered AI. Prevent, limit AI deep fakes that could. They can't pass laws against AI deep fakes. It would basically, he says, give a free reign to corporations to develop AI in any way they wanted to develop Automatic decision making without protecting consumers, workers, workers and kids I have mixed feelings about this.

28:28 - Cathy Gellis (Guest)
Um, I am generally a fan of federal law preemption. When it comes to particularly communicative technologies, um, but maybe even technology in general, like this is a big issue that comes up with internet regulation. It is a problem when each state comes up with its own rules. This is why Section 230 is so useful, because it preempts the ability of individual states to try to regulate the internet. Because if we look at like, are we going to get the same regulation out of California that we're going to get out of Texas? No, and what does a company do when the two sets of regulation don't match and don't meet? You can also see things where the patchwork problem, the patchwork well, that's also a.

29:13
There's no way you can make them all happy, um, particularly for the cross-border communications thing. But then we also see, in terms of privacy law, that every state wants to do it their way. Then it's patchwork, and then you've still got an enormous compliance problem.

29:27 - Leo Laporte (Host)
Because even if the rules aren't inconsistent there's too many rules. If the feds say well, we're going to pass a law regulating AI that preempts state law, but this creates a vacuum.

29:38 - Larry Magid (Guest)
It's not only preempt, it's important.

29:40 - Leo Laporte (Host)
There's no federal AI regulation either.

29:44 - Larry Magid (Guest)
Kathy, I agree with you and in my article, which you'll find on the right column in the front page of Connect Safely, I make the same point that generally speaking, I don't like state laws because of the patchwork and conflicting each other, and it's impossible. It's hard enough in a global environment. But you know an American government. But in this case, my theory is and I'm not even a theory. I mean, if you look at what's going on with the Trump administration and their tech bros, this is an attempt for them to essentially eliminate the regulation of AI.

30:16 - Leo Laporte (Host)
And if the federal government this may be the last thing Elon did before he burned his bridges.

30:21 - Sam Abuelsamid (Guest)
Right, and this is an administration that is fundamentally opposed at every level to any kind of regulation. Right, if they could, they would get rid of every federal regulation which while I, again, I agree with you. Uh, kathy, you know that a federal regulation that does preempt the state regulations would be better. Well, could, could be, could be better, wouldn't necessarily be better Wouldn't necessarily be better.

30:45 - Leo Laporte (Host)
Look at Maria Cantwell's privacy law which, as part of it, did in fact preempt all state privacy laws, but it didn't replace them with a better law, it effectively defanged them with that same argument. Well, you can't have a patchwork quilt of laws without providing a better alternative.

31:04 - Cathy Gellis (Guest)
Well, better also tends to be relative. I do think that is important actually, that we still do get a federal privacy bill, and I think one of the critical things that we get out of it is preemption, because I think we've got some real problems with this.

31:20
I think I might take it, even if it wasn't necessarily a good law, because I think there are some externalities.

31:26
We're kind of glossing over with what happens when every state does it its own way and it gets really inconsistent and we have some big not insignificant impingements on speech that can happen, and there's other collateral problems, but also just the you know, a huge compliance problem where you stifle innovation as well because you have to have more lawyers on your staff instead of more engineers, among other problems. But in terms of this, I think one of the things that's emerging from this conversation is the sense of it's one thing to preempt prospectively when there's nothing and it's another thing to preempt when you have something. The Congress actually is speaking in that area and this is just kind of creating a vacuum. It's probably both under inclusive and over inclusive. The definitions are iffy and I get the sense that this may have been proposed and promulgated by people I like who should have known better, because it kind of seems a little bit more like making a statement about a lot of the dumbness that comes up with how technology gets regulated.

32:31 - Leo Laporte (Host)
You give a good example in your article about california's ab-3030, the healthcare services artificial intelligence act, which requires healthcare providers to disclaim it when they're using ai, for instance. I just was in the doctor. He's got a little sign in his office saying I'm using ai uh, to record this conversation and analyze the conversation, but lead immediately afterwards. I use this for notes so I can pay attention to you instead of typing the whole time, which you used to do. Yeah, that would be a good law and the fact that it's only in california.

33:03
How often does California make a law that, because it's a good law, ends up changing the way AI works or some business works?

33:12 - Larry Magid (Guest)
nationally and it was overwhelmingly supported by both parties. And, by the way, marsha Blackburn, the Republican senator from Tennessee, is concerned that this bill would, or clause would, preempt the Elvis law.

33:25 - Cathy Gellis (Guest)
Which it should which it should. Yes, I don't know.

33:29 - Larry Magid (Guest)
But the point is that there is bipartisan concern about it.

33:31 - Leo Laporte (Host)
The Elvis law, which is a law in Tennessee, the Ensuring Likeness, voice and Image Security Act protects Elvis, I guess, against AI impersonation. Is that the point?

33:44 - Cathy Gellis (Guest)
I think the reason that this law got, that, that this particular language managed to get promulgated, um and provided to somebody who was able to get it into the bills, is because there's so much stupid. There's stupid in the states, they're stupid in the federal government and the idea is, I think, to just kind of like press pause a little bit so that you know, do policymaking in this space, but do it sensibly from a federally peremptive perspective.

34:13 - Leo Laporte (Host)
In favor of light touch regulation, though on the Internet and now on emerging technologies, because we don't know yet, right.

34:21 - Cathy Gellis (Guest)
Well and like. So the bill that you're talking about from California. Ok, that provision kind of sounds pretty benign and probably sufficient for, you know, even states to be able to do, because you're really sort of regulating something so local, so there are some good things that can come out of the states. But we also know that the states, including California, looking at you, have a habit of taking gigantic hammers to things and producing really dumb things. So, in which case you really want the federal preemption because you want the states to stop doing dumb things. So I think I think this was a little bit of a stunt and mostly to kind of just press pause on things so you could get more space.

35:00 - Leo Laporte (Host)
That's going to be a federal law, though.

35:02 - Cathy Gellis (Guest)
And I think the problem is, yeah, it's a stunt that is probably going to escape, you know, control and end up becoming law and we'll have to live with the consequences. But maybe the consequences let's. Let's look at the silver lining If the states can't act, maybe it does prompt the Fed to do something. Now the downside is they could do something that's terrible, and now they've got more political pressure to do something anything and are more likely to do something that's terrible. But on the other hand, we won't have 51 versions of terrible, because we'll just now have the one version of terrible.

35:33 - Leo Laporte (Host)
I'll give you an example of terrible. Federal judge blocks florida from enforcing social media ban for kids. Uh well, florida banned the use, banned, outright banned the use of social media for children under 14. Right, and I mean you. Even if the parent says, yes, you can't. And if you have a 14 or 15 year old, all the parents can get permission. Yeah, uh, that is unconstitutional, clearly.

36:01 - Cathy Gellis (Guest)
Yes, well, we hope, clearly. But and it's good that this court saw it but occasionally we hit courts like looking at you, fifth circuit that tend to struggle on these things right. So there's other cases that are um free speech, coalition versus media. Matters came out of the fifth circuit, where they did not recognize the unconstitutionality of these things, and there's a new litigation that's just hit it about another one of the age-gating legislation that came out of Texas.

36:30 - Larry Magid (Guest)
We actually hired a policy consultant because of this. I mean, we're an internet safety organization, but we also believe in things like the First Amendment and the UN Convention of the Rights of the Child and other silly little documents that say you don't have to be 18 years old to enjoy the right of free speech. So there's a lot of these things that we wind up having to fight, even though they're done in the name of what we advocate.

36:52 - Leo Laporte (Host)
I'll tie on this Hill. Tell me what you think of it, larry. That it really should be not the government's job but the parents' job.

36:58 - Larry Magid (Guest)
It should be the parents' job and also we should recognize that there are young people, but some of them are LGBTQ. Some of them have religious ideas or political ideas that differ from their parents, and they should be able to explore and express themselves even if their parents disagree. Now, I'm not saying that there should be a free reign, that you know four-year-olds should get to go to porn sites, but I am suggesting that preteens and teens deserve some privacy and some free speech.

37:24 - Leo Laporte (Host)
I agree on that, but I don't think it's practical for the government to say oh, parents, you can't block your kids from LGBTQ sites either. So I think that the reasonable compromise is parents should be given the tools they need to make these decisions for their kids. I think it's really clear that, in fact, apple's already implemented this. Apple should have an api, a way for an app to query the device saying is this kid old enough in the parents eyes to use this right?

37:58 - Larry Magid (Guest)
and the big controversy between apple and meta over that well, there's a difference.

38:02 - Leo Laporte (Host)
Because meta wants apple to block stuff in the app store. They want them to be the arbiter. I think the parents should be the arbiter. Apple gives them a tool to you. You should have on android and ios a tool as a parent that says my kid emotionally is eight or yeah, I don't care what chronological date, but but emotionally they're an adult and you should have that ability as a parent to set that. And then apps should have a way of querying it and would be required to adhere to it. That would give parents control over it. Now, I agree with you, larry. There are many cases where a parent shouldn't have control, but sorry, that's the way it is where a parent shouldn't have control.

38:47 - Cathy Gellis (Guest)
But sorry, that's the way it is. I'm, I'm, I'm, I'm, I'm more on the side of you. Know, legally, a proposal like that may have fewer of the problems that a lot of the age gating does, because I think you're just saying that it's a check for whether the parent has given permission, as opposed to whether it's recording anybody's true identity and verified things like that right and you don't have to give age verification of any kind as an adult or a child, and I also understand that a kid could go to his neighbor's house and use the their phone.

39:12
There's no perfect solution to this well, there are lots of imperfect solutions I do think, like what larry was starting to get at, that there is a risk vector itself, if um, to the rights of kids too, if um, if they're, if there's somebody, their parent can completely overrule and essentially cut them off from the world that there is a problem there, although that's the way it is right now, in every respect in a child's life not when it comes to reproductive health.

39:39 - Larry Magid (Guest)
California, I believe a minor can have an abortion without parental permission. I believe that's the case.

39:44 - Leo Laporte (Host)
I don't know if anybody can update me on that one, but if a parent wants to have their kid get a certain kind of education, wants to keep them from leaving the house, wants them to dress like they belong in the Handmaid's Tale. A parent can do that and there's no law against that, and there shouldn't be, I believe, because that was overreached by the government although in europe there are, you know europe, children have privacy protections vis-a-vis their own parents.

40:10 - Larry Magid (Guest)
In america those, those laws don't exist.

40:12 - Leo Laporte (Host)
No, they do lisa, you know, I remember lisa, when she took her 13 year old to the doctor. The doctor said okay, mom, bye. Yeah, yeah and she can't get her son's medical records. I remember I couldn't get my son's radical records, uh, when he was under 18. So there are protections like that. I suppose they're good. I hate to see government tell parents what to do. I admit there are bad parents, but I don't think that.

40:37 - Cathy Gellis (Guest)
I think that's, I don't think government can fix that problem the problem is that the with this policy problem, what you're basically saying actually, let me make two points. One is you're basically saying that if the government can pass a law saying we should have this apparatus, you've already gotten the government involved and the government is placing its finger on the scale of how this is all going to get worked out. If Apple itself wants to offer this as a feature because it thinks it's a marketable feature and parents are going to want to buy, if you're going to buy your kid a phone, buy one where you can have this ability to cut them off from your kids, off from certain things. And fine if it's a private, if it's a truly private choice between the consumer and the and the company. You know, whatever.

41:20
There may be some social externalities to study, but I don't. I'm not as bothered by it from a legal standpoint. But if you have the state saying there needs to be a law to require Apple to do it, then I think we've really sort of right. The state is starting to navigate these really fraught issues in ways that are going to cause harm, even if they might.

41:39 - Leo Laporte (Host)
In this case? In this case, we have we have societal norms, we have moral authority. We don't need a law.

41:48 - Larry Magid (Guest)
Apple, do the right thing right, that's all we have to do, but how far is the right thing for Apple?

41:55 - Sam Abuelsamid (Guest)
Should it be just this emotional age type of threshold, or should they also give parents a list, a list of here's topics that you can block?

42:08 - Larry Magid (Guest)
uh, how far, how far should they go with emotion, emotional age, which I, actually, everybody who's an expert in this field agrees. That 13, 18, all these numbers arbitrary, arbitrary meaningless but it's pretty hard to regulate around emotional age, so there will be an age.

42:24 - Leo Laporte (Host)
The reason you have to do that but, I'm talking about any other regulation I, any other system is going to require some sort of age verification. That's right out. You can't do it. That's. There's no way to do that safely or privately. That's just a bad idea. So it has to be the parent is the judge. It doesn't even have to say emotional age, what's your child's birth date? But the parent can set it no.

42:44 - Cathy Gellis (Guest)
With that, that's uh, then you're starting to get into uh, age verification. If you're starting to specifically no because it doesn't distribute the birth date.

42:51 - Leo Laporte (Host)
All it is is an api that says can? Is this child allowed to use this?

42:56 - Sam Abuelsamid (Guest)
you don't know what the birth date is and there's no, no, no, no, that's the only person responsible for that as a parent.

43:01 - Leo Laporte (Host)
There is no verification. The only person responsible that's the parent.

43:04 - Larry Magid (Guest)
I think the argument that Meta makes and it's a pretty good one actually is that when you buy a child a phone, you're handing them the phone, you're the parent. You can set their age. When they go off today and sign up for an Instagram account, they tell Instagram how old they want Instagram to think they are, and there's really no checks on that Now. There's really no checks on that. Now there's all sorts of technology that's being built and that's around that's called age assurance.

43:28 - Leo Laporte (Host)
But I understand Meta does not want to do it.

43:30 - Bill Atkinson (Interviewee)
Yeah.

43:31 - Leo Laporte (Host)
They want Apple to do it in the App Store. What Meta should say is Apple, you need an API that we can query that says is it OK or not? Right, and that's all. And we just query it Is it OK or not? And the parent could say never okay to install Instagram, sam. You could have a setting that says never okay to install social media apps, whatever Apple wants to do. In that light, the only thing Meta needs to know is is it okay or not? It's a binary choice.

43:56 - Cathy Gellis (Guest)
I have a concern about that, though, because if you have in the bad parent situation, if the parent on the phone context is able to veto that, facebook can be installed, that's going to carry over to any other avenue that the kid is going to potentially need access to, because once the parent on the phone has said, oh, you don't get it, how are they going to be able to access their Facebook account? Anywhere else, including if they needed to.

44:24 - Leo Laporte (Host)
This can't be solved on the app store level. That's why you have child protective services. That's why you have a. You have state and local means to protect children. The app store is not going to be doing that Well, it's also why I've.

44:38 - Larry Magid (Guest)
it's basically what my whole career has been dedicated. You also need to educate.

44:42 - Leo Laporte (Host)
Education, I agree, and ultimately, as I said but making these companies responsible for it is not the path Right and I think 30 years ago.

44:49 - Larry Magid (Guest)
The best filter is the one that runs between the ears. You've got to train your children to be critical. Yeah.

44:55 - Cathy Gellis (Guest)
This is a sociological problem, not a legal or a technical problem, although I would point out. The second point I was going to make is we are essentially hitting the heart of Section 230, and what caused it to be passed in the first place which was one of the features of Section 230 was to make sure that filtering technologies could be developed and used where it was trying to create these space where those could exist, and you know we like innovation here. There's cool things that technology can do. I'm not necessarily averse to like seeing if there's some technology that can be helpful in some context, but I think it is a mistake and I think we're agreeing on this that this idea that technology can definitively solve it, especially if it does things in a certain way. I think that's pie in the sky and potentially damaging to some very legitimate interests.

45:44 - Leo Laporte (Host)
Yeah, all right, we're going to take a break. We'll come back with more. Got lots of conversation to do. Larry maggid uh, the ceo and founder of connect safelyorg. It's all about keeping kids safe online. Lots of good information there. Connect safelyorg is the website, and if you google, family guide to parental controls.

46:02 - Larry Magid (Guest)
You're going to get a lot of information on this, including some advice on what parents can do.

46:07 - Leo Laporte (Host)
Do you think adequate the Parental Controls Apple and Google offer today are adequate?

46:12 - Larry Magid (Guest)
They're moving actually closer to the meta model. I think they're good. Adequate is the problem with adequate? It depends on the child. I think it's adequate for most kids, Right, but there's always exceptions.

46:24 - Leo Laporte (Host)
But that's all you can hope for. Look, this is a big world. It can't be perfect, I don't think. Uh, kathy gillis also here, as you can see a legal expert, and it's always good to have a legal expert when you're talking court cases here. Nice to have you, kathy, and we'll get some car stuff for you. Sam will say that wheel bearings podcast and vp research at telemetry. Stay tuned. Lots more to come.

46:48
The show brought to you today by shopify. It's difficult when you're starting a new business. You know I, you can. I know us by heart, learning all the new hats you have to wear. It seems like your to-do list is growing every day with new tasks that can easily begin to overrun everything else your entire life. Right, finding a tool that not only helps you out but simplifies everything can change everything for millions of businesses. That tool Shopify.

47:23
I love Shopify and you know, so do my kids. Both my kids have online businesses using ShopPay Shopify's store. They're not alone. Shopify is the commerce platform behind millions of businesses around the world. 10% of all e-commerce in the US, from the big guys like Mattel and Gymshark to brands just getting started like Salt Hank, use Shopify.

47:49
It's so cool that Henry could go from being, you know, a TikTok celebrity to creating an online store where he sells his salts and his pickles, and and and apparel is, and he has. You know, I? I know he's my son, but he's not exactly a web designer. He was able to make a great site with the help of Shopify. My daughter same thing, in fact. I just helped her set up her website for her book that she just published and she's selling online. And she said, dad, I need to connect my domain name to Shopify. And I said easy to do. And she did it. Get started with your own design studio with hundreds of ready-to-use templates.

48:32
Shopify helps you build a beautiful online store to match your brand style. Again, I know this directly. You can accelerate your content creation. Shopify is packed with helpful AI tools so they can write product descriptions, page headlines. They can even enhance the product photography. Get the word out like. You have a whole marketing team behind you. You do. You have Shopify. Easily create email and social media campaigns wherever your customers are scrolling or strolling. And, best yet, shopify is your commerce expert, with world-class expertise in everything from managing inventory to international shipping, to processing returns and beyond.

49:11
If you're ready to sell, you're ready for shopify. Turn your big business idea into. With shopify on your side, sign up for your $1 per month trial and start selling today at shopifycom slash twit. Go to shopifycom slash twit. Shopifycom slash twit. I am seriously a Shopify fan. I think this is empowering technology. S-h-o-p-i-f-ycom slash twit. Take a look at what they can do for you. Shopify. I love that sound. I don't know, does henry hear that every time he sells some salt he must be going crazy right now. Oh hey, just some good news. Uh, while we're on the subject of uh, government tick tock. Uh is back. Uh, it's a 75 days at a time. The the president has once again extended the uh the pause and the tick tock ban. Remember, congress passed this. The president, former president, signed it. President trump immediately put it on pause talks. President Trump immediately put it on pause talks, supposedly for a sale underway.

50:28 - Cathy Gellis (Guest)
I haven't seen any movement there. Yeah, he doesn't lawfully have the ability to put it on pause because the predicate things that needed to happen in order to put it on pause haven't actually happened. And I remember sitting in like the front row of the Supreme Court while oral argument was happening and listening to Justice Sotomayor talking about. I don't even want to think about what it means for the president to not be enforcing a valid law, and yet here we are.

50:53 - Leo Laporte (Host)
The Supreme Court already said he was immune from prosecution for any.

51:01 - Cathy Gellis (Guest)
Well, she did, didn't, she didn't but the majority of the court did.

51:06 - Larry Magid (Guest)
Yeah, well, we go by majority rule here in this country well, this is one of those cases where I agree with the president's decision but disagree with his right to make it yes, I think that sums it up, yeah it's kind of the opposite of free speech, where you disagree with what they say, but you know he had no right there is so there is so much other unlawful stuff going on that this seems de minimis compared to all the other things.

51:29 - Cathy Gellis (Guest)
The problem is, yes, there are and there's vastly bigger fish to fry. But when you start chipping away at like, not caring about like law actually doing its law thing, then it becomes really difficult to have law be available to do the things you really needed to do historians may look back and say it all started with tick tock I? I think. No, they'll, they'll. The clock will start earlier than that.

51:53 - Leo Laporte (Host)
Garcia is going to be part of that as well well, uh, I don't know when did they write the insurrection act 1807?

52:00 - Sam Abuelsamid (Guest)
uh, wasn't it like?

52:01 - Cathy Gellis (Guest)
no, it was earlier than that. Yeah, yeah 1998 I think I.

52:05 - Leo Laporte (Host)
It seems odd to me that that oh ancient bill would still be invocable today. And where's the insurrection anyway? But anyway, that's the big issue, though.

52:14 - Larry Magid (Guest)
The constitution was written a long time ago, but you're right, where's the insurrection?

52:18 - Cathy Gellis (Guest)
oh, I was thinking not the insurrection act, I was thinking the ae aliens sedition act another fine anti-first amendment bill they passed.

52:26 - Leo Laporte (Host)
It's so funny the founders passed the first amendment and then within years they created the aliens seditions act, which is exactly the antithesis of the first amendment.

52:37 - Cathy Gellis (Guest)
One of the things that that happened, if you look at that history is, uh, john adams, I think, was big on pushing these things and you know he didn't have the supreme court, wasn't quite booted up the way it has been booted up, so it wasn't challenged, but it was politically incredibly unpopular and it did in his political career to the point that, like, some of this just ended up falling off the books. So, um, there is a cautionary tale of like even if you can manage to squeeze these kinds of autocratic things through, um, you know, the people don't actually like them.

53:11 - Leo Laporte (Host)
So Supreme Court has allowed doge to access Social Security data. This is a very tangled web here, because you've got the social security data, the irs data, the opm, the office of personnel management data, all of which are supposed to be siloed because they contain lots of information, personal, personal information about people. Then you have doge coming in saying we want to combine it all, and now we learned that Palantir has just made a deal to apply its database searching techniques to these databases. The US Supreme Court granted the Trump administration's request to access sensitive social security information. The protections that kept that information out of the hands of other agencies are gone.

54:07 - Cathy Gellis (Guest)
So I was talking about this case the last time I was here, because there had been the injunction saying that they had to stay out of the Social Security Administration, and I was pointing out one of the issues that was kind of percolating out of it, which is there's a fourth amendment problem, which is the government is collecting all this data that it doesn't have a warrant for.

54:30 - Leo Laporte (Host)
so it's really collecting it via consent, but the consent is sort of contingent on the privacy act keeping it in these when I turned 65, I gave the government a lot of information, including ids, because I had to use id me which I wasn't thrilled about to sign up and all this stuff for my social security. Uh, but I thought I was giving it to only the Social Security Administration and it was being siloed and carefully protected. Now some 20 something named big balls has access to it yeah, um, I, and what happened is not so.

55:05 - Cathy Gellis (Guest)
Once again, the Supreme Court did a tiny thing that ended up huge, but they think it's tiny, and in what context? It is where there was an injunction and what they have done is stayed the injunction, so they're treating it as kind of this minor thing. They haven't overturned it, they haven't looked at the substance of the underlying case, they haven't said anything. It's just that the injunction doesn't get to function while the litigation continues. But functionally, that means now there's no functioning injunction, so all the people who are just restrained are no longer restrained I think I saw that on uh, the function injunction.

55:41 - Leo Laporte (Host)
Uh, I'm just a bill on capital I. I feel like that might have been one of the songs on Schoolhouse Rocks. I might be wrong.

55:48 - Larry Magid (Guest)
I thought it was a 60s sub-comedy about trains and something junction.

55:55 - Leo Laporte (Host)
So here's the question. No one knows the answer. This is purely philosophical. But does Doge still exist now that he's on his left town?

56:02
Yeah, it's still there Is Big Ball still embedded in all of these agencies. Does Doge still exist now that he's on his left town? Yeah, it's still there. Is Big Ball still embedded in all of these agencies? I believe so. Yeah, you know, a month or two ago there was a whistleblower at the Veterans Administration who said that when Doge got access to all the database of the Veterans Administration, they turned off all logging first thing. They did, uh, and then exfiltrated 10 gigabytes of data to some place we don't know and then, within minutes, russian accounts were logging in using doge credentials into the veterans administration.

56:40
Right, this whistleblower was a sysadmin, uh, it got a lot of attention. Npr covered it. I saw uh. He was interviewed by rachel matto on msnbc. He was threatened uh, there was a note taped to his door with a picture of him walking his dog from a drone shot, saying you might want to shut up about this. And then we've. And then, by the way, the uh I can't remember, was it sisa one of the federal agencies said yeah, we're not going to investigate this, and I haven't heard a word about it since now. I don't know what happened. I would like to know what happened, and it is is of, I think, a good reason to be concerned that these doge guys, who some of them have very shady backgrounds, including relationships with Russian hackers, are in there and are being given access to our most private information Information we gave the government with an understanding that it would remain private government with an understanding that it would remain private.

57:47 - Cathy Gellis (Guest)
And one of the things about these injunctions is that like the one for the Social Security administration, I believe was basically it didn't keep them out completely, it just sort of throttled them through. You only get onboarded into these systems if you go through all the trainings a normal person would have to go through. Like you know, if, if you're sensible, you're trained and you're independently trustworthy, you still can come in. But they couldn't come in for, necessarily, things that were not particular to, uh, the way the office was normally run. Run, but they weren't even completely shut out. It was a very reasonable kind of judicially supervised arrangement of where they could still get access if they could sort of make the claim that they had some legitimate need for it.

58:27 - Leo Laporte (Host)
Incidentally, doge is protected. Uh, the people have argued that they should be considered an agency and then therefore, under the federal freedom of information act, we should be able to see a wide range of their records. The trump administration says no, no, doge just plays an advisory role. They're exempt from foia right this is completely.

58:47 - Sam Abuelsamid (Guest)
Why do they have access to all of this information, right? Yeah, if they're, if they're nothing but advisors, they shouldn't even be in here.

58:54 - Cathy Gellis (Guest)
Yeah, and some of the district court judges have called them on this, to say you're only an agency when you want to be an agency and then you're not an agency.

59:01 - Leo Laporte (Host)
So the inconsistencies have been noted but just like elon is only an employee of the government when he feels like it, yeah but this is the supreme court kathy, this is it right.

59:12 - Cathy Gellis (Guest)
It's over, there's no um, well, it's over ish. Um, the problem with these things is they were all shadow docket issues and and the justice.

59:22 - Leo Laporte (Host)
There's no opinions written.

59:23 - Cathy Gellis (Guest)
They're just a vote no opinion was written, there was a very, very little, a little bit of text that came out with with like, there was a rush to the shadow docket to petition for the stay and the stay was granted with very little text.

59:38
But justice Jackson wrote a dissent and is basically calling on them and saying you know, we are doing tremendous things with tremendous effects, with no merits, no briefing, you know, way too early in the process.

59:53
Well, we're still learning and the record is being established and and she's just pointing out that we have that it's ridiculous for the Supreme Court to do so much of such magnitude in its effect, so carelessly and via this, you know, really shortcutted, abbreviated, non process. So I don't know what happens in these, but I wonder if the district courts are going to start getting creative. I also think that in a lot of the litigation, the litigants are going to stop asking for injunctions and start to ask for expedited briefing on the merits, which is what we saw with the Institute for Peace who lost the TRO, which I think was an abomination. Ok, we're basically going to do briefings on summary judgment within weeks and now there has been a decision on the merits, and that kind of puts the whole thing in it and with summary judgment on the merits, they got their building back because now she doesn't have to do a preliminary injunction, she can do a real injunction because they've been found to have acted unlawfully, that's possible with buildings.

01:01:03
Well, it was more than the data is exfiltrated.

01:01:05 - Leo Laporte (Host)
It's not possible.

01:01:06 - Cathy Gellis (Guest)
Well, that's true too, yeah.

01:01:08 - Leo Laporte (Host)
And if they have which they certainly have by now copied the database it's out. I don't care what the court says in two weeks or two years, it's done and odds are it was copied a long time ago before this.

01:01:23 - Sam Abuelsamid (Guest)
yeah, this ruling came up right, yeah, yeah what's extra garbage about it is.

01:01:29 - Cathy Gellis (Guest)
It's not like preliminary injunctions are a thing that only rarely exists. They are a major point of, you know, litigation. We have them because you want to be able to preserve the status quo and make sure that harm doesn't happen. And the supreme court is basically completely upending centuries of I mean literally centuries of law, because, you know, injunctive relief comes from the common law in england, so like they're just completely upending all the all the rules that ended up governing that and we no longer have them available as a tool.

01:02:01 - Leo Laporte (Host)
Let's cheer everybody up. I remember Elon tweeting that the IRS direct file was dead. That day I immediately went and tried to use it. It was before April 15th and it wasn't dead. It was fine. I created that nice ID me account and all that. Well, it's dead.

01:02:20
This is something that Intuit, the creators of TurboTax, h&r, block and other commercial tax filing software, did not want the federal government to have a free filing system because that would be bad for business, wouldn't it? Well, they got their wish and the IRS direct file system was killed. However, thanks to Jason Keebler writing in 404, direct file has now been open sourced. It was originally created by the us digital services and, of course, doge has replaced usds and that software, which was created by usds. Oh, and 18f, that's the uh, the other company that doge killed other agency. Doge killed um, apparently 300 000 people used it last year as a limited pilot program. Very positive reviews, uh, because jason writes it's free and because it's an example of government working.

01:03:22
Direct file and irs is free file program more broadly been the subject of years of lobbying efforts by financial technology giants like intuit to kill it. Doge sought to kill direct file again. Elon said it's dead. Uh, there is no language in the great big ugly bill that would kill direct file, language in the great big ugly bill that would kill direct file. But the irs, you know, kind of I think, like the ships are burning, quick, save the treasure have released the droids in the uh escape pod yeah, the droids are in the escape.

01:03:58
Pod the source code's on github but what about updates?

01:04:01 - Larry Magid (Guest)
every the you know, turbo tax has to be updated at least once a year, every year, yeah, yeah, there's a lot more change. So who's going to make?

01:04:07 - Leo Laporte (Host)
well, this will be an interesting uh thing. You know, the three people who worked on it at usds chris given, jen thomas, marichi vinton have left the government to join the economic security project's future of tax filing fellowship. They'll be joined by gabriel zucker, who also worked on direct file as part of code for america. So there is an effort to in effect, create a, a non-profit that will continue to support this outside of government that's good, but I think the government should do it.

01:04:43 - Larry Magid (Guest)
I mean, of course, the government should do it.

01:04:44 - Leo Laporte (Host)
I mean of course, the government should do it, make it easy to process it.

01:04:47 - Cathy Gellis (Guest)
You obviously do not work for intuit, mr you're not well, one of the things that kind of excites me about this is so much has been lost that we spent money on with paying the right personnel to build good applications. So what I'm hoping is that this sort of archives it a little bit so we go through this period and then, when we have a normal government again, we don't have to start from scratch because we have something that you know hasn't been lost forever, Cause I worry about that.

01:05:14
That to replicate all this infrastructure that's been destroyed is massively expensive and not a savings at all.

01:05:21 - Leo Laporte (Host)
Well, check out the economic security project. That is a group that is trying to keep this alive, uh, as an open source project. I think there's enough people who care enough about this. Uh, who would that? Perhaps this will survive also one other nice thing.

01:05:40 - Sam Abuelsamid (Guest)
Wouldn't it require some? Uh, I assume it requires some cooperation from the IRS, some approval, some certification.

01:05:47 - Leo Laporte (Host)
They publish. Whenever these rules are changed, they have to be published and that's all Intuit has. That's all you know H&R Block gets. They look at the rules and they incorporate into the software. I don't think you have to get the IRS to help you.

01:06:00 - Cathy Gellis (Guest)
I mean, I suppose you may have to plug in in terms of if you were going to file the return you probably need some api and then pay. You need some sort of api access so that that I think may be true, but you've got direct file. I mean there's going to be maybe all you need.

01:06:14 - Larry Magid (Guest)
Yeah, yeah let intuit compete by being better. I mean, that's yeah no, no, no, no.

01:06:20 - Sam Abuelsamid (Guest)
That's not the way capital is are you new around here?

01:06:23 - Leo Laporte (Host)
we don't do that it turns out that this was always a part of the plan open sourcing. The project was required by a law called the share it act. Uh, again one of the things that'll probably be thrown out as soon as trump realizes it exists. But um, you know, there's some good stuff.

01:06:45 - Cathy Gellis (Guest)
He can't just things that are statutorily mandated. I mean, he's been ignoring a whole lot of that you mean, like banning TikTok. Yeah Well, things that are statutorily mandated. We're having some issues with him obeying the law, don't you?

01:06:58 - Leo Laporte (Host)
know it's an emergency that the EO trumps the EO.

01:07:01 - Cathy Gellis (Guest)
It's a different thing entirely of getting rid of the law entirely. So the law remains on the books. You'd have to get a whole bunch of his republican friends in congress to get rid of it.

01:07:11
So a whole bunch of republican friends, yeah I just not that many yeah I mean they only have a three-seat majority in each house yeah, well, sometimes, um, when it comes to tech, they get some democrat people who don't know what they're doing and they opt for the bad things too. But there was one other Doge thing. Before we move on from it, there was another decision that I think is worth keeping an eye on, which is most of the lawsuits have been attacking Doge, either in terms of Musk and Doge and also the agency that that is at issue with that particular case. But there were a couple that were suing Doge and also the agency that is at issue with that particular case. But there were a couple that were suing Doge in general as saying you are not an entity that has any business running around doing any of the things that you're doing.

01:07:54
And one of the big cases was brought by a bunch of states, led by New Mexico, and that was an early case where the judge there kind of had a sense of this does not seem like something you should be able to do, but it was dicta at the time there was no injunction. But now they've survived the case has survived a motion to dismiss where there's at least enough, or now a judge has said you know, I have a finding that, at least in terms of what we evaluate at this stage of the case, it looks like you have no business doing the things that you were doing. So that's an interesting benchmark that we've just gone through and that case will proceed and we'll see where that goes. But that was. If you are not a fan of the doging, this was a good ruling to be aware of.

01:08:40 - Leo Laporte (Host)
You're just trying to cheer me up. Yeah, yeah.

01:08:46 - Sam Abuelsamid (Guest)
Well, you know I I mean this whole doge thing has been such a sham from the beginning. You know, they've they've already acknowledged that they're not going to save very much money of any at all. And there was a an npr interview this week.

01:08:57 - Leo Laporte (Host)
What the point of doge?

01:08:58 - Sam Abuelsamid (Guest)
was a former staffer, uh, that was uh working in in the va um a doge staffer that was working in the VA, a Doge staffer, and you know he acknowledged that. You know they found that federal waste, fraud and abuse that the agency was that they were supposed to uncover was relatively non-existent. I quote I was. I personally was pretty surprised actually at how efficient the government was. And you know I mean this is true of a lot of agents like Medicare. The Medicare system is way more efficient than any of the private health care health insurance companies. You know they pay and the US Medicare system pays about three percent, spends about three percent on administrative costs, compared to about 17 to 18 percent for most private health insurers.

01:09:43 - Larry Magid (Guest)
Well, you know, the irony about the Doge savings is that on the day that Elon and Trump got into their little pissing match, tesla lost almost as much market value as he claims to have saved in Doge. He claimed to have saved about $180 billion. Tesla went down $140 billion that same day, so it's up a little bit, and even that $ billion is is is subject to a great deal of skepticism.

01:10:10 - Leo Laporte (Host)
Yeah, yeah uh, I have, uh, an interesting story we're going to take a break and come back and talk about that. May explain this is from quartz why there have been so many layoffs in the tech sector over the last two years? Uh, because I you know it's. You look at, for instance, the most recent example Microsoft laying off 6 000 uh programmers the day after they announced record profits, and it really makes you wonder. You know what's going on. Why are these companies? Is it just to appease the stock market? Uh, or or do they really have too many employees? What's going on? Why are so many people being put out on the streets? Well, it could have to do with the tax code. We will get to that in just a moment.

01:10:57
Larry maggott is here, uh, kathy gellis, samuel bull salmon we're glad you're here too. We do twit every sunday afternoon. You, canmon, we're glad you're here too. We do twit every Sunday afternoon. You can always watch us live if you're around, around 2 pm Pacific, 5 pm Eastern, 2100 UTC. We stream it on eight different platforms, including our Discord for our club members, but also on YouTube, twitch, xcom, hi Xcom How's Elon doing? Huh, tiktok, kik. Facebook, facebook how's mark doing? And linkedin on all the platforms, so watch live if you want. After the fact, though, on demand versions of the show available on our website, twittv and wherever you get your podcasts. We're glad you listen, whether it's live or after the fact. We appreciate, appreciate you being here.

01:11:43
Our show today, brought to you by ZipRecruiter oh man, summer's here. Seasonal businesses are hiring Everything from mule packers to drama camp leaders. That means people with pretty specific skills are in high demand and probably not that easy to find. Whether you're hiring for one of those roles or any other role, how about you find the top talent before the competition gets to them? And how do we do it? Zip recruiter right now you could try zip recruiter for free at zip recruitercom slash twit. Zip recruiters powerful matching technology. There's a deep dive to identify top talent for your roles. Immediately after you post to your job, ziprecruiter's smart technology starts showing you qualified people for it. Gear up for summer with ZipRecruiter's high-speed hiring tools. See why four out of five employers who post on ZipRecruiter get a quality candidate within the first day. Just go to this exclusive web address right now to try ZipRecruiter for free. Ziprecruitercom slash twit. Again, that's ziprecruitercom slash T-W-I-T. Ziprecruiter the smartest way to hire.

01:13:01
You know this is from katherine bob writing at courts. It sounds credible to me, but I I am not sure. But I thought maybe I'd run it by you all, not not just our panel, but our listeners. She writes between 2022.

01:13:18
Three years ago and today, a little notice tweaked to the us tax code has quietly rewired the financial logic of how american companies invest in research and development outside of cfo and accounting circles. Almost no one knew it existed, um, so this was buried deep in the 2017 tax law under trump right. The delayed change to a decades-old tax provision has contributed to the loss of hundreds of thousands of high-paying white-collar jobs, half of more since the start of 2023 started last year, or a year and a half ago. More than half a million tech workers have been laid off. According to industry tallies, 500,000 people. What's known as Section 174, 174 of the US Tax Code has helped gut in-house software and product development teams everywhere, from Microsoft to Met meta to much smaller direct-to-consumer other internet first companies.

01:14:30
There's a bipartisan effort to repeal the 174 change. It's moving through congress right now, the first, so it's a little complicated, but I think it's simple to understand. For the last 70 years, companies could deduct 100% of the qualified research and development spending. In the same year they incurred the costs. So in the same year that they paid programmers, contractors, they could deduct that fully. 100% comes off the top of your income. The deduction was guaranteed by the tax code, section 174, the tax code of 1954, and since then rnd has flourished in the us. It's been very good for tech companies. It was repealed.

01:15:20
Yeah, congress packs the tax and cuts job act in 2017, slashing the corporate tax rate from 35 to 21. In order to make it comply. This is the. This is the interesting thing. In order to make that comply with the senate budget rules, they had to offset the cost, so they added future tax hikes. That wouldn't kick in right away, so the businesses wouldn't go hey, that could, in theory, be repealed later. So they made a change to 174 saying you had to amortize your R&D expenses over five or even 15 years. That's a big change from deducting it the year you spent it, but it wouldn't kick in until 2022. When did the firings kick in?

01:16:08
2022 yeah, uh, now there are people who are trying to fix that change, but right now it's the law. Rnd is not an immediate write-off. The tax benefits of salaries for engineers, product managers, project managers, data scientists, even some UX and marketing staff. The cost has to be spread out over five or 15-year periods. That's a big deal. Any thoughts on this? I've never heard of this.

01:16:42 - Sam Abuelsamid (Guest)
This is not unique to uh the um tech industry either. Um, over the last couple of years there have been thousands of engineers laid off from the auto industry as well. Gm um last year, last summer, laid off a thousand software engineers. You know this in an era when they are supposed to be transitioning to software-defined vehicles. You know we've had similar layoffs and buyouts of engineers from most of the automakers over the last couple of years. So I wouldn't be surprised if the same logic is applying in other industries as well.

01:17:19 - Leo Laporte (Host)
In the 2023 annual report, meta described salaries as its single biggest R&D expense. Between the first and second years that the 174 change began, meta cut its total workforce by 25%. Microsoft 7%. Twilio 22%. Shopify cut almost 30% of its staff. Coinbase 36%. This makes sense. This is because you can't deduct them anymore. Now we could argue about whether you should have been able to deduct it or not, but that's that seems to me a very credible yeah and if you're going to change it, you need to change it carefully to absorb whatever you know all the energy in the system yeah, yeah.

01:18:03 - Cathy Gellis (Guest)
Well, a secret change, but a secret change that doesn't manage what the fallout is, because now it's really disadvantaged labor. So even if you were to say you know, on balance, maybe this would be a better long-term way to do it, you probably need to, like, soften the landing in some way, and since nobody was aware of it, really it didn't soften the landing at all, and now you're losing hundreds of thousands of people from the labor economy and that's going to catch up with us in a really bad way.

01:18:28 - Leo Laporte (Host)
I'd love to hear from CFOs or somebody about what all this. If this is a sensible story, it kind of makes sense to me anyway. Well, how about this, the AI startup? It was an. So we've been talking a lot about vibe coding. In fact, we did a whole episode in our uh club on friday of uh. We have a ai user group that meets every first friday of the month on a on vibe coding. It was so much fun we're gonna do more of that.

01:18:56
I, after that, I started vibe coding a twit app. I was able to get cloud code to understand our api, connect with it, start downloading information it's kind of amazing without writing any code myself. Uh, there wasn't a startup in india that claimed to be a kind of a vibe coding builderai. Uh, turns out there was no ai in builderai. They actually had hundreds of human workers in India pretending to be chatbots. The London-based company was valued at one and a half billion dollars. They were overstating their income by significant as much as 300%. The company claimed 220 million in revenue, just made 50, and it wasn't an ai company at all.

01:19:50 - Sam Abuelsamid (Guest)
It was just 200 programmers in india it's, it's probably, you know, the vibe coding equivalent of uh, let me google that for you.

01:19:58 - Leo Laporte (Host)
Yeah, yeah, instead of instead of googling the other uh going in they might have done all right if they'd said hey look, we've got 200 great programmers in India you can use. I don't know. There are people who probably would have preferred that to an AI. I'm a little worried as a podcaster. Google's Notebook LM now has gone public. It lets you so this was actually. I know we played with it on the shows and talked about it. You could upload documents and then it would analyze those documents. You could ask a question of it. But, even more popular in the public eye you could create a podcast with two AI voices talking back and forth about the content. Well, now you can actually share those publicly and I think we're in deep trouble.

01:20:45 - Larry Magid (Guest)
Have you? Have you loaded in your bio, leo? What do you mean? Well, I loaded in my bio. I am so impressed with me myself. I had no idea I was a spectacular engineer.

01:20:58 - Leo Laporte (Host)
You know this Larry Magid guy is amazing. It is a little sycophantic.

01:21:04 - Larry Magid (Guest)
And then, just to make sure that I didn't, I you know I wasn't going too much, uh, too egotistic I I found on the web a bio of some random plumber somewhere. You know a good plumber, no question about it put his bio into that and he was amazing too.

01:21:17 - Leo Laporte (Host)
So you know, everybody's amazing I'll give you a little example I uploaded to Notebook LM. I uploaded it was only like 10 shows, the transcripts from Security Now and then it actually takes a little while to load the conversation. Let's see if I can load it in. And then I made a little podcast out of it and it kind of I I could I could see listening to this here. Listen, see if you would listen to this show welcome to the deep dive.

01:21:47 - Cathy Gellis (Guest)
We're here to unpack the latest in security and tech, pulling out the key insights for you, the learner yep, lots to cover today.

01:21:54 - Leo Laporte (Host)
We've got sources touch on everything from browser security pitfalls to how your car might be spying on you and encryption battles, hardware security failures it's quite a mix.

01:22:04 - Sam Abuelsamid (Guest)
Our goal is to make sense of it all without getting bogged down in jargon so that's it.

01:22:07 - Leo Laporte (Host)
Those are ai voices, ai podcasters, doing content directly from the security. Now transcripts uh, it's only a matter of time. You know there was an article in vanity fair about the ceo of microsoft, satya nadella. He said, oh, I love podcasts. And they asked him well, which ones do you listen to? So well, I don't know. I just I load them all into Copilot, have it, analyze it, and then I play the analysis back in my car and have interactions with it. So he doesn't listen to podcasts.

01:22:34 - Larry Magid (Guest)
The problem is, you know there's this term garbage in, garbage out. But what I discovered today is you can actually put in you know really fine cuisine and get garbage out. You can upload documents. I was telling you this off the air. I uploaded. I was concerned about my hearing and I have some minor hearing deficits. So I loaded in a bunch of audiograms into ChatGPT, one of which was created by Apple AirPods today, and it told me I had severe hearing loss at a certain frequency and I said said wow, that's kind of scary. So then I went and looked at the audiogram and it just got it wrong.

01:23:09
There was not just misinterpreted it it misinterpreted the data that I gave it. Yeah, and.

01:23:14 - Leo Laporte (Host)
I know it hallucinates.

01:23:15 - Larry Magid (Guest)
But I figure, if I'm going to feed of the document, it ought to get that right you think so.

01:23:20 - Leo Laporte (Host)
Yeah, uh, jeffrey Fowler at the Washington Post, uh, I guess he's done this before. He took all the main llms and he did a couple of things. He fed it a novel and asked it to summarize the the main points. And then, as a judge, he brought in the author of the novel. He said, okay, I'm doing this under protest. He got permission from the publishers to do this and he said some of them were good, some of them weren't good. He said, uh, they really got the epilogue downright. They got all the, all the emotion out of it. But one thing and then fowler did it with illegal documents. You'd probably be interested in that, kathy.

01:23:58
Um, the only ai of all of them that did not hallucinate at all was claude's. Ai of all of them that did not hallucinate at all was claude's anthropics. Claude everything else hallucinated would make up characters. In one case in this novel it's a civil war novel a guy who, just a soldier, just got his leg amputated, showed up on somebody's doorstep. Um, completely made up, completely fantasy. So, but maybe I don't know, try it with claude and see, because I don't.

01:24:25 - Larry Magid (Guest)
I don't understand it because I understand how, if I ask it a question, it's going to go off on the web and get all sorts of extraneous and you're going to get weird if I feed it data you shouldn't make. It shouldn't make that mistake. It should just reinterpret that data. I don't know there was.

01:24:37 - Cathy Gellis (Guest)
I forget if we talked on it about the last time I was here, but there was an incident where a lawyer got in trouble for a hallucinated citation. But it wasn't in this case, really the lawyer's fault because they didn't do legal research where the AI made up cases that didn't exist. The lawyer already knew the material that they needed to be cited and they just fed it through and I it might have been clawed, but I'm not positive, um, that they fed it through the ai for citation formatting, which should be like food, for you know you do it with a script so like why should a bot be any worse? And it didn't do it in any sort of true way. It invented and hallucinated a new source and so the lawyer filed the thing and got in trouble for it because when the sites were checked, it produced a bad result.

01:25:30
It was a, it was a hallucinated thing, but because the the AI substituted the hallucinated thing it made up and deleted the actual source that the lawyer was trying to get into the filing. So if you can't trust the AI to do basic tasks without just wandering and frolicking and detouring, as lawyers say, and coming up with all sorts of garbage, the lawyer should have supervised it. There should have been a site check and stuff to find the problem before it went through the courthouse door. But that should not have been a thing that went wrong. The AI was supposed to make the lawyer's job easier and it made the lawyer's job now terrible.

01:26:09 - Larry Magid (Guest)
You thought that Robert Kennedy healthy human services report, which was-.

01:26:13 - Leo Laporte (Host)
Yes, it was a formatting error.

01:26:14 - Larry Magid (Guest)
Formatting error. Well, maybe that's what they did, I don't know.

01:26:18 - Sam Abuelsamid (Guest)
I think one thing we should probably stop doing is using the word hallucination.

01:26:25 - Leo Laporte (Host)
I agree, because should probably stop doing is using the word hallucination.

01:26:26 - Sam Abuelsamid (Guest)
I agree, because that's anthropomorphizing it. I agree, and I. I read a good paper a couple of months ago called chat gpt is bullshit, um where yes, good paper yeah, you know that and the the context there is.

01:26:36
They're putting, they're analyzing llms in the context of harry frankfurt's on bullshit. You know where the definition of bullshit is not. You know that it's trying to mislead you. It's just that it's a complete disregard for the truth. It doesn't care and and that's what the way, fundamentally the way lms work they just don't care, so they actually don't even that care is not even a word you should use.

01:27:02 - Leo Laporte (Host)
Yeah, they, they. There's no, they're not intended to be factual, right period?

01:27:09 - Sam Abuelsamid (Guest)
yeah, it's, it's just, it's just a probability engine. Uh, what? What should the next word in this sequence there's no intent on? Yeah, there's no thinking right, and that's that's why it fits with that that, uh.

01:27:21 - Cathy Gellis (Guest)
Frankfurt definition of yeah I have mixed feelings about this. I tend to like using hallucination to describe what it is that we end up describing about it, because I think it's a term that is very specific in terms of the effect of what it's doing. But I otherwise agree with the larger point, which is that, um, you know, these are soulless things. This is like a sociopathic technology, and I worry about, in context where people do have relationships with their bots, that this is worse than dealing with a con artist, where somebody who's just feeding you a line to kind of manipulate you and you're sort of like we're human beings. We tend to presume there's a soul and a conscience on the other side and there's not. And this is where people really get hurt when the when a soulless sociopath, you know, tells them what they want to hear and doesn't care how they're misleading them. And this is in that sense I think your caution is correct, because I think we really need to break those human uh relation these social relationships that are coming up between people and their AI.

01:28:25
So I agree very much there, but I do kind of like hallucination in terms of really describing what the effect is of what's getting produced by it.

01:28:35 - Larry Magid (Guest)
As a child of the 60s, I consider that cultural appropriation it's not necessarily bad in the real world as also a child of the 60s, if you can remember it, you weren't there.

01:28:47 - Leo Laporte (Host)
No, well, maybe I was there. I was there. I remember it too, or I?

01:28:53 - Larry Magid (Guest)
was there to the extent that anybody was there. Yeah, we weren't completely there, floating over it occasionally. Yeah, mostly there.

01:29:00 - Leo Laporte (Host)
You experienced it firsthand. I agree with you, Kathy. Actually, that's a good point, Because I don't. In general, I try to discourage anthropomorphizing AIs because it kind of leads to a misunderstanding of what's going on. But you're right, From the point of view of the user it is kind of like that.

01:29:17 - Cathy Gellis (Guest)
It is a hallucination. Yeah, it focuses on the harm, yeah, the problem with it, is it often?

01:29:22 - Larry Magid (Guest)
it always sounds very reasonable because it is. You know, that's his job and so it's very believable. And you know, as I've always said, if it's an important decision you need to vet it. But on the other hand, you know it's a great way to start a research project. I use it all the time.

01:29:39 - Leo Laporte (Host)
Perplexity just added a deep research thing which takes 10 minutes. Before I went to the doctor I gave it all my medical information and had questions and it really gave me a great research paper that I was able to go into. The doctor kind of prepared and I think there are.

01:29:57 - Larry Magid (Guest)
I mean, look, I use Now the New York Times. Know your entire medical history?

01:30:00 - Leo Laporte (Host)
Yeah, I know I use AI all the time, for everything. You just have to be judicious in it, that's all I think to sam's point.

01:30:07 - Cathy Gellis (Guest)
You know, maybe the I don't know what would we say instead? Because this is the technological equivalent of pulled out of its technological ass and and like, but that's too many words and you still have sort of a personification thing. It's like something random has been invented it's incorrect, but it's incorrect in a way that I think we have to convey why it is incorrect.

01:30:29 - Larry Magid (Guest)
It's invented it's a little bit like full self-driving, and the reason I say that is that when full self-driving first came out, it was a joke and I just paid no attention to it.

01:30:39 - Leo Laporte (Host)
I you have a tesla, you drive it's very good it's.

01:30:43 - Larry Magid (Guest)
It's far from perfect. It's only going to get you killed one out of a few hundred times you do keep your hands on the wheel well, or at least your eyes on the road. But my point is that chat gpt is the same thing. It's gotten so good that most of the time you can rely on it.

01:30:57 - Sam Abuelsamid (Guest)
But you can't rely on it, and that's the part, and that's in some ways the most dangerous yeah, you're getting into that uncanny valley where, if you get to the point where you're relying on it, you are very likely going to miss the times when you shouldn't be relying on it Exactly Right, right.

01:31:14 - Larry Magid (Guest)
And that's the problem with full self-driving If you're not paying attention, it can kill you. It's the problem with GPS. If you are, it's actually very driver-resistant.

01:31:20 - Leo Laporte (Host)
I don't know how to drive anymore, because if the GPS doesn't tell me to turn left, I'm just going to, and so we've all become more or less dependent on these automated tools, and that is something to be aware of Instead of robots taking over the world.

01:31:33 - Larry Magid (Guest)
They're turning us into robots. Is what you're saying. I saw WALL-E. They're turning us stupid.

01:31:38 - Leo Laporte (Host)
We're going to be driving around, slurping our slurpees in our little hover cars, watching, watching days of our lives and we used to have a tesla and of course the fsd wasn't as good as it is now.

01:31:48 - Larry Magid (Guest)
But I have no, I have to time. I, I, I I'm kind of watching to make sure nobody I'm not going to crash into anything, but I actually have no idea where I'm going. I use, I have a bmw i5.

01:31:57 - Leo Laporte (Host)
They have a similar adapt system that I, or a desk system. That I think is very good. It's highways only, right, sam? I do like that, like Blue Cruise and these other technologies. They don't try to do it on city streets, that's really challenging for FSD, but it does just fine on the highway.

01:32:16
It has a feature that's a little creepy. It will pop up a thing saying you know you could be going faster. If you check your side mirrors, I will change change lanes for you definitely didn't even ask you to it just does it, it just does.

01:32:29 - Cathy Gellis (Guest)
Yeah, I don't even like a number of systems now yeah, I drove a car with like lane policing and it made me really uncomfortable oh, I like that.

01:32:38 - Leo Laporte (Host)
I live with that. Now you're right because we, we, I think, as humans you're right, because we, we, I think, as humans we tend to drive perhaps a little closer to one of the lines than the other. Yeah, it's going right down the middle, and that often, especially if there's a truck on on the left. I want to go a little bit to the right I don't want to.

01:32:56 - Cathy Gellis (Guest)
Actually, that's my reasons for being over here. There were important reasons.

01:33:01 - Sam Abuelsamid (Guest)
Stop yelling at me actually a lot of these systems now do that, like for blue cruise, um, just super cruise, um, you know, when it detects that it's a truck, you're passing a large truck, it'll go, it will, it will nudge you over, it'll stay within the lane but it'll nudge you over towards the opposite side of the lane, um, and that that is helpful I didn't.

01:33:22 - Cathy Gellis (Guest)
I didn't even like being yelled at, but I definitely don't want to lose control of the car where I'm fighting with the car over it. I don't want something breaking for me. I don't want something steering for me.

01:33:33 - Sam Abuelsamid (Guest)
Yeah, the Tesla systems sometimes do fight back against driver intervention.

01:33:40 - Leo Laporte (Host)
Not my BMW. I can grab the wheel and take it out.

01:33:43 - Sam Abuelsamid (Guest)
Yeah, intervention, not my BMW. I can grab the wheel. Other systems yeah yeah, the other systems generally. You know, if you give it any force at all on the steering wheel or if you hit the brake.

01:33:52 - Larry Magid (Guest)
I've never had Tesla prevent me from taking control. I mean either by hitting the brake, touching the problem is it's, it's.

01:33:59 - Sam Abuelsamid (Guest)
The system is very inconsistent. Most of the time it doesn't, but sometimes, every once in a while, as, but sometimes every once in a while, as you said, every once in a while it will do something very, very wrong. And and it and it's not, it, you know again, partly because it's so overwhelmingly controlled by AI is that it? It doesn't behave the same. You know, if it was a deterministic system, then it would tend to behave the same way under the same circumstances all the time, but it doesn't, and you can be in the same circumstances. There's been a lot of testing that's independent testing that's been done that shows going down the same road on same time of day, same conditions. Uh, you know, it will do totally different things, which is not what you want but for the human beings right, but yeah, but if you know, yes, human beings do it.

01:34:51
But if you're going to replace the human with the technology, the technology should be better better in a related, otherwise there's no point.

01:34:58 - Leo Laporte (Host)
Yeah, the courts in california have decided that it is illegal for you to hold your cell phone while driving, even if you're looking at the map and that's probably a good thing. I think it's a good thing I'd agree with it.

01:35:12 - Sam Abuelsamid (Guest)
I mean, wasn't it already illegal to hold?

01:35:14 - Leo Laporte (Host)
your phone at all, prohibits drivers from texting or talking on a cell phone while driving I thought you're not even allowed to hold the phone while you're driving uh, yeah, it's also. Let's see the court. Yeah, I think the court upheld the law basically. Okay, yeah, it's illegal to hold a phone. Yes, as I remember when they passed the law that everybody said don't even touch the phone, because the chp will give you a ticket if you touch the you have to touch it sometimes for gps unless you get well three yeah, um, that's a good question, like if it's in a, if it's in a holder, you're not holding it.

01:35:49
Uh, legislatures. The legislator adopted the current state law prohibiting drivers from quote operating a cell phone while driving, whatever operating operating.

01:35:58 - Larry Magid (Guest)
What if you're doing hand three?

01:35:59 - Cathy Gellis (Guest)
that's okay yeah, what is operating, and that's there's probably some case law that establishes what operating means exactly.

01:36:05 - Leo Laporte (Host)
Yeah, that's okay, yeah, what is operating and that's. There's probably some case law that establishes what operating means exactly.

01:36:07 - Cathy Gellis (Guest)
Yeah, that's because that's going to come up with? Um, yeah, I would look in like dui. Uh, litigation, somebody's probably sued for. What does it mean to operate?

01:36:15 - Leo Laporte (Host)
well, this is what happened. A guy was ticketed after looking at a mapping application while holding his phone in his left hand and driving. He contested the ticket, lost his initial court appearance, ordered to pay 158 fine. But he said I'm right. So he went to the appellate division of the santa clara county superior court. The decision was reversed. The superior court said operating a cell phone required active use or manipulation of the device. Merely observing GPS directions does not infract. But California Court of Appeal for the Sixth Appellate District disagreed and on Tuesday reversed the decision. So you're right, kathy.

01:36:56 - Larry Magid (Guest)
That's because Android and iOS are both built in Santa Clara County. We want people to use our phones all the time.

01:37:02 - Leo Laporte (Host)
I mean, I have GPS. Look, every car now has GPS screens, right, which can also be distracting. I'm looking at. Yeah, definitely, I'm looking at it. Sometimes I have a. This is another thing I like about that BMW has a heads-up display that will show me the instructions, so I don't have to look away from that.

01:37:18 - Sam Abuelsamid (Guest)
Yeah, the uh, the new Cadillac Vistic has an augmented reality heads-up display, so it's so. It's a multi-plane display, so normally your HUD. If you look at your heads-up display, it's projecting the data to be roughly about six or seven feet away from you, but roughly around where the front bumper is. This new one projects information in multiple planes, so when you're using navigation, for example, as you're approaching a turn, it'll show a big arrow float hovering over the intersection where you're supposed to turn.

01:37:56
Yeah, and as you get closer it appears to get closer to you as you approach it.

01:37:59 - Larry Magid (Guest)
It doesn't block your view of the intersection, though.

01:38:01 - Sam Abuelsamid (Guest)
No, it's translucent, so you can see through and around it.

01:38:05 - Leo Laporte (Host)
So it's not the new apple ui they're going to announce tomorrow. Yeah, all right, I'm going to take a break. We will talk about that. We will talk about volvo's iot seat belts I don't know what that means, but sam will explain it and, of course, the fastest selling game device in its first day ever. All of that still to come.

01:38:25
You're watching this week in tech as we talk about the week's tech news, our show today, brought to you by threat locker man. You don't have to listen to the security now to know the ransomware is killing businesses worldwide. They get in all sorts of ways phishing emails, infected downloads, malicious websites, rdp exploits. Look, this is the last thing you want to have happen to your business. That's why you need threat lockers. Zero trust platform and zero trust is the key. It takes a proactive listen to these three words deny by default approach every action is blocked unless it's explicitly authorized protects you from known and unknown threats because they, they just. It's like hitting a brick wall to them. It is exactly the kind of protection a city government, a school district needs. Big businesses, uh, infrastructure companies like jet blue. Jet JetBlue uses ThreatLocker because they just they can't, you know, they can't let their business get shut down. The Port of Vancouver, talk about infrastructure. They use ThreatLocker to shield themselves from zero-day exploits and supply chain attacks. And very nice, providing complete audit trails for compliance. You know exactly what you've approved, what you haven't approved, who used nice, providing complete audit trails for compliance. You know exactly what you've approved, what you haven't approved, who used what when, where.

01:39:49
Threatlocker's innovative ring fencing technology isolates those critical applications from weaponization. It stops ransomware cold. It also limits lateral movement within your network. See, that's part of the problem. Bad guys get into your network. On average, it takes 91 days for a company to realize they've been breached, so the bad guys at that time can wander around at will, go and look at anything, run applications. We had a story on security now about a ransomware team that got in to a company's network. The company had some sort of protection software, so they couldn't do everything they wanted, but then they were able to move around. That's what you want to stop that lateral movement. And they found a camera, a security camera that was running as many do linux and had a little bit of memory and a little bit of a processor, but it was enough to inject their malware into it and to ransomware the entire operation using the camera. You need threat locker to stop that lateral movement within your network.

01:40:50
Threatlocker works across all industries. It supports Mac environments as well as Windows, provides 24-7 good quality, us-based support and enables comprehensive visibility and control. I'll give you another example. City governments are bombarded by ransomware all the time. Mark Tolson is the IT director for the city of Champaign, illinois. He said enough is enough. He installed ThreatLocker. This is what he says about ThreatLocker, and I'm quoting. Threatlocker provides that extra key to block anomalies that nothing else can do. If bad actors got in and tried to execute something, I take comfort knowing ThreatLocker will stop that. Stop worrying about cyber threats. Get unprecedented protection quickly, easily and cost-effectively with ThreatLocker. Visit ThreatLockercom slash twit. Get a free 30-day trial. Learn more about how ThreatLocker can help mitigate unknown threats and ensure compliance. That's nice, threatlockercom slash twit. You will be amazed at how affordable it is. You can really protect yourself with threat locker. Thank you, threat locker, for uh supporting twit. What are iot seat belts, sam?

01:42:01 - Sam Abuelsamid (Guest)
so iot is probably a bit of a misnomer here um, basically, they're um smarter seat belts. So what? What volvo is doing here with these new seatbelts that are coming next year on the new EX60, is they're taking advantage of the various sensors they have on the car and trying to better judge the situation both outside the vehicle and inside the vehicle. Both outside the vehicle and inside the vehicle. For some time now, most seatbelts in modern cars have already had some smarts built into them so that they adjust depending on the impact force and various other things.

01:42:44 - Leo Laporte (Host)
That's almost mechanical, though right, like if it's leaned forward.

01:42:48 - Sam Abuelsamid (Guest)
Well, originally it was, but now it's not. Oh so, so seatbelts, modern seatbelts in cars for the last 20 ish years have all had pretensioners on them. So when there's an impact, the original, the first ones, actually used a pyrotechnic device to retract the seatbelt to pull you back tight you know, so you didn't hit the steering wheel Right, and that was a good thing.

01:43:18 - Leo Laporte (Host)
Might have broken some bones, but at least you're alive.

01:43:22 - Sam Abuelsamid (Guest)
Yeah, I mean it wasn't that powerful, or at least it wasn't intended to be. Sometimes it could, but it was better than the alternative. But it was kind of a one-shot thing and there was no adjustment for different body sizes where you were sitting relative to the steering wheel. So depending on how tall you are, if you're shorter you're going to be sitting closer to the steering wheel. It didn't have adjustments for that.

01:43:50
So what they're doing now with these new systems is they're taking advantage of looking at where's your seat positioned, looking at using the driver monitor system, because a lot of modern cars now have an infrared camera, like your BMW does. That is looking at you to make you know, to check that you're watching the road and you know hands on the steering wheel, various other things. But also looking outside, looking at what is the direction of the impact. For example, you know what you know, you know it knows. Okay, I'm going to have you know what's called a small offset impact. You know, if somebody crosses the center line and you have a glancing blow, that's a different impact than a direct head on impact, and so it responds differently based on all of these different factors and it adjusts the amount of pre-tensioning, um, you know, based on on your uh, your, your, your size, your body type, um, and the type of impact that's going on wow, that's kind of cool yeah, this is a video from volvo.

01:45:03 - Leo Laporte (Host)
Yeah, and you know, back in the 1950s oh really

01:45:08 - Cathy Gellis (Guest)
yeah they, they create, they were the first shoulder, shoulder plus uh lap yeah didn't it just involve a physical mechanism involving some pendulum type action that when the money that was, that was the.

01:45:21 - Sam Abuelsamid (Guest)
That was the early thing, yeah, it was a pendulum type action, uh, but now you know. And then they went to a pyrotechnic system, uh, and now they're going to an electromechanical system that can adjust the amount of tensioning force on the seatbelt automatically, based on all these other factors.

01:45:38 - Leo Laporte (Host)
This is an example of how everything, now you know, there's cameras, there's sensors, there's radar. There's so much information now, yeah, and it makes sense for the auto to use it. I mean, it's expensive, though, doesn't it add greatly to the cost?

01:45:54 - Sam Abuelsamid (Guest)
um it it does. I mean we're putting these sensors on the car anyway, so what? What we're seeing here.

01:46:01 - Bill Atkinson (Interviewee)
It's just you know, we're finding new ways to use the sensors okay, um, you know okay.

01:46:05 - Sam Abuelsamid (Guest)
So we've got, we put these sensors on the car. What else can we do with them to enhance safety? And so they're using it in a bunch of different ways. Another example for Volvo is what they call their driver understanding system. So, leo, you remember back in 2017, I think, when I first took you for a ride in the Cadillac CT6 with Super Cruise, yeah, and it has that infrared camera on the steering column I was looking at my eyes.

01:46:34 - Leo Laporte (Host)
I had to cover my eyes to get it to warn me.

01:46:37 - Sam Abuelsamid (Guest)
Yeah, and that was a fairly simple system that all it did was look at your eye, gaze and head position Right, just to make sure you know, are you?

01:46:45 - Bill Atkinson (Interviewee)
watching the road. Yeah, yeah.

01:46:47 - Sam Abuelsamid (Guest)
And if you were looking, if you look away from the road for more than a few seconds, it would alert you and it would start to if you didn't respond, it would start to disengage Super Cruise. Well, what?

01:46:56 - Leo Laporte (Host)
Volvo is doing. It also shook the seat.

01:46:59 - Sam Abuelsamid (Guest)
Yeah, yeah, it was a haptic seat too.

01:47:01 - Leo Laporte (Host)
It was shaking you to wake up.

01:47:03 - Sam Abuelsamid (Guest)
Yeah, you know, what Volvo is doing now is what they call driver understanding system. So they've done a lot of research, you know, because one of the things that happens when you use these systems like Super Cruise or FSD, you know where you're supposed to be watching the road. If you're on a long road trip, what can often happen is your eyes might be on the road but you're not actually attentive, your brain's not. And this is one of the problems is, you know, you start to get that thousand-yard stare, so your eyes are pointed towards the road, but you're not ready to take control if you need to. And so what Volvo?

01:47:43
You know they've got some anthropologists and behavioral scientists that have been looking at this and trying to model human behavior to better understand, not just on where your eyes are looking at any moment in time, but also looking for patterns of. You know, are you occasionally glancing over at the mirrors, at the screen? You know? You know, are there other signs that you are actually alert rather than just completely zoned out? Your eyes may be open, but you're not you're?

01:48:11 - Leo Laporte (Host)
you're not, how would it know your pupils are dilating or it's looking for a variety of factors, you know.

01:48:19 - Sam Abuelsamid (Guest)
Like I say, it's looking for eye motion this is, this is ai, isn't it?

01:48:23 - Leo Laporte (Host)
this is ai.

01:48:24 - Larry Magid (Guest)
Part of part of it is ai, yes, that can happen even without driver assistance. I remember, you know, passing an exit on the highway because I was just sort of spaced out listening to music, thinking about other things All the time.

01:48:34 - Sam Abuelsamid (Guest)
Yeah, and there's an increasing number of vehicles that, like a lot of Toyotas, a lot of Hyundai and Kia vehicles now, even though they don't have a hands-free system like Super Cruise or Blue Cruise or FSD, they do have that driver monitor camera that's just looking for distraction, just trying to make sure that you are paying attention to the road, and if you're not, then it'll start to provide some alerts and try to get your attention.

01:49:06 - Larry Magid (Guest)
My 2016 Prius yells at me if it thinks that I'm not paying attention, but it gets it wrong, like if I'm on a certain road in yosemite that's highly curvy we'll constantly yell at it.

01:49:17 - Leo Laporte (Host)
Yeah, the systems were very crude bmw yells at me. If I pick my nose, I swear to god it and I I apologize, it didn't, yeah, yeah. It was a scratch. It was a scratch. It yells at me because I guess it. I don't know. It's really interesting, my wife yells at me if I do that?

01:49:37 - Larry Magid (Guest)
Yeah, it should.

01:49:39 - Sam Abuelsamid (Guest)
I've seen some recent research that picking your nose can lead to dementia. Yeah.

01:49:43 - Leo Laporte (Host)
I know, don't go too far up there. That's basically the so. Okay. So are cars safer. They're safer we know they're safer.

01:49:52 - Sam Abuelsamid (Guest)
They are absolutely safer for occupants than they ever have been.

01:49:55 - Leo Laporte (Host)
Are deaths, are there? Oh, not so good for pedestrians, they're huge.

01:49:57 - Cathy Gellis (Guest)
They run down all the people outside them.

01:50:01 - Leo Laporte (Host)
Yeah, if you look at the data, I walk by those trucks all the time where the hood is higher than my head.

01:50:15 - Sam Abuelsamid (Guest)
If you look at the data from the last decade, since the 1960s we've been on a steady downward trend for traffic fatality rates, except until the last decade. And what's happened is, in the last decade the occupant fatalities have continued to decline, but pedestrian and cyclist fatalities have increased because everybody's getting into these giant trucks and SUVs and they can't see people and and you've got, you know, worse visibility and, um, it's, it's really. The cars are bigger and heavier too.

01:50:39 - Cathy Gellis (Guest)
So if there's contact, the contact is more likely to be destructive. Um, I I saw a headline the other day of I don't remember where some girl got hit by a car on her bike and she died, and that hit me like emotionally, because I got hit by a car on my bike when I was 14 years old, but it was one of those like big 1980s buicks, so when I couldn't avoid the car I went on its hood and I rolled off on the other side and didn't get run over or bounced too hard or something.

01:51:10
So, um, I mean, I didn't see that in the car, didn't see me but now I wouldn't, I would just end up under the car, and then I would have no chance yeah, yeah, now you get.

01:51:19 - Sam Abuelsamid (Guest)
You know people getting hit by cyber trucks and raptors and you know, and I know, people who are buying suvs just for defensive reasons, because they just right everybody yeah, well, it's protecting the people on the inside, but it's killing people on the outside what do you make of the?

01:51:33 - Larry Magid (Guest)
you know the, the kind of fighting over whether tesla is safer than other cars or less safe. I've seen data suggesting it gets into more accidents and why is that?

01:51:42 - Sam Abuelsamid (Guest)
you know the the problem.

01:51:43 - Leo Laporte (Host)
The problem is, I'll tell you why larry, it's people like you, it's tesla drivers that's the problem.

01:51:49 - Larry Magid (Guest)
I'm actually listening.

01:51:50 - Cathy Gellis (Guest)
Yeah, I was embarrassed to drive a bmw for a long time because bmw I've noticed that issue with, like you know, driving with teslas on the road and, with all due respect to my friends who drive teslas, I always noticed it was something kind of absent about the way it was getting driven. It almost seemed like the drivers had like just turned over a little bit to too much control to the, to the, to the car and just sort of expected the car will make it fine.

01:52:22
It's like there's other people on the road. We're busy trying to like merge with you.

01:52:26 - Larry Magid (Guest)
We need you know, you know, we need you to be on the ball and I'm curious that sam knows why it is that the, the you know what, what's behind this, these reports so so we've you know the.

01:52:36 - Sam Abuelsamid (Guest)
The problem is trying to get valid data to get you know, apples to apples data, because one you know one of the things uh, you know tesla has often put out reports saying our cars get into, you know, one quarter as many many crashes as other cars. But what they don't tell you is they're you know. It's the classic thing of you know. Tell me which side of the argument you're on and I'll give you the statistics to prove you're right In this case, you know Tesla.

01:53:01
We think we don't know for sure because they don't actually provide any context.

01:53:05
They just give a headline number and nothing else.

01:53:15
But most likely, what they're doing is they are comparing the rate of crashes of their vehicles to the entire vehicle population, and there's 290 million registered vehicles in the United States with an average age of almost 13 years now, and most of those don't have advanced ADAS systems and other safety features.

01:53:27
You know, if you compared Teslas, which 90% of all the Teslas have been built since 2017, so the last eight years, or actually probably more than 90%, and they all have these ADAS features If you compared them only to the population of vehicles built in the same time period and especially more expensive premium vehicles that tend to have these types of features. The numbers would be very, very different. Interesting Because a lot of the crashes are happening with older vehicles, and when studies have been done trying to make those comparisons like-for-like comparisons they have found that Teslas tend to get into more crashes. They have found that Teslas tend to get into more crashes, and while it's hard to determine the exact causation, there is certainly a correlation with use of features like autopilot and FSD, where the drivers are not as attentive, and Tesla drivers in general, because they tend to be more tech forward, tend to be more trusting of the technology, and so they're more likely to utilize it, perhaps beyond what its true capabilities would justify.

01:54:39 - Cathy Gellis (Guest)
I just always felt like it was an abdication, like it's not necessarily that they were using the active features, but that they were so convinced about the car being the better driver and that the car will take care of them that they stopped being, as is, involved with what their role was with the cars operation about 98 of the time.

01:54:57 - Sam Abuelsamid (Guest)
But it doesn't take, doesn't take much to kill you well and, and the thing is, the vast majority of driving we don't have crashes. You know and right. This is one of the one of the one of the problems around automotive safety has been the way that it's been talked about for so long. Everybody likes to trot out this. You know, 94% of crashes are caused by human error, which is true up to a point. 94% of crashes human error is the last in a chain of events that led to that crash, chain of events that led to that crash. What you don't hear is that you know we drive 3.2, 3.3 trillion miles a year in the United States alone. That's trillion with a T, and we have about one fatal crash for every 100 million miles of driving. For all crashes, it's about six and a half million crashes a year, which works out to roughly about one crash every half a million miles for average drivers driving 15,000 miles a year.

01:55:55
That's one crash every 30 years on average, so we as much as we humans are often the last in a chain of events that lead to a crash. We actually are really good at this incredibly complex task. Driving is very hard, but we are actually much better at it than we give ourselves credit for is it harder than hitting a baseball?

01:56:22 - Leo Laporte (Host)
um, yeah, I would say it is okay you know, because, because there's so many more variables, that's a a silly question.

01:56:29 - Larry Magid (Guest)
By the way, my homemade bumper sticker says I knew he was crazy. I didn't know he was dangerous. Oh, I like it. I've known he was crazy for a long time. I think I'm going to send that to Donald Trump. He could put it on the back of his Tesla.

01:56:41 - Sam Abuelsamid (Guest)
Well, he's getting rid of that though.

01:56:43 - Larry Magid (Guest)
Yeah, apparently he's going to sell it.

01:56:45 - Leo Laporte (Host)
Is he selling his Tesla?

01:56:50 - Larry Magid (Guest)
is he selling his tesla? Apparently he's gonna sell it, yeah, I I'd love to own that tesla.

01:56:52 - Leo Laporte (Host)
I think that'd be a great collector's item. This is this is a great. First of all, I doubt very much donald trump has ever driven a car in his life, right, right or walked in the grass with bare feet.

01:57:01 - Benito (Announcement)
I've decided that that would explain a lot hi this is benito before we move on from cars, I want to ask sam something about modern cars right now. Why are the headlights getting brighter and brighter?

01:57:14 - Sam Abuelsamid (Guest)
To try to better illuminate the scene around you.

01:57:17 - Cathy Gellis (Guest)
Not to blind me everywhere I go oh the better to see you with my new, the ones that see me from behind.

01:57:23 - Benito (Announcement)
they're blinding me into my mirrors. The ones that are coming towards me are blinding me ahead of me.

01:57:27 - Sam Abuelsamid (Guest)
It's like I feel like it's more dangerous.

01:57:30
Yeah, lighting technology the ones that are coming towards me are blinding me ahead of me, it's like I feel like it's more dangerous. Yeah, I mean, lighting technology has changed a lot. You know. We've gone from incandescence to halogens, you know, and now we're using LED lights, headlights and they are much brighter. One of the problems that has happened, particularly in the US, is our lighting regulations did not keep up with it For a long time.

01:57:53
In Europe they've allowed adaptive lighting systems, so the lights used on a lot of European cars. If you're familiar with DLP projectors, they work a lot like that, where you've got the lighting elements and there's basically mirrors that create the lighting pattern and it can adjust in real time. You know, using the cameras and other sensors on the vehicle, it can adjust the lighting pattern so as you're driving as a car's coming up towards you or you're coming up behind another car, it will automatically lower the beams, you know, so that you're not blinding other cars. They only recently changed that regulation in the US and so far the first manufacturer to actually implement adaptive headlights like that in the US market is Rivian. They have the mechanism, they have not.

01:58:52 - Leo Laporte (Host)
I don't believe they've actually implemented it very recently I think that's a funny thing, because I think my bmw also has it, but it's not a lot, it's not on yeah, and a lot of european cars have the hardware, but they haven't enabled it in software yet yeah like the last time I was driving a rivian a few months back.

01:59:10 - Sam Abuelsamid (Guest)
I was driving at night and as a car was approaching me I could see. The high beams were on, it was illuminating the entire scene in front of me and all of a sudden it was basically a notch in the upper left-hand corner.

01:59:23 - Leo Laporte (Host)
That's so cool.

01:59:24 - Sam Abuelsamid (Guest)
You could see the lights go down where that car was, so it wasn't blinding the car't just so you wouldn't find the driver, but you could still see everything going on.

01:59:32 - Leo Laporte (Host)
We're gonna. We have to take a break. I have lots more stuff. We'll do a speed round because we're almost done, but there's so many big stories this week. Uh, great to have you, sam, a bull salmon car guy. Uh, what's the best self-driving vehicle on the market today?

01:59:46 - Sam Abuelsamid (Guest)
uh, there is no self-driving vehicle on the market, okay thank you you cannot. You cannot buy an actual car that will drive itself without human supervision. What's the best?

01:59:55 - Leo Laporte (Host)
next best thing? Uh, gm supercruise system. Okay, there you have the answer. See, that's. I like a definitive answer like that. Thank you, sam kathy gallus is also here. She talks about the law, so obviously there's no definitive answer, are?

02:00:09 - Larry Magid (Guest)
you going to ask me? What is the best law. What's the best?

02:00:12 - Cathy Gellis (Guest)
law in the world, section 230. Yes, I agree. Section 230.

02:00:18 - Leo Laporte (Host)
By the way, we're trying to get Mike Mazdik on. You write for Tech Dirt and I love Mike and love your writing. He has vibe coded his own to-do list app and and he's like thrilled with it, so I want to get him on a little bit to talk about it. I think we have to schedule him uh at a time because he's a busy, busy guy, but uh, very cool. What he said done. Anyway, it's great to talk to you, kathy, and, of course, larry maggot from connect safelyorg, this week's show, brought to you by NetSuite. Hello, netsuite.

02:00:53
It's an interesting time for business. Tariffs and trade policies are dynamic. How about that? Supply chains are squeezed and cash flow, yeah, tighter than ever.

02:01:06
If your business cannot adapt in real time, you are in a world of hurt. Now more than ever, you need total visibility, from global shipments to tariff impacts, to real-time cash flow. How do you get it? Netsuite by oracle, your ai powered business management suite trusted by over 41 000 businesses. Netsuite is the number one cloud ERP and for a lot of reasons, it brings accounting, financial management, inventory and HR all into one suite. That means you have one source of truth giving you the visibility and control you need to make quick decisions. With real-time forecasting, you're peering into the future. With actionable data and with AI embedded throughout, you can automate a lot of these everyday tasks, letting your teams stay strategic. Netsuite helps you know what's stuck, what it's costing you and how to pivot fast. Don't you need that? It's one system, full control. Tame the chaos with NetSuite if your revenues are at least in the seven figures. Download the free ebook navigating global trade three insights for leaders at netsuitecom slash twit. That's netsuitecom slash twit. N-e-t-s-u-i-t-e dot com slash twit. We thank NetSuite so much for supporting our network. We thank you especially, uh, for supporting our network as well.

02:02:33
You club twit listeners. Uh, we have raised the price of club to it first time in four years. We felt we kind of had to, what with tariffs and all. Uh, no, tariffs don't impact us yet, but they impact a lot of our advertisers, and so we, uh, we thought you know, 10 bucks is still a pretty good deal.

02:02:50
You get all the shows plus all the content we do in the club, all of it ad-free. You get access to the Club Twit Discord. The keynotes are all in the club. Now Tomorrow we're going to do the WWDC keynote. Micah Sargent and I will start 10 am Pacific in the club. So you have to be in the club to watch our coverage and in the Club Twit Discord. The good news is that means the discord can also participate in our coverage. So we'll have your thoughts and your chat and your takes and all this as we go. Uh, and because we're doing it in the club, we're also going to bring our lunch, have a, have a round table conversation in the break and then cover the other keynote which we never covered before the state of the union keynote. That follows at 1 o'clock Pacific tomorrow afternoon. So a full day of Apple coverage for club members, if you're not a member, make this.

02:03:38 - Larry Magid (Guest)
You're not importing podcasts from China. I just want to be clear about that. No, we are not. No tariffs, save some money, you know.

02:03:45 - Leo Laporte (Host)
You know how would you. I don't know how you would do. Tariffs on intellectual property.

02:03:48 - Larry Magid (Guest)
That's a really good question.

02:04:01 - Cathy Gellis (Guest)
Actually, that's a major issue that trump is not thinking about and should be. Twittv, slash, club, twit, we will not charge you a tariff.

02:04:04 - Leo Laporte (Host)
Actually, I think the bigger issue is, I think he did try to do something along those lines.

02:04:06 - Cathy Gellis (Guest)
And movies, right, yeah, with movies. And there is text that says, no, you don't get to do it with that type of material.

02:04:11 - Leo Laporte (Host)
But it was that got dropped like a hot potato.

02:04:13 - Cathy Gellis (Guest)
It was and I think it had to be. I mean, there's degrees of illegality of various things that are attempted, but some of them are running. How would you, even into some very explicit language, saying no, you can't do that to this?

02:04:23 - Leo Laporte (Host)
and how would you, even logistically, could you tariff movies?

02:04:28 - Cathy Gellis (Guest)
well, and it was also a really bad idea. That would have really hurt hollywood in a very direct way, where, like china would have been like okay, we'll not use your stuff, and it was just. It was sure deal.

02:04:39 - Leo Laporte (Host)
There were a lot of issues yeah, uh, so far no tariffs on pockets, although I have to say it is very confusing that that oracle ad it really underscores it. Lisa wants to order some jewelry from france. She said, well, what's the tariff on france? I said I think it's 25. I actually had to go look it up. It was 25, but it's temporarily paused until later in july, so it's 10 right now, but it's good do they query a database, or trump's brain, or how do they?

02:05:12 - Larry Magid (Guest)
I don't know how I asked?

02:05:13 - Leo Laporte (Host)
you know, I'll be honest, I asked perplexity. I hope it was right perplexity, yeah, hey, here's one area of confusion that might be cleared up thanks to microsoft usbc cables, for, crying out loud, you never know what that cable is capable of, what that port is for. Microsoft, at least, is trying to do their bit. They have something called the Windows hardware. Um, what is it? Whcp? Uh, I can't remember what it stands for, but if you're a PC manufacturer, you want to be whcp approved and Microsoft says, okay, if you that you're going to have to label your ports, they're going to have to work in a certain way and you will have to advertise it in a certain way. Here is the USB-C WHCP requirements, this chart itself. It's confusing, but at least you'll know you're going get this as a minimum.

02:06:14
Okay, uh, and I? It's amazing that this is the windows hardware compatibility program. There it is. It's amazing that we haven't had anything like this until now. There's just it's there's no consistency at all. It's been driving people crazy. It's nuts. Uh, continuing the speed round. Walmart is going to expand its drone delivery to 100 stores. Get ready, you will see walmart drones making deliveries in atlanta, charlotte, houston, orlando, tampa. They're already doing it in dallas and bentonville, arkansas.

02:06:46 - Larry Magid (Guest)
Arkansas, the home, the hometown of walmart well, they land in your backyard, or what do they?

02:06:55 - Leo Laporte (Host)
walmart. Well, they land in your backyard, or what do they? What do they? I think they drop it by a cable. They're working, uh, with the drone startup zipline. That kind of tells you what they're going to do. Uh, this amazon is is really only doing it two sites, so they and in fact, they paused drone deliveries for two months.

02:07:19 - Larry Magid (Guest)
Uh citing, uh uh safety issues and the lindsay is doing it as well, but not not with uh presence, with there.

02:07:22 - Leo Laporte (Host)
Well, it's being used to good effect in disaster areas to get medicines to people. Uh, that makes a lot of sense.

02:07:29 - Bill Atkinson (Interviewee)
I don't know uh, walmart deliveries?

02:07:32 - Leo Laporte (Host)
I don't know, we'll see. I don't know.

02:07:36 - Larry Magid (Guest)
Walmart deliveries? I don't know, we'll see. Get ready, look up in the sky. I live near an airport, so I don't even think they can legally fly in my neighborhood.

02:07:41 - Leo Laporte (Host)
Yeah, you might not be able to. You might have to actually get in the car and drive down to your local Walmart. If you do, would you pick up a Nintendo Switch for me? Sure, sure, it is hard to find them. Uh, walmart has them. Target, has them? Uh, the switch 2 has reportedly beaten the record for the most sold out console within 24 hours it came out, uh, on thursday. It's on track to shatter the two-month record three million units on launch day. The most successful hardware, one of the most successful hardware, one of the most successful hardware releases of all times has Microsoft sold 3 million Xbox Ones, yet.

02:08:20
Don't. No, you're being mean now You're just being mean Buzzy. The Buzzy launch, according to Barron's or no. Bloomberg drew long lines at retailers like Walmart, target, best Buy and GameStop. Long lines at retailers like walmart, target, best buy and gamestop. Here. Usa today has some pictures of people sitting. This is like the iphone when the iphone came out.

02:08:43 - Benito (Announcement)
These are people outside a best buy store waiting for midnight so they could buy the switch too yeah, there are actually a lot of people on my feeds this weekend um being like, wow, it feels like 2010 again. I, I, they were like excited to wait out in line for a midnight launch. It was weird did you benito?

02:08:59 - Leo Laporte (Host)
did you wait? Did you wait?

02:09:01 - Sam Abuelsamid (Guest)
I did not, I did not know on the gadget podcast this week davindra was talking about he went to midnight launch at his local best buy, I think, or somewhere good for davindra yeah, he couldn't get a pre-order.

02:09:12 - Leo Laporte (Host)
So he went and went, went to the midnight line this reminds me so much of 2007, when people, would you know, come out of the store. I got it, I got one I was.

02:09:21 - Larry Magid (Guest)
I was covering that and it was. It was phenomenal watching that there was also the.

02:09:25 - Benito (Announcement)
The nintendo store had just opened downtown recently, so there was something down there too there too okay, look, he's got one musicians playing

02:09:44 - Bill Atkinson (Interviewee)
at a f-nac darty store.

02:09:45 - Leo Laporte (Host)
I don't know, what f-nac darty is during the launch store.

02:09:47 - Cathy Gellis (Guest)
Oh, I didn't realize it was the same store. They must have merged, but those are two big stores in france.

02:09:54 - Leo Laporte (Host)
Here are mario and luigi dancing. That's in paris also.

02:10:01 - Benito (Announcement)
Paris, france france not texas yeah uh, I don't know anybody in our does anybody in our chat uh have uh? Anthony says he got one, anthony got one.

02:10:11 - Leo Laporte (Host)
He says, his brother got one oh and Anthony got one too. Anthony, I'm very jelly, You're going to have to come over. He got it with a pre-order on Walmart Nintendo in order to avoid scalpers actually did, I think, a pretty good thing. They may have also created this frenzy by saying you can't pre-order it unless you've had a nintendo account, uh prior to the end of last year and more than 50 hours using your switch wow, that's a good question. I wonder how did the scalpers get any?

02:10:45 - Sam Abuelsamid (Guest)
I'm, that's a good question well, I'm sure they did, but like it doesn't seem like a big thing. I don't see an out no right now.

02:10:50 - Larry Magid (Guest)
Yeah, no, it sounds like a loyalty program you know, if you're a loyal heavy user, you get rewarded.

02:10:57 - Sam Abuelsamid (Guest)
Yeah, it makes sense right let me just look like buying a ferrari you know, ferrari won't sell you a ferrari unless you've previously bought the only reason I didn't buy one last week.

02:11:05 - Leo Laporte (Host)
That's the only reason, all right let me see if I can get. Okay. Here's a nintendo switch 2 on ebay, brand new, sealed in hand, ready to ship. It's only 200 over retail.

02:11:19 - Cathy Gellis (Guest)
Here's what actually only 100 over retail I mean, maybe some of it is people who are like I can get one did some private speculation as opposed to yeah, or I bought one for all the kids oh, they bought their own so they are on ebay, but you're going to pay $100 to $200 more.

02:11:36 - Leo Laporte (Host)
It's not like the double.

02:11:37 - Larry Magid (Guest)
Pre-orders. It's not the double you can buy them for below retail.

02:11:41 - Leo Laporte (Host)
I'm sorry, Benito, what did you say?

02:11:43 - Benito (Announcement)
It's not the double that scalpers usually charge when they get a hold of all this talk. Yeah.

02:11:50 - Leo Laporte (Host)
And I, you know, I tried to sign up and I guess I didn't have the requisite 50 hours on my old Switch. But I signed up for the Mario Kart World Bundle and all that too. It would have been fun. Maybe someday I'll get one. Anthony, you'll have to tell us how you like it. Have you played with it? But yeah, tariff worries might have been an issue. Nintendo prepared for that by raising the cost up front. So, um, I this kind of merges kathy's interests and, uh, sam's.

02:12:27 - Cathy Gellis (Guest)
You cannot name your car, eleanor no, no, no, um, eleanor, the car is um, let me not say the character it's from the movie gone in 60 seconds in the movie gone with in 60 seconds.

02:12:42
In that whole franchise there is a recurrence of a car referred to as eleanor yes of course, in all these movies eleanor looks quite different like sometimes it's yellow, sometimes it's silver, and sometimes it's new and shiny and sometimes it's old and dirty and broken. Anyway, in this is a case I wrote a post on texter about a court case from last week or the week before where the ninth circuit said eleanor is not a copyrightable character. You can't like stop other people from copy, from using a car and calling it.

02:13:17 - Leo Laporte (Host)
Eleanor, why not you could copyright one of the cars in the cars movie.

02:13:22 - Cathy Gellis (Guest)
Well, so the why not goes back to so character copyright is a really weird and pretty stupid area of copyright, because normally copyright is you can copyright the expression of something, but you can't copyright the idea of something.

02:13:39
If you could do the latter, you would be able to like oh, I had this idea that inspired this, and you could cut off everybody else from being able to express something with the same inspiration. But the problem with characters is you kind of have that thing where if a character emerges in a copyrighted work, not only is the work copyrightable, but potentially the character as well. So, for instance, like Sherlock Holmes, when the Sherlock Holmes stories were copyrighted or still under copyright, nobody could use Sherlock Holmes, because Sherlock Holmes was a character that also had its own copyright. Roughly speaking, there's a little bit around the edges, but so that was what was happening. But then there had been a case in, I think, 2008, where somebody had sued over whether Eleanor the car could qualify as a copyrightable character. Now, unlike the characters in the movie Cars, which were personified and had personalities and did things and stuff, like that Actual characters.

02:14:34
Yeah, they were personified and had personalities and did things like that they were. Yeah, they function. They were personified and they function like personified characters. They just happened to have had automotive form. Eleanor the car never really had a personality, it just sort of had characteristics. And a court then said you know, that's enough, you get your copyright in Eleanor the car. You know, that's enough, you get your copyright in Eleanor the car.

02:14:57
Then, years later there was another case where somebody sued for copyright infringement. Dc Comics sued for copyright infringement in the Batmobile because there was a guy who was making life size models of the Batmobile and DC Comics managed to sue and there's a lot wrong with that particular lawsuit. But one of the problems was this idea that the Batmobile was a copyrightable character given that it looks different in every movie and every comic that it shows up. There's no consistency to the Batmobile. It has no personality, it just sort of is a stylized prop that keeps showing up in a vaguely reminiscent stylized way.

02:15:36
But there was a case called DC Comics versus I don't know how to pronounce it TOWLE, which said you know, if Eleanor the car was copyrighted or copyrightable, then there's really no reason that the Batmobile shouldn't have been copyrightable as well. The Batmobile shouldn't have been copyrightable as well. But now the people who were suing each other over the Ellen or the car, they had a settlement agreement and the settlement agreement went out the window. So they sued each other again and now it's led to the Ninth Circuit saying that's nuts, there's no way this is copyrightable. So they have this decision using a test that the DC that the DC Comics case had actually come up with and said it fails all three prongs of this test. It's not distinctive enough, it's not persistent enough.

02:16:19 - Leo Laporte (Host)
I'm there were four different Eleanors, four different colors and four different gone in 60 Second films. Uh, carol Shelby licensing was suing uh, denise halicki saying that halicki's copyright claims over the eleanor ford mustangs were invalid. Halicki countersued, saying shelby's gt500 cr mustangs infringed her copyright in eleanor. I don't know who won.

02:16:48 - Cathy Gellis (Guest)
I don't know who won is um, the people who want to be able to use another Mustang and call it any Mustang and name it Eleanor.

02:16:58
There is not enough in Eleanor the car that would qualify for copyright in Eleanor the car, even if the movies that it appeared in are still under copyright. So it's a good decision. It dials back copyright, ability and characters, particularly for these inanimate objects that are really nothing but props. But meanwhile the batmobile case is still on the books and arguably that was a huge injustice and it was predicated a little bit on the or actually in pretty significant amount on this earlier case that's now been completely undone if you copyright the batmobile, you would be just copywriting the name, because what, which batmobile would you copy?

02:17:34
they don't don't bring logic into it. The court there didn't. Um, and you're talking about, if you're copywriting just the name, that's really a trademark. So maybe there would be some issues with trademark infringement. But no, the batmobile case was a copyright case where a guy you modeled it off of two of the batmobiles that appeared in movies that DC Comics had only licensed to the movie makers, so they didn't even make design those particular cars for those movies. It was really it's a screwed up case.

02:18:08 - Sam Abuelsamid (Guest)
A lot of it was that from the Tim Burton movies or the 60s Batmobile I think it was the tim burton ones a reasonable presumption of confusion.

02:18:18 - Larry Magid (Guest)
For example, I always thought it was interesting that madonna no, that matters for trademark, it doesn't matter for copyright yeah, because madonna protected her name, but the catholic church had it a long time before.

02:18:27 - Leo Laporte (Host)
Yeah yeah, there was one madonna before you that really, maybe even arguably, was better known.

02:18:34 - Cathy Gellis (Guest)
That is trademark and although I recently had to do a project where I was reading a lot about character, copyright and trademark is supposed to be distinctive from copyright and operate on a different axis with a different rationale. But it does appear that if there is consumer confusion, that does seem to end up informing the copyright analysis, even though it's not supposed to wasn't that the whole apple, apple music versus apple computer dispute? Because that was trademark, though when apple got, they made a deal, though.

02:19:03 - Leo Laporte (Host)
Yeah, they made a deal got upset well, you know, we, I learned a little bit about this, about something called reverse confusion, which is apparently something you can really get some damages for. We, uh, we have. We got a little tussle with uh, twitter over the name twit and we proceeded. Twit came first, then there was twitter and uh. The problem is reverse confusion, which is that people would assume we copied twitter, which is more damaging to us than even the fact that our name was, you know, copied. So you know, apparently that's the one you got to really watch out for.

02:19:38 - Bill Atkinson (Interviewee)
Yeah, you know what?

02:19:39 - Leo Laporte (Host)
else we want to watch out for in the skies over sonoma, not walmart drones, the county drones in our county, the apparently and my way. I have to say I have a little connection to this because the guy who's doing it is a buddy of mine, jt wick. He's the director of permit sonoma, the county's land use and planning department. He's being sued right now. They, they, they sent drones in the sky to look for pot farms, which apparently is okay, but then they started to find, like little you know violations of code, that residential properties there's a word that is really important for this analysis.

02:20:22 - Cathy Gellis (Guest)
This is a Fourth Amendment issue. Yeah, case law has gotten kind of weak on this point, but the key word is a word called curtilage, which has to do with what part of the property is visible from overhead curtilage, which has to do with what part of the property is visible from overhead.

02:20:37 - Leo Laporte (Host)
Well, apparently, uh, they have issued citations to people for having, uh, illegal horse operations. Uh, there's a woman who says that. Uh, she. She says she noticed a drone hovering over her farm. She's deaf so she didn't hear it. Somebody else said what's that? What's that drone doing up there?

02:20:57
At which point she ran inside her bedroom where she could see the drone photographing the property through her blind. She was looking out, they weren't looking in. She said she feared that the drone may have photographed her earlier while nude after bathing. So she's also a part of the lawsuit. I'm afraid to open my blinds or go outside because she's deaf. She wouldn't even know there's a drone overhead. Who knows if the county drone could be spying on me? A second plaintiff said the county used the drone program to photograph her horse stable and issue code violations. The only way she found out the county employee mentioned they had photos of her property. She filed a public records request. This is why these are so important, by the way, which left her stunned, according to San Francisco Chronicle, after seeing the county employees were monitoring her private property, including photographing her outdoor bathtub and shower.

02:21:50 - Benito (Announcement)
Hey, Kathy, how legal is it to shoot these out of the sky?

02:21:53 - Cathy Gellis (Guest)
Oh, is it? Probably not legal at all, but in terms of pure constitutional theory about what is OK and what is not OK, yeah, don't do it, because somebody is going to get you with some sort of code, can I? Train crows to attack it. How about that?

02:22:11 - Larry Magid (Guest)
Yes.

02:22:12 - Cathy Gellis (Guest)
Yes, I could definitely do that. What about radio jamming? I don't know If the drone is in your yard that incessantly.

02:22:18 - Leo Laporte (Host)
Then Radio jamming how about that? Yeah, what about radio jamming? That's a good question, Okay.

02:22:23 - Cathy Gellis (Guest)
You know I am a lawyer. I'm not your lawyer. None of this is legal advice and one of the things that is an issue is that. You know, fourth Amendment law and other related laws were sort of developing and then as technology hit them, sometimes they kind of veered off and started to develop in ways that were not necessarily protective of individual rights. They tended to defer a little bit more to the state action. I wonder if those are going to get rethought in some way, but in the meantime I don't pick that fight with the state. I don't think it sure sounds like mission creep.

02:22:58 - Larry Magid (Guest)
And if I were a resident of Sonoma, if it's Sonoma County, I would go to my board of supervisors and say, hey, wait a minute.

02:23:05 - Cathy Gellis (Guest)
Yeah, I mean, I think to some extent I think to some extent that is the way to do it, because right now it's a political solution will take care of this. But for the people who've got citations, they could potentially challenge them and challenge them on a fourth amendment basis. Um the case law. I'm not enough of a fourth amendment expert. I know that, like I took some, you know some classes in law school.

02:23:29 - Leo Laporte (Host)
Their curtilage is a word that will be relevant to the, to the dispute, but but is that when you have like a gallon of milk for a little too long and it curtilages, is that? Is that what you're talking about?

02:23:43 - Cathy Gellis (Guest)
You know, I don't think you should.

02:23:46 - Larry Magid (Guest)
Don't make cheese out of anything here. Ok, but it'll impress people that you know the term.

02:23:52 - Cathy Gellis (Guest)
That's why I'm basically offering it to everybody.

02:23:54 - Larry Magid (Guest)
I can't give you the Fourth Amendment analysis, but here's a nice fancy word. I think it is informing one enclosure with it. That's according to some dictionary.

02:24:03 - Sam Abuelsamid (Guest)
Just just just consult with your own attorney and you know, mention the word curtilage and say here I heard this word.

02:24:15 - Cathy Gellis (Guest)
All right. So the word with curtilage is. I remember being taught a case where curtilage mattered when I was in law school. So damn it, it needs to matter now, but but it's, it's a.

02:24:20
You know a lot of stuff in terms of the surveillance you can do with technology. We already had the development of a Fourth Amendment law and then we added technology to it and sometimes you would end up with some case law that would be very protective of the liberty interest. Sometimes you would end up with some case law that would be very protective of the liberty interest. Uh, like, uh, justice scalia wrote a case called kilo or kylo, I think that's what it was um, which involved like, could you use like a heat gun, um on a house to like surveil what was going on in the house? And he had said no, but but that was sort of like the opening volley. And then there's been a lot of creep where the the courts, increasingly, have said yes and yes and yes to more technological surveillance, and maybe we need to start seeing a snap back. To hang on a second, the liberty interests here were pretty important and this is going to be a problem if we keep saying yes to it especially with cameras, literally everywhere today.

02:25:10 - Leo Laporte (Host)
I mean yeah, yeah well, um, I want to thank you all for allowing me to learn a little bit of law here today. It's been, it's been a great time. Uh, thank you, larry maggid. Everybody should go to connect safelyorg. We didn't get to your story about aipcs, but I think that would be well worth reading if people go to larrysworldcom a laptop built for ai, but how long oh, this was from the merc, yeah, oh, nice, all right, um, thank you, larry, appreciate it. What would you like to plug? Anything? Well, I was going to say I would.

02:25:44 - Larry Magid (Guest)
I'd ask my attorney about curtilage, but it would probably cost me a thousand dollars for him to look at that guarantee.

02:25:50
Yeah, no, I what I just asked chat gpt, that's what I did those of you who are worried about your kids and the apps they're using. Go to connect safely and click on explore by app and read about those apps. Read our parents guides, because we can't solve all your problems. We can at least tell you what they offer. What controls are there, and maybe we can assuage some of your fears or create new fears you haven't even thought of yet. But you know it's a way to educate yourself.

02:26:14 - Leo Laporte (Host)
Very nice. It's great to see you, larry. Thank you for being here. The number eight now number eight most visited podcast guest, or maybe you're. Maybe you're seven. Maybe you're moving up on the charts. Yeah, thank you, larry. Kathy Gellis soon to be on the top of the charts. We love having you on. She writes for TechDirt and you can find her at CGCouncil C-O-U-N-S-E-L dot com. Anything you want to plug, kathy.

02:26:42 - Cathy Gellis (Guest)
Not currently, but lots of TechDirt posts and hopefully lots more soon.

02:26:46 - Leo Laporte (Host)
Love what you do on TechDirt, really appreciate it. And, of course, sam Bull Sam and my car guy. You know we got to do a club event, uh, about cars. If you would like to do that sometime, we'll schedule it absolutely, because I think a lot of people miss you on the radio show and we don't get to talk cars as much as I'd like to. For instance, I wanted to ask you about that 20 000 truck that doesn't have a radio or anything else in it.

02:27:13
Maybe that'd be a good subject and and there are plenty of others car people love to talk about cars, so let's, let's get that on the books. Okay, sounds good. His podcast, wheelbearingsmedia with nicole and robbie, and it's a great show. You must. You must listen if you love cars. And he is a vp of research at telemetry.

02:27:30 - Sam Abuelsamid (Guest)
I want to thank all. Do another podcast now too. A short daily two-minute hit every day called the telemetry. I want to thank all. Do another podcast now too. A short daily two minute hit every day called the telemetry transportation daily.

02:27:39 - Leo Laporte (Host)
Oh, that's cool yeah, is that for people in the biz or would every anybody be interested?

02:27:44 - Sam Abuelsamid (Guest)
anybody who's interested in the auto industry or vehicles. Uh, might I mean it's only two minutes, you know give it a listen. You might find it interesting. Might uh get some useful insights out of it.

02:27:53 - Leo Laporte (Host)
What a good idea, so it's just like a quick update on what's happening. I'm going to have to listen. Is it available everywhere?

02:28:03 - Sam Abuelsamid (Guest)
It's available everywhere. Ignore that I used one of the. I think I used Gemini or ChatGPT or something to generate some JavaScript code to put an embedded player in there and it was working for a while, but I think maybe the proxy that it was using to generate the player feed wasn't working. But you can find it in. You know, Apple Podcasts, Overcast, it's everywhere. Just Telemetry Transportation Daily it's out there.

02:28:38 - Leo Laporte (Host)
Very nice. Yeah, congratulations, thank you, that's great. Thank you Sam, thank you Kathy, thank you Larry, thanks to all of you for joining us, especially our Club Twit members who make this show possible every week. Don't forget tomorrow, big day for Apple fans the Apple WWDC, the Worldwide Developers Conference, convenes. We will have coverage in the Club Twit Discord for both keynotes the morning keynote that's open to the public, and the State of the Union keynote following. Micah Sargent and I will be talking about what's new with Apple. It should be interesting to see how little they mention AI after getting all that egg on their face by overdoing it last year. We'll be watching with great interest.

02:29:22
We do Twitter, as I mentioned, every Sunday, 2 to 5 pm Pacific. You can get the show live or get it after the fact. There's also a YouTube channel for the video, but the best way, of course, to subscribe is in your favorite podcast player. Leave us a five-star review if you would let the world know about the longest running and, I think, best tech podcast in the world. Thank you everybody. We'll see you next time. As I've said for the last 20 years, another twit is in the can bye-bye.