Transcripts

This Week in Tech Episode 929 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.

Leo Laporte (00:00:00):
It's time for Tweet this week in Tech. Oh, we have a great panel for you. Owen Thomas is here from the San Francisco Examiner Glenn Fleischman from the incomparable and from Windows Central. The editor-in-Chief, Daniel Rubino will talk about Microsoft's build some big AI announcements there. We'll, we'll talk about Elon Musk's disaster and what happens when your lawyer uses chat. G p T. It's all coming up next on Twitter.

TWIT Intro (00:00:30):
Podcasts

Leo Laporte (00:00:31):
You love

TWIT Intro (00:00:32):
From people you trust. This is Twit.

Leo Laporte (00:00:44):
This is twit this week in Tech. Episode 929 Recorded Sunday, May 28th. 2023 sous women and $4 potatoes. This week in Tech is brought to you by aci Learning IT Skills are outdated in about 18 months. Stay ahead of the curve and futureproof your business competitiveness with customizable entertaining training. Fill out the form@go.acilearning.com slash twitch for more information on a free two week training trial for your team. And buy worldwide technology with an innovative culture, thousands of it engineers, application developers, unmatched labs, and integration centers for testing and deploying technology at scale. WW t helps customers bridge the gap between strategy and execution. To learn more about wwt, visit wwt.com/twi and by lookout, whether on a device or in the cloud, your business data is always on the move, minimize risk, increase visibility, and ensure compliance with lookouts Unified platform. Visit lookout.com today.

(00:02:03):
It's time Forwe. This week at Tech. The show covers the week's news and we covers it today. <Laugh> with three excellent panelist. Glen Fleischman joins us from Parts Northeast. He's the editor of a brand new book, Kickstarter Northwest, what did I say? Northeast? Crazy <laugh>. I was thinking You and Maine. You're going to Maine North to print your book New North of Yes. Hello, hello, hello. Glen Fleischman, editor of Shift Happens. Good to see you again. Thank you. Also joining us from windows Central Magazine, the editor in chief. Do you still call it a magazine, Daniel? No, no. No. Blog. Just Windows Central. Just a thing. It's a thing. Yeah. Tell me

Daniel Rubino (00:02:48):
About I do joke around people. I'm like, yeah, I just, I'm I blog. That's

Leo Laporte (00:02:51):
All I do. <Laugh>. I'm a blog. Are you in your mother's basement, Daniel? Where are you?

TWIT Intro (00:02:55):
<Laugh>.

Leo Laporte (00:02:56):
Exactly. What are you doing there? Good to have you here. Of course, Microsoft Build was this week. Lots of AI news. We'll get to that in a bit. But before we do, let's introduce panelist number three. My good friend Owen Thomas, columnist from San Francisco Examiner. He's Alex Wilhelm's Twitter. Daddy, do you still say that now? Is Twitter over? It's done. Oh, you're muted. I, I,

Owen Thomas (00:03:19):
I think, I think you need to say Twitter Dad, not Twitter, daddy. That's different.

Leo Laporte (00:03:23):
That's a different, two different things. That's,

Owen Thomas (00:03:25):
That's more only fans.

Leo Laporte (00:03:27):
Okay. <laugh> is Twitter. Is Twitter furry? Is that right? No, that's wrong. <Laugh>. That's also wrong. It's very

Owen Thomas (00:03:33):
Raw. It's

Leo Laporte (00:03:34):
Very wrong this week. Twitter had a little attention on Wednesday. As Elon constantly reminded us Rhon DeSantis governor of Florida was to announce his 2024 candidacy for presidency didn't go exactly as planned. I, somebody noted that Twitter ha had fired 97 out of the 100 people who worked at spaces, which might have had something to do with it. <Laugh> lot of echoing. They

Owen Thomas (00:04:14):
Also, they also stopped paying for the the service that oh. Helped keep spaces online that, you

Leo Laporte (00:04:22):
Know, they, was it DDoS protection kind of a service or, I

Owen Thomas (00:04:27):
I think it was, I think it was like a more, like a cdn or

Glenn Fleishman (00:04:31):
Right. They literally didn't have enough servers spun up. I, I think they're gonna, when you look up puric Victory and the Diction Store, I think they're gonna put Elon's face in the next edition. But, so it's pretty amazing. The cell phone there. His,

Owen Thomas (00:04:44):
His cost containment strategy at Twitter seems to be, I'm going to randomly stop paying for bills and randomly shutting servers down and see what breaks. And this is an example.

Glenn Fleishman (00:04:54):
It's that story that came out, was it just last, this is last week about reporting what people inside the company are responsible for real estate and other bills were were talking about like what they were asked to do which seemed you know, a little, a little wild

Owen Thomas (00:05:09):
What? Right. And, and, and in the filings that, you know, they were saying like, look, I, my career is in real estate. I will not be able to work again in this industry if you have me, like, try to stiff our landlord. Yeah. And, you know, and the answer was, well, we don't care. You know, management, I, Elon Musk don't care. Or

Glenn Fleishman (00:05:28):
I might be actually sentenced, you know, to jail or the Yeah. I

Owen Thomas (00:05:32):
Might be personally sued.

Glenn Fleishman (00:05:33):
Yeah. Persecuted or, you know, prosecuted or persecuted. Yeah. It's but I mean, isn't this, this is what, you know, I think he maintains when he's talked about it in public it seems like he's you know, this is what rich people allegedly do, is that they don't pay bills strategically to see which ones they have to pay. And that this seems, you know, I don't think that's ethically the case for everybody, but you know, certainly some other prominent people have done that. So it's amazing to see

Leo Laporte (00:05:56):
It blatantly and on such a huge scale. It did something happen when Trump got elected president that it just completely changed. Like it's, it, it slid the ethical Overton window over about a thousand places. Cuz he's very famous also for not paying bills. And it, it feels like he made so a lot of things. Okay.

Daniel Rubino (00:06:19):
Yeah. And there's no repercussions, right?

Leo Laporte (00:06:21):
That's, you get away

Daniel Rubino (00:06:22):
With, if I don't pay bills, right. If I don't pay bills, I'll Right. I'm gonna get in trouble. This is

Leo Laporte (00:06:26):
Like, we

Daniel Rubino (00:06:27):
Talk about this bunch of bad things happen. We

Leo Laporte (00:06:28):
Talk about politics and stuff, but there's also a the president is also in a way, a a leader of how, what people think. And you can really establish kind of a, a, a point of view that yeah, it doesn't matter. You can now say anything you want. You can, you can do whatever you want. And it doesn't matter. You get away with it. I think Elon must have watched this with interest and is now sort of adopting the art of the deal, as it were. On the other hand, there has been some very good news for Elon. That was the bad news, by the way. I gotta, I gotta point out the Subtweeting that the New York Times did. This is an article by Ryan Mack and Tiffany Shoe DeSantis. His Twitter event falls short of the reach of past life live streams.

(00:07:15):
Within hours of the event participants celebrated the achievement. David Sachs, who's been a big donor to the DeSantis campaign, declared it by far the biggest room ever ho held on social media. Desantis himself said there were probably about 10 million people who would've watched the event. The time says the very kind of crisp single paragraph sentence, they were wrong on both counts. <Laugh>, they were wrong on both counts according to Twitter's metrics. High of 300,000 concurrent listeners that is actually not. And 3.4 million afterwards. Listen to recording. Not quite 10 million nor the biggest room ever held on social media. I mean, Facebook had an 800,000 person Facebook Live event which by the way, featured two Buzzfeed employees placing rubber bands around a watermelon until they exploded. Oh yeah. That was great. Oh, that was great. That was some entertainment. 5 million people watched the entire event at some point.

(00:08:20):
The 2017 times is really being mean here. Livestream of a pregnant giraffe on YouTube brought in 5 million viewers a day <laugh> in fact, even Twitter spaces had done more. 3 million people last month listened to an interview of Elon by a BBC reporter and a Twitter space. So yeah, this is, this is severe degradation from what spaces had been, and certainly nothing compared to other AOC had to tweet. More people watched me play a game on Twitch <laugh> and watched your announcement, which is true. Was she playing among us? I think she was. Yeah. but that's pretty exciting as is watching a watermelon about to explode. So either way I can understand the interest. Good news for I Is there

Glenn Fleishman (00:09:12):
A are, are you making some kind of DeSantis metaphor there?

Leo Laporte (00:09:15):
<Laugh> watermelons explosion, you think? No, I wouldn't, never. <Laugh> No, I'm not. That's any re any resemblance to a person living or dead is accidental. I should say that the became of every show. Good news for Elon, though, the FDA has approved neural inx study of human brain implants. Woohoo. And anybody, do you wanna get in line for this one?

Daniel Rubino (00:09:46):
It's gonna be a real life blue check <laugh> for people. <Laugh>.

Glenn Fleishman (00:09:50):
Because

Daniel Rubino (00:09:51):
You know what? Those are gonna be deferred. People sign up for this. Oh, yeah.

Leo Laporte (00:09:54):
It's gonna be Elon stance. Right. And then

Daniel Rubino (00:09:56):
They'll be on Twitter. They'll have their little icon. It'll be a blue check and then a little brain next to it. You know, it's gonna be awesome. Let them go. It's fine. Yeah.

Leo Laporte (00:10:03):
I have Elon in my mind. I mean, to his credit, the whole idea is really to help people with brain injuries live more normal lives. So it's a, I'm sure that's why the F fda cleared it. They cleared Neuralink to use its brain implant and surgical robot <laugh>. Just what could go, what could possibly go wrong, <laugh> for on patients, but declined to provide more details, should point out Neurolink has been in trouble there, in trouble with the United States Department of Agriculture for animal welfare violations. They're in trouble with the Department of Transportation. This is a weird one. They're probing whether Neurolink illegally transported dangerous pathogens on chips removed from monkey brains without, oh my God. Without proper containment measures. Okay.

Glenn Fleishman (00:10:55):
It's like a walking set of science fiction movies about to be made. It,

Leo Laporte (00:10:58):
He, it's like he, I I don't know why he doesn't have a secret layer in a volcano. It doesn't, it doesn't

Glenn Fleishman (00:11:05):
Seem You don't know that. He doesn't. He could know if it's secret.

Leo Laporte (00:11:09):
That's

Glenn Fleishman (00:11:09):
True. You know, I, the exciting news though, you know, the neural lake thing is, is brain-based, I realized. But I was incredibly excited about the the fellow who had been paralyzed, who was able to regain the ability to use his legs. Yeah. The bypass that like a bridge, a bridge technology from spine to spine bypassing the disconnection. And I'm like, that is, you know, whenever, whenever we talk about the future, I'm always like, well, you know, I don't, we don't need flying cars necessarily. We don't need, I don't know, AI-based recipe generators, but it's like people being able with aphasia, being able to speak, people, being able to walk again, this stuff is fantastic. And the question is, will anybody be able to afford it? Or will it be, you know, 10 million solutions that insurance companies can't reasonably cover for, you know, the people who need it. So it's, but it's, I think that's, so I'm excited. So Neurolink as a concept, I don't know what his end game is, but I'm excited if it leads to people being able to resume the lives that they had before.

Leo Laporte (00:12:04):
That's, yeah. I mean, incredibly, and I'm sure that's why the FDA was a little more open to the work. It's not, it's not. Just so you can go in Mark Zuckerberg's metaverse Swiss researchers implanted an electronic device in a paraplegic skull on top of the region of the brain responsible for controlling leg movements. And then this is the AI part using algorithms based on adaptive AI methods. This is from the Financial Times movement. Intentions are decoded in real time from the brain and then transmitted wirelessly to a neurostimulator, connected to an electrode ray over the part of the spinal cord that controls leg movement that's below the injury site. And he was able to walk.

Glenn Fleishman (00:12:49):
Yeah. And then also even be able to do that. Do I remember this right? It was even after, after a lot of training, he was able to do it without the connection in place, I think was the case

Leo Laporte (00:12:59):
Too. Oh, that's interesting.

Glenn Fleishman (00:13:00):
So even though the signals aren't there, he was able to, to produce the response required. I mean, just this is, you know, it is his science fiction that we're living in, which is makes the good kind the non dystopian kind. Yeah. Occasionally you get it.

Leo Laporte (00:13:14):
The Dutch man's spine was injured 11 years ago in a bike accident. They upgraded his digital bridge. He said, this feels radically different. Before I felt that the stimulation was controlling me, now I am controlling the stimulation myself. I could take steps that feel natural and there's no external computer. It's all in, you know, he's, it's all in his head, so to speak. That's, that is, yeah. Thank you for bringing that up. That's a, that we should talk about the positive stuff, really.

Glenn Fleishman (00:13:44):
That feel good story for sure.

Leo Laporte (00:13:46):
Yeah.

Daniel Rubino (00:13:46):
Yeah. It is. Feel good. But I'll also point out that there's down the road security risks with this as well. Oh, yeah. Oh lord. When you have, when you have wireless devices planted in your brain, people are gonna hack those for good and for bad. You know? So that's the other side. We do need to think about, just like ai, just as we're building out ai, we need to think about both things. How do we detect the bad stuff using AI and how we use AI for good stuff. We gotta do the same here with these implants and brain stuff. We gotta be thinking about security and how We'll, because man, now we're, you're actually really messing with people's lives potentially. Right?

Leo Laporte (00:14:24):
That's right. Well, we've seen that with the pacemakers that have that were happy. Right, right. Right. Yep. That, that, yeah, that's something to be, obviously the good news is I think we're much more aware of security than we used to be. And presumably medical researchers who are creating, especially wireless technologies like this, I would hope are, are really paying attention to that. It's not, not impossible to encrypt those signals and so forth. I,

Glenn Fleishman (00:14:47):
I wish I had Cory doctor to pull out from behind a bush right here and say, but wait. All the systems proprietary and no one can investigate the code. And we're relying on Oh, good point. The FDA's underfunded groups to vet that they're secure and these companies never actually invest in security. Then I'll push Cory back behind the bush.

Leo Laporte (00:15:01):
Ah, that's exactly right. They're good.

Glenn Fleishman (00:15:03):
Yep. Right. He's right.

Leo Laporte (00:15:04):
Yep, he's right. My

Glenn Fleishman (00:15:05):
Wife has a bone anchored hearing aid, which is great, but, you know, every few years they're like, we've ended end of life to this thing that costs several thousand dollars. We're fortunate that it got covered by insurance. Other insurance payers are not because the premiums reflect that, but there's no open source thing. It literally snaps onto an anchor on her body. And and that

Leo Laporte (00:15:25):
Was surgically third

Glenn Fleishman (00:15:26):
Party

Leo Laporte (00:15:27):
Implanted.

Glenn Fleishman (00:15:27):
The anchor is surgery planted. Now they implant a magnet under the skin. And so it's magnetically connected apparently. But it's a very simple device in in reality, the implant is the thing that is medical. And but it's not open. Like, you know there's, there's a Baja Bone anchored hearing aid, but there's all those other kinds of hearing aids available. What's the one that's is the one you wear an external part on? I forgot what it's called. A, a a, forgot the name of it, but it's one that goes directly and stimulates the inner ear where there's an implant.

Leo Laporte (00:15:55):
Oh, the cochlear implant. Yeah.

Glenn Fleishman (00:15:57):
Cochlear implant. Yeah. And so all these, none of these devices have any requirement or necessity for access outside the company's proprietary interest. Cause they developed them. And that's a shame, so forth. So when the company goes outta business, you saw the thing where some people had gained regained the limited ability to see like some shades of color. And the company had to shut down, I think. And there was no future for that implant, which is or even supporting them medically for implant side effects. So un unwanted consequences from closed source development in the medical world.

Leo Laporte (00:16:31):
This is where the communist in me comes out and just points out that for-profit medicine is is destined Oh, yeah. To disappoint in the long run. This is just not, there's some things that should not be capitalistic and medicine is one of them. Yeah. Yep. Absolutely. Yeah. Oh. Well, <laugh> <laugh>, oh, well, sorry about that. You were talking about chat, G P t hallucinating. We all saw the story. Jeff Jarvis leaped on this. He's already ri he's writing a book with it. You already started the chapter. A New York law lawyer is in trouble because his firm used chat G P T for legal research. The case involved a man suing an airline over personal injury. His legal team submitted a brief that cited previous court cases saying there's a precedent for why this should move forward. The airlines lawyers wrote to the judge said, what? We, we can't find these cases, <laugh>. We can't find 'em. Six of the submitted cases appear to be, this is the judge speaking bogus judicial decisions with bogus quotes and bogus internal citations.

Glenn Fleishman (00:17:49):
That's

Leo Laporte (00:17:49):
Amazing. Over the, he demanded that the legal team explained itself over the course of several filings. It emerged, this is from the B bbc, that the research had not been prepared by the lawyer, but by a colleague at his firm who used chat G p T to find precedents. Go to the library, man. It's in the basement. We bought those books for a reason.

Glenn Fleishman (00:18:11):
My favorite part is when the guy said, but I asked chat, G p t whether the citations were real. And it said they were, and then he attached screen captures of him asking chat, g p t that, and assuring him they were correct

Leo Laporte (00:18:24):
Screenshots attached. The filing appear to show a conversation between Mr. Schwartz, the lawyer who did the research and chat. G p t in, in one message says, is Vaga a real case referring to the quoted President Vaga versus China Southern Airlines, one of the cases no other lawyer could find chat. G P T says, yeah, <laugh>,

Owen Thomas (00:18:46):
Would I lie? Yeah,

Leo Laporte (00:18:48):
It's real. So then Schwartz said, well, what's your source? After quote, double checking. Chad says, no, no, it's real. You could find it on Lexus Nexus or Westlaw <laugh>.

Owen Thomas (00:19:01):
That's awesome.

Glenn Fleishman (00:19:02):
I don't know why the guy didn't search. I don't Yeah, there's a lot of things about that. But while that, it's like the laziest thing that a lawyer, well, they're

Leo Laporte (00:19:08):
In trouble ever does. The judge is not happy. Yeah. There's gonna be a disciplinary hearing on June 8th.

Owen Thomas (00:19:15):
And doesn't, doesn't this open him up to a malpractice suit from his client,

Glenn Fleishman (00:19:22):
Probably you kind of think, and also being disbarred or, or sanctioned. And I love that. His thing is like, I didn't realize that chat G p t would give false results. And every time you use chat b chat, g d p chat, g p d at the bottom, it says, it may not offer real results. You know, careful. And you're like, all right. Think the disclaimer's pretty good. But I think, I'm sorry.

Owen Thomas (00:19:41):
Yeah. Oh, I, I, I think the, the context here though is, is interesting. Because this comes as as meta and Google and and Twitter have all vastly reduced their resources devoted to fact checking, right? So bots like chat, G P t are training on a corpus of publicly available internet data that we know is increasingly faulty, increasingly un fact checked. And, you know, and, and also being fed by AI more and more so we have chat bots creating made up internet nonsense training on made up internet nonsense. And it risks becoming this like vortex of like verbal vomit that, you know, that will doom our ability to trust anything we we read online ever.

Leo Laporte (00:20:35):
I blame us. It's our fault. And I don't mean twit, I mean, all is too, because the tech, we have been singing its praises. The New York Times has been Kevin Rouse said chat, G P T fell in love with me. We have been acting as if this is a thinking entity. And no makes mistakes once in a while. But, you know, and so I think we oversold it to the point that this lawyer said, well, everybody says this thing's really good and, and really didn't. He didn't know that it was so risky. I bet ya

Owen Thomas (00:21:12):
Shouldn't we be as a, you know, as a species more sophisticated about this. I mean, I remember as a, as a kid in like the, the 1980s hand coding Eliza, you know, the Yes. The the computer psychotherapist. Like this was a, a, this was a program you could like type, it was that simple. In basic. Yeah. So, you know, you got a computer magazine, you typed, you typed out all the,

Leo Laporte (00:21:39):
Eliza, by the way, is still in Eex. I can, if you want, I'll show you <laugh>. It's still around, but it was dopey. You knew it was dopey because Eliza, you would say, well, how's your day been? And it would say, well, how's your day been? You know? I mean, it was, it was just echoing kind of stuff back to you.

Owen Thomas (00:21:57):
But, and yeah, people, people were fooled by it. Oh,

Leo Laporte (00:21:59):
Yeah. There was a lot of conversation. I may, I remember back then about, well, could this replace therapists and things like that. Great piece I wanna recommend in the i e e journal spectrum by Rodney Brooks, who is oh yeah, a roboticist. He was until 2007, director of the Computer Science and Artificial Intelligence lab at m I t. He's he's taught at Stanford and Carnegie Mellon. And he's at three robotics startups in the title. Just calm down <laugh> about Jet G p t four already. But the, this is really important and stop confusing performance with competence. You read, you read the article, Glen.

Glenn Fleishman (00:22:44):
Yeah. He's, I mean, he's always a voice of reason. He's not a hyper, you know, I think he, even when when he started getting into robotics, he was very excited about it. But I don't think he ever, I'm trying to remember like over, I mean, almost 20 years of following his work, I don't feel he ever said, like, you know, did a segue kind of announcement. This is gonna change transportation as cities or whatever. He is like, well, it'll have applications. And then he was behind iRobot. So, you know, there are, I don't know how many millions, hundred tens of millions of Roombas and and knockoffs out there. It did change home vacuuming. It's a narrow scope. And I think that was kind of,

Leo Laporte (00:23:20):
But you would never ask a Roomba for legal citations in a lawsuit, right?

Glenn Fleishman (00:23:25):
Oh, just wait till they add the plugins.

Leo Laporte (00:23:27):
<Laugh>.

Daniel Rubino (00:23:27):
Yeah. Maybe. I mean, to be fair though, paralegal jobs during one of the, the top on the list for AI to like, take over, like when people talk about AI replacing jobs, paralegals is gonna be one of those professions because it is just essentially research and summarizing data. Well, AI is gonna be, will be good at that. It is good at that. But there will be errors for now. And right now that's where we are. Right. But the thing to remember too is that even though we have these, you know, these goofs and screwups will make news and come out, but you know, the rate at which the stuff is improving is exponential. And it's just gonna be very quickly, these stories will go away. And we won. Be, I'm gonna dis

Glenn Fleishman (00:24:09):
I gotta disagree with you. It's not exponential though. It's not even geometric. It's, it's thi this is where I will, I will, I'll go off on you. Let's have an argument. No, in just a very polite, okay. <Laugh> is is, I think there is a tendency, I think there's a hu very human thing. And also in technology where everything is exponentially know chip growth and speed and processors a lot of hardware things to assume that certain kinds of things, or not soon, but to feel that they're going to be gonna continue at the rate of pace in artificial intelligence. We've seen again and again that there have been significant in, particularly in artificial intelligence, but in other software related ca capacities, we've seen again and again that things have improved to a certain point and then leveled off and only eked out improvements after that.

(00:24:49):
And I will, I remember very distinctly interviewing Jan Lacoon almost 10 years ago, who's one of the fathers of deep learning. Deep learning came out of the stalls. They had in previous you know, machine learning algorithms that could never get beyond, I dunno what it was, 80 something percent recognition on photo matching, on voice recognition. Deep learning was, ah, it's the new thing. It's can get to a hundred percent. And he said, even in the middle of all this, it's like 2015 or something he said, well, deep learning's kind of, we're reaching the end. It's not gonna improve that much more. We're eeking it out. Maybe this new kind of learning I forgot what it's called. It's not reflexive, but it's like a, a responsive learning thing with the cycle. I think it might be more like the generative adversarial network approach that, that developed.

(00:25:32):
He said, this has some promise, but I don't know. Like, he wasn't like, throw away deep learning. Like, we got to this, we got so much better, but it's not there. I am absolutely convinced we're gonna see this level off in the same way. So I don't think there is a 99.99% for large language models or any combination of things, but I think we will, we'll certainly be better than the state we are now. And chat G p T four is so much better than 3.5. You can compare them side by side. But I don't think it's gonna be a, i I would strongly disagree that it's gonna be an exponential improvement. I think we are, we are near, we're getting towards the top of the significant improvements that can be made.

Daniel Rubino (00:26:10):
I think it's gonna continue to improve. But I mean, I will say in terms of the human brain and what we know, this stuff is very far from that, you know, these things still have issues with inference. They cannot do it. They can't do basic tasks that even a child can do yet. So there is, and I don't think a large language models are gonna improve upon that. I'm saying though, in terms of what we're doing right now with them, they're going to improve very, very quickly simply because of the exposure amount of people that are using them. And the data that's being fed into 'em has been done at this scale. And so I think we'll see large improvements. I'm very skeptical of AI currently in the sense that it's going to reach, you know our agi you know, I think that's extremely far off. We, we don't even understand how the human brain works ourselves, let alone being able to figure it out in these kind of systems. So I'm not saying we're gonna reach that part, but I think within what we're doing right now with LLMs, I think it's, we're gonna see a dramatic increase over the next few months. And especially when we go to the next level of G P T. But it'll be confined. You're right. I mean, we're not gonna have like robots walking around in two years, you know, like, you know, acting like people.

Leo Laporte (00:27:19):
This is

Glenn Fleishman (00:27:19):
Jermaine.

Leo Laporte (00:27:20):
This is Jermaine. I'll fight you over that. The biggest story for Microsoft build was ai. I mean, this was the AI conference starting with Satcha Nadela saying, we've got 50 plus ai in products to show you. I'm gonna talk about five. They're putting AI in search, they're putting AI in Microsoft 365 in office. They're putting it in Windows even. Yeah. And, you know, in all the screens, Microsoft showed at the very bottom in fine print, you could barely read it. Said that disclaimer, you know, this is generated by AI and may not be accurate. Yeah. that's a lot to put on something that may not be accurate. Rodney Brooks wrote an article to, to solve the problem between you two. I will, I will throw this in this, he wrote this in 2017, so he wrote this six years ago, the seven deadly sins of AI predictions, <laugh>.

Daniel Rubino (00:28:15):
Hell

Leo Laporte (00:28:16):
Yeah. And, and one of them really is, and I think this is what's going on right now. We're seeing how fast we are going from, from Dolly to Dolly to, for, to stable diffusion from chat G P T three to chat G P T four. It looks like, and it might even be Glen Exponential at this point, but it, but I think you're not saying it's not exponential or it's not geometric. What we don't know is, is it continuous and well, it's the

Glenn Fleishman (00:28:42):
Improvement factor, right? It's like, how, what is your, what's perfection? And perfection would be a hundred percent. I mean, it's

Leo Laporte (00:28:48):
Cnos paradox. You're not gonna get there, right?

Glenn Fleishman (00:28:50):
But a hundred, let's say it's 99.9% voice recognition, right? I just did a big roundup for for charter on ai transcription tools. And I did earlier this year, did a lot of you know, checking major services that are available to individuals. And some of them also included chat G p t summaries or similar ones. And you know, a few months went by, the review just went up and I was doing a bunch of interviews doing a history of Yahoo pipes, by the way, for those, remember

Leo Laporte (00:29:18):
Yahoo Pipes? Oh my God, I missed Yahoo Pipes. Yeah. That was you. Excuse me. That was really great. <Laugh>,

Glenn Fleishman (00:29:26):
I'll, I'll,

Leo Laporte (00:29:27):
It was so great. I'll let you know the interview. Perfectly. <laugh> dropped an F bomb. That's how great it was. Awesome.

Glenn Fleishman (00:29:33):
But, so I interviewed you know, I don't know, 10 people all the original folks and had these 40 minute, hour and a half conversations, record it all as I can do now, and that I take detailed notes. I try to keep verbatim quotations. I've got a recording, but I feed it into the, one of the best of these AI recognition things. And I hadn't seen the chat G P T four version, the summary, because I'm giving it, I'm large language models are really good at taking a corpus of text and comparing it to a corpus of text. It does so well with that as opposed to extrapolate. So I'll give it the interview and it says, here's a summary of what you talked about. And I thought, if I do this myself, it would take me 30 to 60 minutes to come up with a list that was that good.

(00:30:14):
And, and it was very good. And it was also a a hundred percent accurate. There was nothing introduced. And I was like, well, here is a use case. Here's something that works for me today that helps me inform my reporting, because it allows me to review that and find the spots in the text in addition to the fact that I'm getting extremely reliable, like 97, 90 8% AI transcription. So that to me is kind of like, oh, well, we're a new goal in age for reporters who need transcriptions and, and and more efficient ways of producing their work. Right. Without giving up the actual writing part.

Leo Laporte (00:30:47):
So there's some useful things. I note that the, the big one where you get in trouble is anytime you use the word think, cuz these are not thinking, they can't, extrapolation is a good word. Mm-Hmm. <affirmative>, they cannot take existing facts and, and extrapolate from it. Brooks uses the, the example of, you know, you can have a picture recognize that that's two guys in the park throwing a Frisbee. And you might say, wow, that's really impressive. But you can't say, well, can you eat a Frisbee? How far can you throw a Frisbee? It doesn't know Frisbee. It's, it doesn't, it doesn't understand cons, the concept of Frisbee. It's not thinking. And I think we owe it to the unwashed masses, normal people to, to kind of really belabor that point. The problem is, we're also excited about this and it fits our science fiction, you know, fantasies that I, I think we've oversold it to people and I think it, we really,

Daniel Rubino (00:31:44):
It's also, I mean, it's just what, it's what humans do, right? Yeah. We

Leo Laporte (00:31:47):
Have, it's

Daniel Rubino (00:31:48):
Normal that we anthropomorphize, you know, or like they're, they're thinking and we, we do it, we see faces everywhere and we see something that's right with the face. We treat it differently, you know, so humans are just wired to sort of see this behavior Yeah. And then extrapolate from it and believe that. So we're just naturally flawed. You're right though. I mean, we do need to get it out there. You know, that this definitely has limits, but like I said, it's being rolled out in so many different ways that it's really starting to make an impact. It's go, it is just gone from a buzzword and there's research and here's a new paper to like, oh, here's this product. Oh, here's another product. And I think that's why it's just getting so much attention, right? I mean, if you look at stable diffusion, looking at video, there was just that, I don't know if you guys saw the video going around with DeSantis in the off playing in the office on the show. They replaced Michael's face with DeSantis face. And it's, it's not a hundred percent, but it looks really pretty good. And so, you know how this, these kind of things affect, you know, elections coming up and everything. It's gonna be, it's gonna be real, you know? So I, I think that's the, the, the threat, the idea though that these things are thinking and all this, it's, you know, we've still got a long way before that happens.

Leo Laporte (00:32:56):
So this is, by the way, it was tweeted by Donald Trump Jr. So, I don't know. Furthermore, I, I don't know. Oh, that looks too, it's pretty uncanny. I mean, it is Ron DeSantis. Wow. That's that's that's a deep fake. That's really really good. Does it have his voice? I I don't wanna turn on the sound cause I know I'll get taken down. You still

Glenn Fleishman (00:33:18):
Yeah.

Daniel Rubino (00:33:19):
It's like, it, it's his voice that's doing Michael's blind.

Leo Laporte (00:33:23):
Okay. Let me, let me, let me,

Ron (00:33:25):
You know what, <laugh> Damn. Would you please tell Daryl that this is not a woman's suit? Oh my God. That's a woman's suit. You're wearing a woman's suit. No, I do. I even, I wear men's suits. Oh,

Leo Laporte (00:33:37):
It is, it is Ron's voice.

Daniel Rubino (00:33:39):
Yeah. Yeah.

Leo Laporte (00:33:40):
Oh, we're in a deep, we're a deep fake heart here, <laugh>. This is, this is just the beginning. Wow. Yeah. You know, this

Glenn Fleishman (00:33:47):
Is

Daniel Rubino (00:33:47):
Fun. Two years.

Glenn Fleishman (00:33:48):
It's the fun part though. That's, I think why it got so much tension is it is a great demo, right? And things that have a great demo you wanna show off and and then people get bored of them. What was the thing? The JibJab that thing that JibJab love JibJab. Yeah, yeah, yeah. And they created some tools where you could make your own JibJab. Yeah. It was like fun for 12 seconds, right? Because the very specific style, or everyone making Simpson, I subscribe spaces of themselves a

Leo Laporte (00:34:10):
Jibjab for a year and use it for a week. <Laugh>

Glenn Fleishman (00:34:15):
It's, but it's, I I think we're seeing the hype curve, you know, is starting to, to fade on it. But it's, there are things like, I'm working on a website right now and I have not found a graphical tool. I like, I'm so old. I remember good graphical tools. They don't exist anymore. And I've kind of forgotten a lot of my HTML c s s and JavaScript. So I'm using chat G p T four. I'm like, Hey, I've for, you know, don't say I've forgotten. Hey, dear Mr. Ms Java chat gtp. I say, you know, how do I make a box that is centered? How do I do this? Whatever. And it spits out, I mean, this is like co-pilot, right? But I'm asking kind of very specific questions. I had a problem on a page. I described it natural language, and I said, why does the page move back and forth when I've got this with this C Ss property? He said, oh, you need to do set overflow property or whatever. And I was like, oh my God, that's it. But I gave it this like paragraph long description of a problem and a I ha again, it's a verifiable solution. It said, this is the thing I tested and it worked. I didn't say like, I've, you know, I'm a hemophiliac and I'm bleeding. What do I do? That's probably not a good use, probably not a good idea by some people's use of it. That's an example. You

Leo Laporte (00:35:20):
Know, I look at this video Don't

Glenn Fleishman (00:35:21):
About me.

Leo Laporte (00:35:21):
I'm Okay. We are heading into an election. This, originally it was created by C c3 p meme and posts on Twitter. We're heading into election where this what is

Ron (00:35:30):
Rightfully through them crazy. And I grabbed one,

Leo Laporte (00:35:33):
This is gonna be,

Speaker 6 (00:35:36):
And it

Leo Laporte (00:35:36):
Fit a problem. I don't, I think this is really gonna be a problem.

Daniel Rubino (00:35:41):
This is funny, but this is funny. It to be difficult for people to tell the difference. I mean, I mean, think of, think of someone in their seventies on Facebook, seeing watch it, these videos.

Leo Laporte (00:35:50):
Watch it Buddy <laugh>.

Daniel Rubino (00:35:52):
I'm just saying they're not gonna see the difference for them. The young Kenney Valley doesn't exist. They're gonna believe in No, you're right. And they're gonna, you know, and if it's a serious thing, you know an event that's happened, you know, there was a Right, we'll get to that story, right about the was it the image that went around the AI generated image of the Pentagon? Yeah. Let's,

Leo Laporte (00:36:11):
Let's take a break. I wanna come back and I still wanna really talk about what Microsoft announced because I think these are all yeah. Really important things. But I need to take a, a little break. We've got a great panel. It's so nice to have you Daniel covering the, the water, the build. But everything else, Glen Fleischman Owen, I see you. I see you, man. And your beautiful painting, which was smuggled into the United States illegally. But we won't mention, we won't mention that. There is a story there though, right? He got it through customs.

Owen Thomas (00:36:43):
Oh, yeah. Oh

Leo Laporte (00:36:44):
Yeah. <Laugh>.

Owen Thomas (00:36:46):
It was, it was work documents.

Leo Laporte (00:36:48):
Why do you have to smuggle a painting in though? I don't understand.

Owen Thomas (00:36:52):
Oh, I, I, I forgot.

Leo Laporte (00:36:54):
Why did you just say maybe didn't want to clear its value? But

Owen Thomas (00:36:58):
I don't know. Yeah, I think it was just, it was, it was just, it was,

Leo Laporte (00:37:02):
It's just a painting man,

Owen Thomas (00:37:04):
Basically. That was the, that was, that was the advice given at the time. Yeah. When the was, was required.

Leo Laporte (00:37:09):
I know, I know, I know. I almost brought a Spanish ham in after a visit Barcelona, but I was talked out of that. That's the good thing. Our show today brought to you by ACI Learning, they sponsor our studio. We're big fans. You may say, well, who's ACI learning? I know you know the name it Pro. They've been with us since they started in 2013. It PRO is now part of ACI Learnings Family IT Pro's capabilities continue to grow. They have those great highly, there's their studio have been there in Gainesville, highly entertaining, literally bingeable short format content that teaches you IT skills so you can get the certifications so that you can get the job. There's over 7,000 hours now in their library, and they're creating new content every day, Monday through Friday. And by the way, great content for enhancing your existing skills if you've got an IT team for building their skills.

(00:38:08):
If you're an Ms P, you must love it. Pro we know IT. And ACI learning. 30% of ACI learners are managed service providers. ACI Learning is dedicated to supporting your s p team through any challenges. An Ms. P, you're selling the expertise of your team, their ability to solve problems for your clients. They will love, they will love ACI Learning. For instance, the ACI Learning Practice Labs a place you can test and experiment before deploying new apps or updates without compromising your live system. That's pretty cool. Try out your skills on Virtual Machine Labs. You could just do it with a browser and, and be running Windows Server desktop clients. You can run it anywhere. Mac os Lennox iOS Windows. You could prepare for challenging certification examinations with practice exams. It's always best to take the test before you take it. For real. You could take and retake tests to make sure you're ready to sit for the certification exam.

(00:39:08):
One, happy Ms. P Team leaders told us I had 110 engineers in the field. Wow. We had dozens of IT PRO accounts last year alone, they passed over 40 certs. That's real success. If your IT training isn't raising your team to the level you need, you need ACI learning. And by the way, ACI learning's content is so bingeable is so high quality. The industry average completion rate on these on videos from other companies, around 30%, 70% of people bailout with IT Pro it's an over 80% completion rate. Why? Cuz these are great. They're entertaining, they're fun, and you're really learning. Don't settle for subpar training. This is, this is the format IT pros want. And with their pro portal, you can assign courses, you can manage seats, you can unassign and assign team members. You can even create custom programs for your team. It's a great way to, in fact, a much needed way to stay compliant with regulations to identify potential risks and weaknesses before they're a problem.

(00:40:12):
To future proof your business to retain top talent, it's a great benefit for your IT team. Upskill that team gain essential insights into training for individuals, teams, and leaders. While other team training companies may not comply with regulatory requirements, you'll be glad to know ACI Learning is ISO certified. So it, it, you are getting the world class training your team deserves. Try it for yourself, then bring the whole team along for individuals. Use the code TWIT 30 for a 30% off on a standard or premium individual IT pro membership. You can even get volume discounts for teams from two to 1000. The volume discounts start at five seats. Look, the best thing to do is learn more about ACA learnings premium training options across audit, across it across cybersecurity readiness@go.acilearning.com slash twi. Go dot aci learning.com/twit. Fill out the form there. You can learn about a free two week trial for training your entire team.

(00:41:12):
Go do ACI learning.com/twit. Somebody asked, I should mention unfortunately, we are not sponsored by Cognizant, by Aramco, by any of the companies on my <laugh> on my shirt or my hat. I, the boss does not sponsor this show. I'm just celebrating a very nice race from my my man Fernando Alanso in the Formula one. I th you know, you wear a team jersey during football games. It's okay, right? I can wear a fire suit if it gets spicy in here. I'm ready. All right, let's get back to ai. You were, you were mentioning the picture. It was, you know, it's funny because I this picture of the Pentagon on Fire, which was, you know, obviously a fake Pentagon was not on fire. The New York Times blamed AI for it, and I really think that was kind of a mistake. It really was Twitter that spread it. Yeah. Right? Sure. you could have done that in Photoshop. Was that an AI thing or a Twitter thing or a, or social media thing?

Daniel Rubino (00:42:32):
Yeah, that's true. I mean, they probably did latch onto the AI for the buzz, but yeah, you're right. I mean, this is the danger of social networks, right? I mean, we saw that <laugh> Yeah. And when they had the verified accounts and, you know, well, was it, oh, that one, one was trying to be a, someone, a pharmaceutical company said they're gonna give away insulin for free. Well,

Leo Laporte (00:42:50):
That's exactly what happened here. This was a fake Bloomberg account, right. That had the blue check tweeted this picture of the Pentagon burning and bec and the stock market plummeted as a result. You know, I think CNN got it right. Verified Twitter accounts share fake image of explosion, causing confusion. It Donio Sullivan knows what he's talking about. He's, he's, he's great. He's a smart guy who was banned from Twitter for a while AI generated. Yeah. But you know what, if you look at it, it's not a, again, we, 70 year olds might not know that that fence is, is is all munged up, but it's pretty obviously fake, whether it was Photoshopped or AI is kind of irrelevant. The real issue was it was tweeted out by an account that looked like it was Bloomberg with a blue check.

Owen Thomas (00:43:42):
I mean, the, the, the issue right with AI is that you could, you could imagine AI being used to generate like 10 different angles, right? Of right. You know, of the same scene. Like, you know, there's an efficiency gained over, over human Photoshop. But yeah, I think the, you know, I, I, I think the information environment is more the issue here. Like our Archie Russia today was one of the accounts that shared it

Leo Laporte (00:44:08):
A legit, by the way, the legitimate RT account, cuz they loved it. <Laugh>, right? <Laugh>, yeah. The, the, the

Owen Thomas (00:44:16):
Legitimate share of legitimate misinformation, right? Shared actual made up you know, right. Made up misinformation,

Leo Laporte (00:44:23):
Right. At stock market, Dow fell 80 points in four minutes, but it fully recovered because it became pretty obvious it was a fake, right? The, the, it took the fire department from Arlington, Virginia to tweet that it and the Pentagon Force Protection Agency. <Laugh> said there is no explosion or incident taking place anywhere near the Pentagon, no hazards to the public. And then the stock market recovered. Yeah. To me that's really more about blue checks and you know, that's something people have to learn too. Blue check on Twitter. It's gonna be doesn't mean the same thing is it used to.

Daniel Rubino (00:45:07):
Yeah, no, that's definitely, that's a whole other thing. But like, I would say with AI too, what's interesting about AI is of course you could generate these images, but you could also use AI to detect these images, <laugh>. Right? Right. Like there are ways to do that, to analyze the photos and videos to and what should happen is all of these companies, Microsoft included, should be investing in technologies that do that as, as part and par parcel of making these other technologies. And what social networks need to do, and maybe be legislated into doing, is implementing these safeguards into the system so that they're automatically detected and flagged as AI generated. Like this is gonna be possible. This is not out of the realm of, you know, like science fiction. But right now, I mean, Elon Musk doesn't care about any of that. He, he, he has a, a meme on his Twitter account right now attributed to Aire.

Leo Laporte (00:46:06):
Oh yeah. It wasn't Volera. I was a white supremacist

Daniel Rubino (00:46:09):
And Yeah. And, and all his followers are like, yeah, but the quote's good <laugh>.

Leo Laporte (00:46:13):
It's like the quote's not good. The quote is always, I know so wrong <laugh>. In fact, when I read it and I didn't know it wasn't Voltaire, I thought, boy, Voltaire really got that wrong. <Laugh>. I have to rethink my whole idea of the father of Enlightenment. I mean the age of reason. It didn't

Owen Thomas (00:46:31):
Didn't Elon Musk buy Twitter to, to conquer the bots and it feels like the bots are

Leo Laporte (00:46:36):
Conquering you? No, he bought Twitter to do exactly what he is doing, which is to shake, to share fake Voltaire quotes from neo-Nazis. The quote by the way, is to learn who, learn who rules over you. Simply find out who you are not allowed to criticize. And you know, if you're not paying, it's

Daniel Rubino (00:46:56):
Hilarious on Twitter, <laugh>.

Leo Laporte (00:46:57):
Yeah, right.

Daniel Rubino (00:46:58):
Free speech.

Leo Laporte (00:46:59):
Free speech who you're not

Daniel Rubino (00:47:00):
Allowed to criticize.

Leo Laporte (00:47:01):
Then of course, it turns out really that it wasn't VoLTE, but this was a neo-Nazi that tweeted it. And actually the first reply was, whoa, well I guess we can't criticize children with leukemia then. So you know, that's obviously a specious quote. It doesn't make it's, you know, well, I think Voltaire would not have said that, to be honest with you. <Laugh>. yeah,

Glenn Fleishman (00:47:24):
I think what was his was I disagree with what you say, but I will fight to the death. You right to say it something. There

Leo Laporte (00:47:29):
You go. Those lines. That's free speech. Yeah. Right. That's

Glenn Fleishman (00:47:32):
A little bit volter. That's

Leo Laporte (00:47:33):
A, by the way, this comes on the heels of Twitter pulling out of the EU disinformation code. Elon has essentially said, no, we loved it. What? No, what's wrong? Nothing wrong with that.

Daniel Rubino (00:47:45):
Voluntarily too. It, it wasn't even like a commitment, right? That kind of agreement. Like they weren't forced to do anything. It was just saying we're, we'll try <laugh> to do stuff. That was too much.

Leo Laporte (00:47:55):
It, it doesn't matter because after August 25th he used internal market commissioner said, doesn't matter, we're gonna enforce this law. So you're gonna have to do something about this information anyway. We, well, we'll

Daniel Rubino (00:48:12):
See, and that's the right approach. I dunno if you guys see, I put in the show notes, there was another funny AI thing that went around, I think it was last night, with politicians cheating. Someone generated a bunch of images of famous politicians, oh

Leo Laporte (00:48:25):
God.

Daniel Rubino (00:48:25):
Cheating on their spouses.

Leo Laporte (00:48:26):
Okay. I just wanna tell you, this is fake. This threat

Daniel Rubino (00:48:29):
Is

Leo Laporte (00:48:29):
<Laugh>,

Daniel Rubino (00:48:30):
This threat is something

Leo Laporte (00:48:31):
Else. This is not real. Okay. I'm just saying, oh my God, this is not, and you know what, you can kind of tell these are not real, right? Yeah. Right. Yeah.

Daniel Rubino (00:48:42):
But, you know, give it a couple more years. Well,

Leo Laporte (00:48:45):
And you and I can tell impressively, but they're pretty good. They're

Glenn Fleishman (00:48:48):
Impressively a they're impressively awkward. There's something really awkward and awkward. Yeah. Bernie

Leo Laporte (00:48:52):
Doesn't look like he's having fun, does he? No. No. It,

Glenn Fleishman (00:48:54):
It makes it seem Is Joe Biden cheating with his own wife? That's what it

Leo Laporte (00:48:58):
Looks like. <Laugh>

Daniel Rubino (00:48:59):
Really <laugh>

Glenn Fleishman (00:49:00):
On his wife. Nothing.

Daniel Rubino (00:49:01):
He doesn't remember <laugh>.

Glenn Fleishman (00:49:02):
It's like, it's just, they didn't make it prettier. If it was prettier, you wouldn't believe it. It's that, it's so, it's so bizarre. Like Skeezy.

Leo Laporte (00:49:10):
Yeah. Yeah.

Glenn Fleishman (00:49:11):
Skeezy. Mm-hmm.

Leo Laporte (00:49:12):
<Affirmative>. Yeah.

Owen Thomas (00:49:12):
You know, what is what is actually very, very well executed and I believe it I, I believe they used AI for this. There's an account called RU Republicans and it's Republicans

Leo Laporte (00:49:25):
In drag dragons drag queen. Oh wow. Yeah, just

Owen Thomas (00:49:28):
Google ru like RuPaul Republicans and and it should pop right up. It's, it's really excellently done.

Leo Laporte (00:49:36):
I will leave that as an exercise for the listener. How about Yeah, <laugh>? Yeah, it's on Insta. Yeah, actually I might follow it. Just, it's a good thing to have around <laugh>. Okay. It is pretty funny. <Laugh>. Okay, I'm sorry. Don't. No, no. <Laugh> don't. The top five announcements from Microsoft Build that you need to know. Alright. it was really an AI show, right? Chat, G P T definitely gets being integration. That's not new, is it?

Daniel Rubino (00:50:14):
Well, so yeah, actually kind of is. So the problem with chat G p t was, it was, you know, locked off basically to the end of 2021 for information. And so if you asked about like the basketball game last night with the Celtics, it would've no idea what you're talking about. Whereas Bing does, Bing has access to the search engine. And so that's the difference between the two. Otherwise they're effectively the same. So what happens now is there's a plugin and they're gonna allow basically chat G p T to have access to Bing. So that becomes more effective. So the difference is between chat g, pt, Bing are really blurring now, which is good for Microsoft because users don't need to choose between them and simp with the plugins.

Leo Laporte (00:50:53):
Oh, see, okay. So I read the I read the headline backwards cuz chat, Bing had already had chat G P T integrated, right? This is now chat, G P T getting Bing integrated. So it has more UpToDate search results, not just Bing either. Quite a few extensions you know, they talked about Open Table, which has been around for a while and Zapier, but they're adding quite a few extensions and this kind of makes sense because now like Wolf from Alpha, that's not new, but that's in there

Glenn Fleishman (00:51:25):
In Instacart. I thought the Instacart was out before. I thought you could

Leo Laporte (00:51:29):
Use Instacart. They announced it. You know, I pay for chat G P T, but I not everybody's getting all the plugins, I don't have access to any of this. Yeah.

Daniel Rubino (00:51:36):
So what happened was chat G P T announced a plugin support week ago, it's like six weeks ago. And then Microsoft announced plugin support a couple weeks ago. And then at Build, what they basically announced was, Hey, our plug-ins, they're interoperable. Oh, oh, they build 'em for Bing, they work on chat g pt, build 'em for chat. G P T, they're gonna work on Bing. Which is great because as a developer, you don't wanna have to choose between these two systems if you have limited resources and you're not gonna have to do that anymore. There is, they're gonna be the same system, which is gonna be really beneficial for Microsoft here because they're going to, you know, reap the benefits.

Leo Laporte (00:52:09):
Yeah. That was, you know

Owen Thomas (00:52:09):
What, this makes me,

Leo Laporte (00:52:11):
Go ahead Owen.

Owen Thomas (00:52:12):
This makes me think of Facebook platform. Like, yeah. Yes. There's gonna be a lot of excitement, a lot of buzz probably, you know some startups are going to get a hundred million dollar plus series A evaluations because they build chat G P T plugins and then we're going to have the privacy violations. Then we're going to have the, you know, runaway spam. Then we're going to have the, you know, the ethical problems and open AI is going to say, oops, we really should have thought about all of that.

Glenn Fleishman (00:52:43):
I'm just switching Task Rabbit Instagram or Insta Instacart. Like, I wanna see the mash of this where you tell it something and your whole life is ruined. <Laugh> cause it out and it reaches into the real world. It books, you trick tickets to Mexico, it hires disposal company.

Leo Laporte (00:53:00):
800 pizzas delivered. Yeah, yeah,

Glenn Fleishman (00:53:02):
Yeah. Orders of sulfuric acid. You know, remember

Owen Thomas (00:53:04):
When Facebook started spamming your friends with all of your e-commerce purchases? My God. And like, God, like it ruined someone's proposal because wow. You know, it posted like, so-and-so just bought an engagement

Leo Laporte (00:53:15):
Ring. Wow. Mm-hmm. <Affirmative>. So yeah, I saw that in fact, platform. You said that that it's like the Ai pla Facebook platform that's in fact such a even explicitly addressed this. He, it

Daniel Rubino (00:53:26):
Is a platform

Leo Laporte (00:53:27):
He brought up bill Gates's definition of a platform, which is a platform is something that the people who use it make more money than the platform makers make in effect. Right? Microsoft created, Microsoft knows better than anybody. The, the, the power of a platform. It created Windows and the people who developed on top of Windows did very well. And, and I guess made more money in aggregate than Microsoft did, but everybody benefited from that. So Microsoft says this is a new platform, essentially. Yes.

Daniel Rubino (00:53:58):
Yeah. Oh yeah, absolutely.

Leo Laporte (00:53:59):
Now they were able to get this plugin format co compatible with chat g p t, cuz they own 47% at chat G B T. Right?

Daniel Rubino (00:54:09):
Yeah. I mean, who knows what's ex exactly. In those agreements. But yeah, obviously the relationship, and I know that fathers Elon Musk, but I think it's just more sour, you know, grapes because he pulled out, he

Leo Laporte (00:54:20):
Founded it and pulled out cuz they were gonna commercialize it and he didn't want 'em to commercialize it.

Daniel Rubino (00:54:25):
Well, the, if you go by semi four, they said he pulled out because he wanted to take over the company. Yeah. He wanted be put in charge. That a lot more

Leo Laporte (00:54:32):
Sense. Yeah, that makes a lot

Daniel Rubino (00:54:33):
More sense. And so he was like, all right, well, I'm gonna go, but then he said, I'll still give you some money. And they're like, all right, cool. And then he didn't give the money, and then all of a sudden they want to scale up because Google was doing these large LLMs. And then that's where Microsoft came in. So Microsoft's like, here's a couple billion dollars. And said, that's how it all happened. So technically this is, yeah, it's all Musk's creation. It's its creation from the beginning of open AI and it's creation where it is now because of his actions. Yeah. So

Leo Laporte (00:55:01):
Microsoft they said put in 10 billion, but I think we learned this week that the 10 billion was really in kind, it was a, it was Azure time. But that's one thing that Yeah. Open AI really needs. It's very expensive to generate these large language models and they're doing it all. And

Daniel Rubino (00:55:18):
They have a super computer. Yeah. Yeah. The supercomputer was that was 2020, I believe that Microsoft and OpenAI announced that they're gonna use the that to train their LLMs. Right. So that's a, that's a big difference here with Microsoft versus some of the other, you know, that's why I worry about Apple. Curious what Apple's doing. Cause they don't have quite have this infrastructure. Oh,

Leo Laporte (00:55:35):
We'll get to Apple because Apple's coming up with a big announcement. And oh yeah. And I'll be very curious how they respond. You know, we've been really watching Google and Microsoft go back and forth over ai, right? There was Google io where it was AI every three seconds. I have to say Bard, Google's chat engine, search engine is, seems a little bit laggard compared to chat G P T, but Google's now integrating it into their own search with Sge. It's a Google Labs thing. So there's the, and then Microsoft comes back with their own volley of AI applications. So that was the first one. B integration. Bing Chat is boosted with plugins. Oh, here's the plugin slide. This is, this is the one with, with all of the plugins from Adobe to Zillow and everything in between Spotify. Yeah. Yep. Yeah. They did the, they said the Spotify thing, like four times <laugh>. It was really, we, we fit, we did we covered both keynotes with Rich Campbell on Day one, and Rich and Paul Thro on day two. And both times they announced that you could generate a Spotify playlist. Like this is the number one thing people want to do. <Laugh> developers. Yeah. Right.

Glenn Fleishman (00:56:52):
You wanna buy a personal computer so you can do recipes, you know? Oh yeah. That's the primary case.

Leo Laporte (00:56:56):
I remember that. Everything. Balance your checkbook. Right? Balance

Glenn Fleishman (00:56:59):
Your checkbook. I mean, I'm still using Quicken, but you know, I I, I'm sure this has been talked about before, but it's, this is continues the culmination of like, Google search results are terrible. And I wonder if one of the reasons that Microsoft's been able to have so much place room to run on this is that, I mean, I am sick of Google results. Like I, even if chat g p T is making stuff up, it's better than when I got outta Google. So I think Bing has an incredible opportunity if they get it. Microsoft doesn't mess it up with all the, with the extensions, with the whole approach that it could dramatically leapfrog. I mean, again, I am completely concerned about the shape of inaccurate results, but at the same time, I get an active results on Google. You, you search in, you search in almost anything on Google now.

(00:57:42):
And often if the first or second result isn't good for me, I have to go down pages or do a ton of refinement. It can take me five or 10 minutes to shape a Google search that actually gets me past the spam. And that is unacceptable. And I also don't blame them because I'm sure the, you know, there's a war in heaven that's always going on between the spamming search results and, and not <laugh>, but it would be great, it would be great if this is one way to cut through it, is that you can't spam the LLMs the way you can at Google Search Index. I,

Daniel Rubino (00:58:11):
It's interesting. I've been hearing that a lot too, you know, and obviously I don't use Google Watch anymore myself. I haven't been quite a while.

Leo Laporte (00:58:16):
You used, you use Bing now, Daniel, is that my, your practice? Yeah,

Glenn Fleishman (00:58:19):
I've always been using, he's Alt A person.

Daniel Rubino (00:58:20):
Yeah. Alt, yeah. No, I've been using Bing for, for many, many years. But

Leo Laporte (00:58:25):
Cause you

Daniel Rubino (00:58:26):
Have, it's, it's

Leo Laporte (00:58:27):
As a job op occupationally, but no, you prefer it.

Daniel Rubino (00:58:30):
I, I actually prefer, and plus I get, you know, there's the, the credits you get on Bing. So I have like 230,000 credits <laugh>. Oh my gosh. And I can go and redeem those for like, or something. I, yeah, I can, I can redeem them for Amazon cards and Spotify cards and Oh man,

Leo Laporte (00:58:45):
I, those big rewards are better than I thought.

Daniel Rubino (00:58:48):
They, they're really good. I mean, it literally does pay you too. And simp with shopping. You got money back. I'm shopping too. So it's really good. In the US I think Microsoft's problem is outside the US it starts to not be as good where Google attempts to have a little bit better global presence. Mm-Hmm. <affirmative>. But I've been hearing this from a lot of people that Google's search results have just been not as good. That said is not a ton of e evidence currently that Microsoft is gaining any significant traction with thing, even with all this hype. They were around two to 4% before. They may be slightly higher now, I don't know, they

Glenn Fleishman (00:59:23):
Launching Pad, this is what I, I mean this is what I, it's exciting to think we could have another search engine war, and it was actually valid. It's like we need, Google has let, its, Google's let itself degrade to what extent it's, its direct responsibility or not. And it'd be great to have great to have a real matchup.

Leo Laporte (00:59:38):
Again, I, I'm very sad because I was using Neva and they just sold and, and they, they're shutting down. I was paying for search. So let's, let's, let's put a pin in the announcements at at Build, cuz I kind of want to, you said something I think that some listeners are, are gonna say what Bing search, I mean, Google search has gone downhill.

Glenn Fleishman (01:00:01):
Don't, yeah. Don't you think,

Leo Laporte (01:00:01):
Do we all agree that to that? That's,

Glenn Fleishman (01:00:03):
I've seen it dramatically. It's what I mean, I don't know how other people feel. Well, you're not using it, Daniel. So you're,

Daniel Rubino (01:00:09):
I use that occasion, but I, yeah, what I've seen people complain about too is all the ads and you know, basically can type in something. Do

Leo Laporte (01:00:17):
You use, what do you think is, is that premise acceptable? Do you Google's search results aren't good anymore.

Owen Thomas (01:00:27):
I, I mean, for commercial results, like, I, I remember looking for you know, looking for a hotel room last fall. And I just, I, I finally found a site that had a good discount on a hotel I wanted to stay at. And I ended up picking up the phone because I just did not trust that the website Google was serving me was, you know, was reputable. You know, it's not, it wasn't Expedia Enterprise line or one that I'd heard of. And I think that, you know, I think that a lot of a lot of searches are kind of degraded at this point. Let

Leo Laporte (01:01:00):
Me, let me, let's do an experiment. Let's do an experiment. So I'm gonna search for Tina Turner, the late Tina Turner passed away this week. This is the Google search for Tina Turner. It says 394 million results, 0.61 seconds. All of all, all, only one of the results is on this page. They're all below the fold. In fact, way below the fold. And this is, it's a way

Glenn Fleishman (01:01:26):
To get away from non-authoritative as you put the stuff that's very authoritative that you've curated. It's Yahoo. It's a Yahoo portal above Fold.

Leo Laporte (01:01:33):
Yeah. There's pictures of Tina Turner. There's one article to the USA today. There's when she was born, when she died, here's a video from YouTube, a Google property, and then the Wikipedia result. Plus a panel that shows where you can listen to her music on YouTube, Spotify, Pandora or YouTube. Music, a knowledge graph about Box, which is crib directly from Wikipedia, I might add. So it's really the same as that Wikipedia result. People also ask, and now I'm below the fold. There's songs people also search for. There's more YouTube videos. Finally, I get top stories. Now Bing, let's look at Bing. It's all, it's links right away. There is a knowledge graph. There is a top searched artists. I feel like

Owen Thomas (01:02:21):
Yeah, it's, it's not, it's not much different.

Leo Laporte (01:02:22):
Well, but I feel like it's more, they're more links here. I mean, at least above the fold, there's 1, 2, 3, 4, 5, 6, 7 pages that are linked. You know, they're all obituaries. But that's, you know, that's kind of, actually they're not all obituaries. That's kind of a good, I think that's a better result. I, I would show you Neva, but I think, I don't think they exist anymore.

Glenn Fleishman (01:02:45):
No. The, the worst one is you search for a how-to thing. You're like, you know, why is my iCloud drive so slow? A question I've asked a lot lately. Yeah. And I think so, and it's, and you get maybe one or two good results, and you get hundreds of pages that are optimized. And they're often have, you know, they're clearly maybe pre AI generated text that was copied and pasted

Leo Laporte (01:03:04):
As generic.

Owen Thomas (01:03:06):
I, I, I'm gonna ask a question that I, that I will then answer, which is why is every how-to question that you search for a YouTube video and I'm pretty sure the answer is money. Because YouTube monetize is better than like creating a, you know an article of how to,

Leo Laporte (01:03:24):
I don't wanna watch a video. I really don't. Exactly. It's inefficient.

Glenn Fleishman (01:03:33):
I but it's more,

Owen Thomas (01:03:34):
It's more profitable.

Leo Laporte (01:03:36):
Oh yeah. Apparently it's good for YouTube.

Glenn Fleishman (01:03:37):
I tell you how I made $1,200 off YouTube by accident. I I made a Air, air airport utility run a walkthrough once years ago. <Laugh> posted it on YouTube and I made $1,200 off without even knowing I was getting the money. Cause it was coming in in tiny amounts. They're like, well, you, you've got a lot of views, you can monetize this before they impose the things I said, yes. And I already had a Google Ad Sense account, so it was just depositing small amounts over years. And one day I looked at where it was like, oh, here's how much money you made. I'm like, how did I make $1,200 off that? So that's,

Leo Laporte (01:04:07):
That's why that's pretty hysterical. That's why thank you. Yeah. For years. John c Devork said his number one YouTube video was one he made of a Barney Doll singing in his bathroom floor. <Laugh>. I'm just, I'm looking for his, I don't think he has a channel anymore, so I know. I can't find out. I bet you it's still number one though, <laugh>, I have to say. So what do we do? What do I mean that's not acceptable? Is is there a a, a replacement coming? Is bing the right right way to go? What is, I guess I'll just start using Bing.

Daniel Rubino (01:04:43):
I think Bing still has its, you know, hangups with people. They think it's kind of a joke, but it is it, you know, I, I think Glen's right, right. If there's an opportunity, if there's gonna be one opportunity for Bing, this is it. There was also the talk that Samsung may adopt Bing as its default browser, but looks like <crosstalk> they changed

Leo Laporte (01:04:59):
Their mind.

Daniel Rubino (01:05:00):
Yeah. Yeah. But stuff like that would be needed, you know? Cause even though most people would still switch away the, the problem is, is people are like, once you're hooked into Google, it's just like browsers. You, you know the thing. But <inaudible> making a, a good effort here. I mean, they're, they're doing all they can. It's a changing sentiment. There's two things that I think are always required in these situations, right? One, and we almost have them. Google is gonna be required to basically really trip up on itself and fail and annoy people enough that they wanna move away. And then there needs to be a really good alternative for people to go to. And we're kind of almost there, but we haven't seen the, the actual push yet. So

Leo Laporte (01:05:41):
I tried

Glenn Fleishman (01:05:42):
Meel Meyer has,

Leo Laporte (01:05:43):
Go ahead.

Glenn Fleishman (01:05:45):
Meel Meyer has answered the question, by the way, of why never former Yahoo, c e o, but also at Google for many years before that. And and she just says that Google results are bad because everything is bad on the web <laugh>. And I'm like, alright, she has a point. It's likes lot. That's fraud today. Yeah. The 20 years ago. The, you know, it, there's just a lot more eco, she says, I think because there's a lot of economic incentive for misinformation, for clicks, for per purchases. And this all makes sense. And this is on an interview with Freakonomics Freakonomics podcast where they asked, is Google getting worse? The answer was, yeah. And they talked to her because they figured, you know, she's actually a slightly disinterested third party now not being at the company and having led it to where it was. So,

Leo Laporte (01:06:26):
Or what do you, what do you use Glen as your phrase to test search engines? You said, how do I back up my

Glenn Fleishman (01:06:35):
Oh, yeah. Or, you know, hotel, you know, what's the best hotel and whatever what what's the best you know, gas grill i, I look at for things like that? Like, like, they're generic enough. And you should get authoritative results because think how many, there's 20 publications that have really good roundup guides on gas grills, and often, not always, but often the top results for things like that are, are polluted by, in my, and maybe I have terrible results, but in my results are polluted by just the this is the best, blah, blah, blah. These, you know, nonsense things from sites that don't ex don't have any credibility, but they show up in Google results, right? So at least water cutters showed up there. The top Oh, always See you. This

Leo Laporte (01:07:14):
Is not Google. So I wanna, I wanna say this is not Google. Oh, this is

Glenn Fleishman (01:07:16):
Not

Leo Laporte (01:07:16):
Google. This is, so what's this? So after Neva went down, I've been looking for replacements and I'm trying one called kgi, K A G I, I think Jason Snell recommended it. It's not free. Neither was Neva. It's five bucks a month, but therefore no ads. And in theory, I guess no need to monetize and this, these are good results, right? This you would agree. These are the results you'd want.

Glenn Fleishman (01:07:40):
Yeah. I'm doing a Google search right now, and I'm getting actually very good gut results for gas grills, actually way down before I start getting, maybe that's the bad one. It's, you know, I do a lot of howto writing, and so when I'm trying to get an answer, I will search forums and things like that. So I start on Google, like to see if anyone's posted anywhere a similar problem. Cause sometimes you'll find somebody writes in with a question and you'll find one other person in, you know, 2008 had this situation, they described it maybe even how to fix it. Well,

Leo Laporte (01:08:06):
That's why people do, there's site colon reddit.com, right? Yeah. That's, that's the best search engine for this kind thing. That stock

Glenn Fleishman (01:08:17):
Exchange thing. Yeah. In

Leo Laporte (01:08:19):
Fact, there's great, if I were looking, if I were looking for the best gas grill, I would search Reddit, not anything, not the wide internet.

Owen Thomas (01:08:26):
You know, it's interesting, everyone from Wikipedia to Reddit to Twitter has been looking to extract more money from Right. The use of their, you know, kind of data corpus much of which is contributed by, you know, ordinary users. But we'll leave that aside. From, among other things ai AI companies looking to train their language models, which I think is smart.

Leo Laporte (01:08:49):
Oh yeah. They're totally scraping Reddit, right? Of course. Yeah, yeah. And stack exchange.

Owen Thomas (01:08:53):
Yeah. And that should not be free.

Leo Laporte (01:08:56):
Yeah, that's a good point. Although as you also point out, we made it damn it, the best, the thing about Reddit is there's, there's, there's 20 people who care more about gas grills than life itself. And those are the people you wanna ask. What's the best gas grill, right?

Owen Thomas (01:09:16):
Yeah. I mean, it, it, it is, it is pretty gutsy for Elon Musk to, you know, look to charge everyone for Yeah.

Leo Laporte (01:09:25):
Oh, he, he is

Owen Thomas (01:09:25):
The, the use of, of our data.

Leo Laporte (01:09:28):
He has really pissed off academics because he not only is charging now, he's saying, if you had Twitter data on your servers that we, you were using for academic research, you must delete it now if you're not gonna pay us for it. And again, that's not, you weren't even there Elon, when most of that data was generated. Yeah. You weren't even there. I guess he wants a 44 billion back. It,

Glenn Fleishman (01:09:52):
What was it? Yahoo Guides. I feel like Yahoo had a pro, had every product, they had so many products that you could find a Yahoo product that did everything. And I think there was a Yahoo Guides or something like that, that had a product experts in different categories who specialized and got paid in some basis before affiliate programs and all that kind of referral revenue. That's the only thing by

Leo Laporte (01:10:11):
Who answers is how is Babi formed? Oh, I love

Owen Thomas (01:10:14):
That. <Laugh>. The,

Leo Laporte (01:10:15):
The,

Glenn Fleishman (01:10:15):
This is the other part is the reason Google results have become poor is because advertising rates went way down. I mean, particularly it was accelerated early in the pandemic, but a lot of publications have found it. I think all of us understand this very well on this episode. I found it difficult for advertising dollars to support most publications. And so some went to subscribers, but a lot followed the wire cutter was a little cutting edge in looking at the how far you could push the, we are totally editorial independent, which they were and are, and the were taking money for referring you to a place in this slightly hands off transaction. And, you know, in, I remember long ago, back in 2000 or something, the Seattle Times said, we're never gonna put advertisements on our cover. And you know, fast forward 15 years of Seattle Times, of course they have ads on their, on their cover. It, it's the same thing. Every site now has affiliate relationships because it's a much more reliable form of income than anything else. So you go on, you're optimizing for Google because you want the

Leo Laporte (01:11:12):
Affiliate from Google so that you can make the money to run your, whatever you're doing. You can't anymore go to unfortunately Yahoo answers, but back in two. No, I know. It's gone. Oh,

Daniel Rubino (01:11:23):
The baby. Whoa.

Leo Laporte (01:11:24):
Back in 2000 r r i p back in 2006 user named Kaia asked the very famous query, how is Babi formed? How Girl Get Pregnant? <Laugh>,

Daniel Rubino (01:11:39):
Someone made a YouTube video like they animated. Oh

Leo Laporte (01:11:42):
Yeah, totally. I mean this is from Know Your Meme and you can find Sight all of the Mimi responses to this my favorite. This is the famous YouTube

Babby meme (01:11:58):
Girl, get Pregnant.

Leo Laporte (01:12:08):
All right, enough. Let's move on to the top five announcements. <Laugh> number three, windows Gets, windows 11 gets its own AI assistant. Microsoft has been using the phrase co-pilot or the brand co-pilot starting I think with GitHub co-pilot, right? That's now two years old. Yeah. and it's now spreading the latest co-pilot is Windows Co-Pilot. What does it, what does it do, Daniel?

Daniel Rubino (01:12:35):
Yeah, so I mean, it's just kind of funny. I mean, we jokingly call it Cortana 2.0 by Microsoft. Yeah.

Leo Laporte (01:12:40):
Didn't they have that in there? Yeah,

Daniel Rubino (01:12:43):
Kind of. Yeah. So it's not that radically different. Obviously doesn't look like Cortana. It's not called that. It does pop out from the side though, and it's gonna be access to the chat being chat as well. So that's not a huge thing. But the bigger deal is it can act upon Windows. You can ask it to do things, you know, put my computer into dark mode and you know, go to do my Windows update. And you can actually do more voice commands that are more advanced to get the operating system to do stuff that you may not have known where it is or whether it could do it or not. And so this is, I think where the first step of where these digital systems may actually become useful. I mean, if you got, like, Siri's been around for so long now and that really is not improved, right? So none of them,

Leo Laporte (01:13:31):
None of them have. In fact, we were talking earlier, Micah and I on asked the tech guys, somebody said I'm in my nursing home. I used to set an alarm with the Google assistant, you know, set alarm for four 50 and then for a long time I could say, how much time is left in the alarm? Now it just says you have an alarm set for four 50, it's gotten dumber. And he said, why is it getting dumber? But this seems like all three assistants have been very disappointed. Serious, serious, serious, serious, terrible, embarrassing. None of them are very good multi-trillion dollar company. What are you doing? But I'm not sure I want chat g p t in there, cuz this gonna go on and on and on. Well let me tell you, <laugh>, you had the alarm, you know, I mean,

Glenn Fleishman (01:14:12):
I'm hoping, I'd just like to be able to say like, I can't Siri currently if I say, you know, send a text message to my child's name, it'll say mobile or email. And I'm like, well I don't know. What's the thing I've used for the last, you know, seven years to use them? Maybe that would be the method. And then if you say mobile or email, it's like, I don't know what you're talking about. Let me

Leo Laporte (01:14:32):
Just ask me. Lemme see what I found on the web at about that. Yeah, that's

Glenn Fleishman (01:14:36):
Right. Email is a,

Daniel Rubino (01:14:37):
This is where I do really look forward to this stuff with machine learning and what we're calling AI is the ability to learn a user's behavior so that it can act upon stuff you, you should proactively, right? It's like, well you usually do this around this time, you know, or like say right now you can get a notification. You have a meeting coming up, cool. Why not? You know, it alerts you did get, launches you into the meeting or gets you set up for it, right? It actually acts upon the thing before you need to do it yourself. I think that's sort of the next level of this stuff that I'm really kinda looking forward to. And you can ask if things probably to, you know, look at your documents. That's be cuz they have copilot now coming out for all this.

Leo Laporte (01:15:16):
Yeah, that's that document summary thing Glen was talking about. I think it'll be good at that. Right? Gosh. They showed an example, a writer who has a bunch of notes and you said, can you summarize my notes on this topic so that I can start my novel or whatever and

Glenn Fleishman (01:15:32):
Oh, that'd be useful. I took 4,000 words of an interview just to see how replaceable I am. I took X only take 4,000 words, <laugh>. So to chat g p d four, write a 500 word article from this using direct quotes and don't use anything. It's not good. It is it good wrote it was not bad. It

Leo Laporte (01:15:45):
Was not, it was not bad. That's, by the way, that's it. It's not bad. It's not

Glenn Fleishman (01:15:49):
Good. If someone could have, if someone had published it, it wouldn't have been entirely inaccurate because it did actually accurately only use quotes from the article. And it used them essentially correctly. It was boring and it didn't have any, there was no reason for the article to exist. It didn't have a conclusive tone to it. Right. And it wasn't analytical, but you

Leo Laporte (01:16:06):
Could add all that. But

Glenn Fleishman (01:16:07):
I was like,

Leo Laporte (01:16:08):
After the

Glenn Fleishman (01:16:08):
Fact, I mean, yeah, I don't, I, I'd look forward to that, to not, to that not being my job. That's why, you know, moving to the,

Leo Laporte (01:16:14):
But you don't want

Glenn Fleishman (01:16:15):
Lucrative field of 19th century print history instead of journalism. It's a much,

Leo Laporte (01:16:19):
But you don't want it to do all the things that you would do. No.

Glenn Fleishman (01:16:24):
For you. No. I, I want it to be an enhancement and I hope it will be. Yeah. And I think, again, as a writer, I have particular you know, we all have particular needs that there are things and, and you know, like transcription is an incredible leg up. Like having a transcript of everything that's, if it's not perfect, I can go listen to the audio and fix it up so it's verbatim in a few seconds. That is an incredible enhancement over. Yeah. You know, I mean that, that alone, like that could save hours on a long article. Very

Daniel Rubino (01:16:53):
Easy. Even with Microsoft teams, you've been advertising the feature where it will not only do a transcription of your meeting, which you expect at this point in technology, but it will also summarize that meeting and provide action points to be

Leo Laporte (01:17:09):
Action items would be great, wouldn't it? Yeah, yeah, yeah. Mm-Hmm.

Daniel Rubino (01:17:12):
<Affirmative>. And so if you missed this meeting, not only do you get a summary, but you get a quick list of the things that needs to be done. That's like the next level of like rarely. Now it's useful.

Leo Laporte (01:17:22):
Can it do that reliably without hallucinating or, or how, how accurate?

Daniel Rubino (01:17:27):
They're rolling it out real soon. So yeah, we'll

Glenn Fleishman (01:17:29):
Find that out. The, the thing that fireflies AI is the transcription service I'm using and it puts action items at the bottom. And they were, again, it's com because it's taking a data input and not extrapolating, it seemed to do a fairly good job. Like, if I were doing, I mean, for an average meeting where you don't need, you would never transcribe, you never pay for a transcriptionist, you never pay for someone to analyze it. Someone might post some notes somewhere. This becomes an incredible company tool. If every meeting can wind up being searchable and have some extracted stuff from it, that, that helps. I mean, think about, I just think meetings are terrible to begin with, but having meetings and then having all the value lost because there's no way to harvest what was discussed except for people's imperfect note taking as it passes through the telephone game of office, you know, stuff like, just a transcript of everything could be helpful. I mean, it might also be bad. It could cause a lot of hurt feelings too, but,

Leo Laporte (01:18:21):
Well, or if you missed an action item that was critical and it didn't know it was an action item, I mean, yeah, I guess it's figuring it out from the words you use or something. But anyway, I wanna take a break semester because there is something we haven't addressed that is really the dirty little secret of all of this. And Microsoft didn't really address it either. I don't think they did. But Daniel Rubinos here from Windows Central. He will, he will correct me if I'm wrong, Owen Thomas, always great to have you from the San Francisco examiner, and we're gonna keep putting pressure on you to bring Dati back cuz we love Dati. And while you're at it, bring back Yahoo Pipes and then everybody will be happy. Yeah. Glen Fleischman. I, I miss Yahoo Pipes so much. I f and miss them <laugh>. That was a

Glenn Fleishman (01:19:05):
Great tool. It's, it's a thing. It's a thing. I think I am trying to think of a piece of technology that people have more nostalgia for than Yahoo Pipes. Partly because it showed a lot of promise and then it was frozen for like seven years before it died, so it never had a chance to decline. It just sat there and Amed, everyone stared at it like a rare fossilized dinosaur bone or something.

Leo Laporte (01:19:26):
Anyway, they killed it. So no point in mourning it. And you should probably learn Python because it's not coming back our showed. Anyway, great to have you too, Glen Fleischman from many, many points. But we will talk about your new book coming out with Marson Witchery that you wrote called Shift Happens. It's imminent. It's imminent, and if you contributed to the Kickstarter, you'll, you'll hear some good news in just a bit. But first word from our sponsor, worldwide Technology, ww t We love WW t. Lisa and I went out there just before the pandemic broke, before lockdown, and we saw the amazing advanced technology center. It, the ATC is at the heart of Worldwide Technology. It's a research and testing lab that brings together half a billion dollars of equipment from the leading OEMs and emerging disruptors, all the companies that enterprises might be using to get business done.

(01:20:26):
Why did they build this? Well, initially, so their engineers could spin up proofs of concepts, could get to know products, could use them in the lab before rolling them out to customers. But there's now a really great feature to the a tc. You can use it to the, a ATC offers hundreds of on-demand and schedulable Labs featuring solutions that include technologies representing the latest advances in enterprise. Everything from cloud to security to networking, primary and secondary storage, data analytics, absolutely ai. If you wanna learn what the cutting edge on AI is, for instance, AI used for business intelligence, they've got it all in the labs. Devops, agile, so much more WWTs engineers and partners use the atc, but you can too. It'll cut your evaluation time for months to weeks. It'll introduce you to new ideas, new technologies. It'll help you do the thing you do best with technology that supports you before you go to market.

(01:21:31):
You can access technical articles, expert insights, demonstration videos, white papers, of course, those hands on labs, all the tools you need to stay up to date. And not only is it a physical lab space, but they've virtualized it. So you don't even have to go to St. Louis like we did to use the ATC platform. Anybody can access these amazing resources anywhere in the world, 365 days a year at no cost to you. So while you're there, make sure to check out WW t's events and communities. We did a great panel event. It was so much fun for more opportunities to learn about technology trends, to hear about the latest research and insights from their experts. We've sent Lu Meeska from this weekend Enterprise Tech out. He did a great keynote last year. Whatever your business needs WWT can deliver scalable, tried and tested, tailored solutions to help you succeed in business.

(01:22:22):
Because WWT understands business. They bring strategy and execution together to make that new world happen. To learn more about worldwide technology, the advanced technology center, and to gain access to all those great free resources, it's simple. Just go to wwt.com/twit wwt.com/twit, create a free account on their platform, and you're g you're in man. And it is really a wonderful place. These are great people. Wwt.Com/Twit. And by the way, please do us a favor and use the the full URL so they know you saw it here, wwt.com/twit. So the, to me, the dirty little secret of all of this is it's hideously expensive and right now it's being subsidized right by Microsoft to gain a foothold into a new market. But that can't continue forever, Daniel.

Daniel Rubino (01:23:20):
No, of course not. It's going to be charged. And I've heard this from Microsoft off the record that, you know, which I shouldn't be probably saying, but like, it's obvious, right? They're gonna charge us somehow that who knows the mechanism, mechanism for it, likely, I would say Microsoft 365, they already have a subscription program that it could come through, you know, it'll start probably with Enterprise since they have deep pockets and typically don't mind spending that. How that trickles down to consumers. There's, you know, Microsoft 365 there where you get a terabyte, OneDrive access to office and all these other perks, you know, they could have another tier, it's an extra couple dollars a month. But yeah, the economics, so the, I think there's two issues here. One, they're currently not charging for it simply because the value isn't there, right?

(01:24:05):
A lot of the stuff we just talked about isn't here yet. So co-pilot's not here yet. It's coming out maybe later this year. Same thing with co-pilot for office. So a lot of the stuff isn't here, so it'd be premature to charge people for it when you're not really getting a lot out of it. You know, you have the big chat, okay, it's cool, but is that worth money? Probably not. They do have big image creator and it does have they call 'em boosts and if you get, you get to boost it, you get a priority on the server and your image renders more quickly because that's the issue here, right? So server time, the server load and who gets access if you pay for chat? G P T, that's one of the things. You get primary access to the server and you, you're, it, it goes really fast at chat G P T if you're paying for it.

(01:24:47):
If you don't pay for it, yeah, you get that lag. So this will trickle down. I think they have time yet to, you know, make sure that there's actual value there that people will pay for. We're not there yet, but you're right. And the other thing they're doing too is they're putting a lot of emphasis on Azure and optimization. So they're really, really trying to, and they've been lowering costs internally on this stuff. And they're going to continue to push that and continue to expect that to happen. So that finally, I think when they do roll this out, it won't be even as expensive as it would be today if they were to do it there. But yeah, obviously they have to charge for the stuff. It is very expensive. A lot of server space.

Leo Laporte (01:25:22):
Well, I mean, when you're talking about building it into Windows 11 and, and we've talked a lot about Windows Weekly really being, maybe that's what Windows 12 will be with, is Windows with ai, right? Then you have a billion people all of a sudden hitting, hammering your servers plus the cost of building these large language models, which is steep. Even though Microsoft owns Azure, it still costs them money to run it. Yeah. but they've, it's your sense. I don't want you to breach your, you know, your, your nda, but it's your sense that they're thinking about this and they think it's gonna be okay.

Daniel Rubino (01:26:03):
Yeah. I, I mean, like I said, they already have a subs, a subscription plan for enterprise and consumers where they could just build these costs into it. I think the question right now is what's the perceived value that people would pay, right? How much would they pay? You can't

Leo Laporte (01:26:16):
Charge 'em. Yeah. And

Daniel Rubino (01:26:16):
Can they reduce? Yeah. Yeah. And, you know, how much can they reduce the actual costs before rolling that out? Glen, you've, but yeah, they're absolutely gonna be

Leo Laporte (01:26:24):
To, you've pinpointed a great use of it. How much would you have paid for that?

Glenn Fleishman (01:26:31):
For just, well, for what? The transcription or for yeah, for the transcription.

Leo Laporte (01:26:34):
Paying the summaries and paying,

Glenn Fleishman (01:26:37):
I'm paying fireflies right now while I'm actively reporting, I'm paying, because I don't always have a big article like that. Right. where have do tons of interviews. I think it's $19 a month, so thousand minutes of uploads and unlimited zoom.

Leo Laporte (01:26:50):
I paid, paid for the, for un for chats with the chat G P T. Yeah, I pay 20 to mid Journey, or maybe more, actually I think it's more like 60 for unlimited images. Do you think they make money on that or

Glenn Fleishman (01:27:05):
I was wondering whether with the fireflies, I'm thinking what engine are they using? Maybe if it's proprietary, like what's the computational right load storage load? It can't be, I mean, I don't know how many people are using substantial amount of the process, but it can't be, you know, AWS services, Amazon services are incredibly cheap per minute. I just don't know. I can't conceptualize it. I guess. I don't know what the, I mean, you can run some of the models you can download and run on a phone, and I know phones are super computers now anyway, but, but legitimately the fact that you can run some of this stuff in very lower low end hardware compared to what you can get at a data center makes me think that the computational cost must be relatively low, but it's not nothing. And when you multiply it by text, right,

Leo Laporte (01:27:50):
It's one, I mean,

Glenn Fleishman (01:27:51):
If I analyze a gigabyte of audio,

Leo Laporte (01:27:52):
If I peg my CPU on an iPhone, well, that's fine. Yeah. But it multiply that by a million people on your servers, that's another matter entirely.

Glenn Fleishman (01:28:02):
I don't know what the growth is. It's a

Leo Laporte (01:28:03):
Whole, Sam seems

Glenn Fleishman (01:28:04):
Like

Leo Laporte (01:28:04):
A Sam Altman, the, the president and CEO of open AI says it's about 10 times the cost of a Google search, a chat GBT query that's a lot more expensive.

Glenn Fleishman (01:28:18):
I did see some fear from the same people who brought you a palm pilot used as a, a free freezer's worth of electricity to operate based on very bad calculations. There's some moral panic that AI will, you know, use more than a bitcoin's worth of

Leo Laporte (01:28:35):
Energy. Well, that's what I'm worried about. We got rid of. So, so we all figured out the crypto was really terrible for the environment. So all the, but all the crypto bros have now moved to AI pros and they continue to burn CPU cycles like crazy. By the way, Nvidia, how much did NVIDIA's stock price go up this week? They became almost a trillion dollar company. And I think a lot of that is because of build

Daniel Rubino (01:29:02):
Yeah, because they they, they're, they need their their GPUs and everything for ai, their, their position, and they do their own ai. So like, you know, Nvidia is positioned really well for the future. They're in automotive, they're in gaming, all these, they're in all every growth category basically.

Leo Laporte (01:29:21):
Here's the

Daniel Rubino (01:29:22):
And they've been releasing

Leo Laporte (01:29:23):
Products. Here's the, the five day trend. This is, well now <laugh>, well now Wednesday, May 24th, right after the Microsoft keynote keynote number two I think that I tr as I remember, they they went up, their market cap went up 200 billion. I think as I remember, it's,

Owen Thomas (01:29:46):
It's kind of crazy that the, you know, the G P U, which was this very kind of specialized niche thing, like mostly of interest to gamers or people who were you know, like computer animators has become this like really desirable computing commodity. Mm-Hmm. <affirmative>. Yeah. And, you know, Intel has badly missed that transition. Oh, no kidding. Mean it's not the only, it's not the only transition that Intel missed, but

Daniel Rubino (01:30:12):
I I would, it's not that bad for Intel, but you, you're right about the gpu. I mean, for instance, the blurring of my background, I always talk about this. This is done by Nvidia. This is Oh

Leo Laporte (01:30:20):
Yeah. You use that that Nvidia technology, don't you? That's really good. Yeah.

Daniel Rubino (01:30:24):
Yeah. So this is using the GPU real time to blur the background. It's also doing my voice in the sense that it filters out my background.

Leo Laporte (01:30:32):
Plus look away from us <laugh>.

Daniel Rubino (01:30:34):
Oh, I do the eye and,

Leo Laporte (01:30:36):
Oh, do the eye. Daniel threw the eye.

Daniel Rubino (01:30:40):
Yeah. Gimme a

Leo Laporte (01:30:41):
Last time you were on, you had eyes looking through your hand. We made that the the the, okay,

Daniel Rubino (01:30:47):
Now

Leo Laporte (01:30:47):
It's on. Okay. Daniel can

Owen Thomas (01:30:49):
Is worth,

Leo Laporte (01:30:50):
He can't look away.

Daniel Rubino (01:30:51):
Now I'm closing one eye. I'm trying to close one eye and it <laugh>, it's, yeah, it

Owen Thomas (01:30:56):
Doesn't like it, huh? He

Leo Laporte (01:30:58):
Cannot look away.

Daniel Rubino (01:30:58):
He tries to, tries to make open.

Leo Laporte (01:31:00):
Yeah. Yeah. He's trying <laugh>, he

Daniel Rubino (01:31:01):
Can't If I'm looking down.

Leo Laporte (01:31:03):
No, no. Eyes through your eyelids. It's gotten better. It actually it's gotten better since the last time. Put your hand in front of your face. Let's see if it's, yeah, let's see if it still does that. <Laugh>. It's trying, there's, it's trying to look through your hand. <Laugh>. This is like a lease video film Suddenly <laugh>.

Daniel Rubino (01:31:20):
Well that's creepy. Yeah. I haven't tried that one before. <Laugh>. But yeah, I mean, so NVIDIA's definitely positioned here. I will say that with Intel, don't forget. So, you know, we're talking about server costs and all that, and that's true, but there's the other half of this that's coming out, right? And it's already kind of out NPUs, right? Neural processing units. So you'll have a localized processing. Well, this

Leo Laporte (01:31:39):
Was a big, big, right now we, using big thing by the way, at Build Bo I think Paul brought this up. A lot of the technologies they were talking about on device tech technologies require an NPU only. I mean, right now only Qualcomm makes NPUs. It's gonna be the fall before the Meteor Lake. Intel processors with neural processing units come out. Of course, apples stuff all has NPUs now cuz they're using their own chips. Microsoft's gonna be a little bit behind because of that. They've gotta get NPUs out. I think just the surface go right or surface x uses

Daniel Rubino (01:32:14):
On Surface Pro nine. Yeah, pro nine because well the the 5G version cuz that's armed. So that can do native background blurring, AI tracking and all that kind of stuff. And that's npu Right now we're relying on GPUs mostly to do that. And occasionally CPUs fine, but it's Bruteforce NPUs are hard coded design to handle ai. Intel does have theirs coming out, but yeah, you're right. Technically right now it's just Qualcomm and they've been doing it for years, right. Because of smartphones, so.

Leo Laporte (01:32:40):
Right. Yeah.

Owen Thomas (01:32:41):
I I just wanna point out that Qualcomm is worth slightly more than Intel. They're both around 120 billion, but Qualcomm's a little more a md, which has long been the underdog runner up to Intel is worth 200 billion. And Nvidia is worth 960 billion. They've been flirting with that $1 trillion.

Leo Laporte (01:33:02):
They're almost trillion. It's trillion. Can you imagine that Eight

Owen Thomas (01:33:04):
Times as valuable as Intel?

Leo Laporte (01:33:07):
Yeah, yeah. Does that, that might just be an overheated market, right? That's not necessarily what they really are worth. That's too their, their price too many. Whoa. Their price to earnings is 202 <laugh>.

Owen Thomas (01:33:20):
It's a, you know what, what, I think that's a

Leo Laporte (01:33:22):
Big ratio.

Owen Thomas (01:33:23):
I think Warren Buffet said that the stock market in the, in the short term is a popularity contest, right? And in the long run it's a voting machine, right? So yeah, I mean is any stock worth what it is? Well, yeah, because someone bought it and sold it at that price. But, you know, that's a good

Leo Laporte (01:33:41):
Point. It's, it's not really tied to the return of the company. It's only tied to the, what people think. It's the return of the company. Yeah.

Glenn Fleishman (01:33:49):
You can tell when the time travelers show up though, is they show up just before it does that, and they place they buy calls on the stock. So if you watch for a lot of calls, then you can identify whether time travel exists.

Daniel Rubino (01:33:59):
That's the, in Intel's saving Grace right now is their foundries, you know, that they build chips and they're gonna be building chips for other companies, right. Including Qualcomm. Right. And they, that's their you know, I think their next big business move as, but their processors are catching up. They're, I think they're positioned long term they're going to, all right.

Leo Laporte (01:34:19):
One of the stories that got buried, cuz it wasn't an AI story, but you guys call it as one of the top five, and I agree Windows 11 is getting cloud powered os backup and restore the kind of thing that Oh yeah, we've got with an iPhone. And you've got with your Android phone windows and we had

Daniel Rubino (01:34:37):
With Windows eight

Leo Laporte (01:34:38):
<Laugh>, well really this was a Windows eight feature that that disappeared, huh? Yeah. Wow.

Daniel Rubino (01:34:43):
Yeah. Yeah. A lot of people upset about that. Yeah. So it's a basic function, right? That when you say you get a new Windows 11 pc, you install it, you log into your account and then the question is, all right, what are the apps I need to install <laugh>? You know, and if you have a pretty previous Windows 11 device, this is just going to populate the stuff from the store on your start menu again, just like a smartphone would when you, you know, transfer your

Leo Laporte (01:35:06):
Old. That's awesome. Awesome. Actually that's a great feature. Yeah. Yeah.

Daniel Rubino (01:35:09):
Pretty basic, you know. But yeah, windows eight used to have it and then it got rid of it and people have been wanting it back ever since. So all

Leo Laporte (01:35:14):
Time. So that was a little pallet cleanser in between AI stories. I just thought we'd throw that in. Yeah. <laugh> just to refresh your pallet actually, you said, I think really this whole arm thing is interesting because windows on Arm historically has kind of not been great. I found it's getting better and better. I've started using it on my Mac in emulation or no, not emulation in vm. Well it is emulation, I guess in, in hypervisor, hypervisors, VMware or no, I'm sorry. It's parallels and it actually runs great. It really runs. Yeah, it's very snappy. Are is, is Microsoft going to, how important is ARM to Microsoft?

Daniel Rubino (01:35:54):
Very important. But Doka Apple has a different strategy. Apple is like we're switching from X 86 to arm and we're just gonna throw that switch and everybody has to come over to it. And because they can do that, they, they do. And they were in just a better position long term with their chips. Microsoft is will support arm, but we're going to keep X 86 round and that's, you know, two different strategies. And so they're getting there. Windows 11 is much more optimized for ARM than Windows 10 was and Windows 12 will be even more of that app build. We saw them bring over Unity now on ARM and a bunch of other tools that are needed for developers to build apps, right? If you're doing a basic app right now, you can do it in arm, not a problem. And you can even do hybrid apps now X 86 and arm.

(01:36:42):
But the problem is we have very complicated apps that rely on ex other libraries. So those libraries don't support arm. Right. You know, it's a problem. And so that's been the journey there combined with Qualcomms chips. Haven't been that impressive. They've been getting better, but they've always been like compared to what last Gen Intel has done, the big thing now we're expecting is of course the Nuvia chips, what are called the Onyx processors. I expect those to be debut later this year in the sense that they'll premier them, they'll talk about them, the architecture, the processing speeds. I don't think we'll see them ship until at least summer 2024, maybe in fall 2024. So we're still far out. But according to Qualcomm, they've had a lot of design wins. Meaning a lot of OEMs are committed to building laptops and devices around these chips more so than any other at that Qualcomm says 2024 will be the pivot point for Windows unarmed. That that's gonna be where we're gonna really start to see it take off. Will that happen? Will Onyx deliver? I think it's a tough challenge, especially again, one chip, but it is the ex Apple engineer, so fingers crossed.

Leo Laporte (01:37:53):
Yeah, it's an interesting story. Guy left who was instrumental in the design of the Apple chips. Left Apple founded Nuvia, apple sued him. He countered suit. I think those suits had been dropped cuz he took a lot of employees with him to Nuvia and then Qualcomm bought him. And so that's how Qualcomm 4 billion Yeah. Got into the race. It's interesting cuz when Apple came out with these apple silicone chips, it it, they, they were performant, they were low power, they had NPUs built in, they had integrated memory. It looked like they had kind of leapfrogged the PC industry. And I imagine that all of this is in response to that. Yeah.

Daniel Rubino (01:38:34):
Oh yeah, absolutely. In fact, so I I put up that story later about <laugh>. So Microsoft's been hiring people to

Leo Laporte (01:38:41):
Build, you said stop being ridiculous. Microsoft is not building its own processor, so knock it

Daniel Rubino (01:38:47):
Off Windows fan and Windows fans have this weird fantasy that Microsoft is just gonna build an arm processor, it's gonna come out the gate Gen one, it's gonna blow everybody away and that's gonna be the end of it. And it's less like that's not how any of this works. <Laugh>. Yeah. Apple came out with the M one, but it took em 10 years. 10

Leo Laporte (01:39:03):
Years. Yeah.

Daniel Rubino (01:39:05):
They, they were doing it for a very long time and they had a lot of devices to test it on and and scale. Microsoft does not have an a four processor to start with, to build up to an M one. Right.

Leo Laporte (01:39:17):
Apple putting these out on the other phone for years.

Daniel Rubino (01:39:20):
Yeah. So I think there's some truth to this in this sense that I do believe Microsoft experiments with chips and chip design that I think that's actually publicly known that they've done that. They've done it with you know, for server chips and have explored that. They have people working on it. They build their own co-processor. That's public knowledge. They have 'em in the surface device. So they do, and they use 'em in HoloLens, they use 'em in Xbox. So they do build microprocessors, but they don't do it at the scale of like what Apple is doing. That said, I wouldn't be surprised that back in 2019 and 20th they were you know, looking into this, which is where the Bloomberg story came out saying that Microsoft was looking to build its own processors to use in Azure and they may trickle down to surface devices.

(01:40:02):
I think what happened though was they were hedging their bets. If Qualcomm couldn't, wasn't up to the task of Windows and committing to it, or they didn't, they were going back out of that market. So I think, but then what happened was Qualcomm in 2021 bought Nuvia for 1.4 billion. I think they did that because it was a strategic decision, but it was also sending a message to Microsoft, Hey, we're serious. We're committing to this. And I think Microsoft was okay, why would we need to build our own chip at this point when we're going and get it from these people? It's, it doesn't make sense for Apple to make their own processor

Leo Laporte (01:40:34):
Or Microsoft. Yeah, yeah, yeah, yeah. Although if they did, it should go in this beautiful cooler master sneaker X computer <laugh> that is,

Daniel Rubino (01:40:43):
I don't get this stuff

Leo Laporte (01:40:44):
<Laugh>. That is now, that's a key. I love

Daniel Rubino (01:40:46):
Casey's vote

Leo Laporte (01:40:47):
<Laugh>. It's only $6,000 <laugh>. I like

Daniel Rubino (01:40:53):
$3,000 over the cost of the parts

Leo Laporte (01:40:55):
For it. Yeah, yeah. No, it's definitely the the the case. You're, you're paying for the case. I'd like to, they should put this next to though a screen or a person or something. Cuz how big is scale? Yeah. I don't know. Is it actually sneaker size? I don't know. No, it is not <laugh>. It's, yeah, it's a giant sneaker. All right, let's take a little, let's take a little break and then Owen, I'm gonna get you back in the conversation because the Surgeon General has told us social media is bad for kids. And I'm going to, I'm shocked. <Laugh> Owen, it's great to have Owen Thomas here from the Examiner. Of course our great friend Glen Fleischman from the incomparable and his new book Shift Happens. We'll tell you about that. Glen.Fun is his website, and of course, Daniel Rubino, editor-in-chief of Windows Central Great panel.

(01:41:49):
Our show today brought to you by Look Out business has changed, right? I mean the, the pandemic changed how we work. People don't, you know, don't necessarily have to be at the office these days. Boundaries to where we work or even how we work have disappeared. But if you are in charge of protecting your company's data, that's scary. Business has changed, and your data is always on the move. Whether it's on a device in the cloud, across networks, at the local coffee shop. Your, your employees love it. Users love it, but it's a challenge for IT security. That's why you need Lookout. Lookout helps you control your data and free your workforce. It's possible to have both. With Lookout, you'll gain complete visibility into all your data so you can minimize risk from external and internal threats. And very importantly, ensure compliance by seamlessly securing hybrid work.

(01:42:43):
Your organization doesn't have to sacrifice productivity or even employee satisfaction for Security. Lookout makes it's job easier too. You know, with IT security these days you might have multiple solutions, none of which work together. You're trying to get it all working and, and you're spending way too many cycles. Getting the tools to work together and too little time on Keeping Secure Lookout solves that. It's a single unified platform that reduces it complexity. So you can focus those cycles, those brain cells on the things you need to whatever comes your way. Good data protection. It does not have to be a cage. It could be a springboard letting you and your organization leap toward a future of your making. Visit lookout.com today and learn how to safeguard data, secure hybrid work, reduce it complexity lookout.com. Lookout, thank you lookout for supporting this week in tech. Thanks to all of you for supporting it. We had a great week. We put together a little video for you to watch.

Leo Laporte (01:45:07):
Ai Leo says, cuz we are living in a generative world, and I'm a generative girl. Oh, please, this is my world. My world, <laugh>. He can, this really fr so <laugh>, they've got this ai Leo, and they can make it say anything. That wasn't me, that was ai. Leo. I'm living in a generative world, please. Holy cow.

Owen Thomas (01:45:31):
I I would never expect a Madonna reference from you <laugh>, if I'm honest.

Leo Laporte (01:45:37):
<Laugh>. Oh, come on. What about Taylor Swift? You know, I'm a swifty.

Owen Thomas (01:45:41):
Oh, yeah. That I can see. Okay.

Leo Laporte (01:45:43):
That I can see. Okay. Surgeon General has warned that Taylor Swift may harm children and adolescents. Wait a minute. No, sorry. Social media <laugh>, although Surgeon General, there are things in this world that could harm them much worse than spending a little time on Instagram. Dr. Vivek Murti cited a profound risk of harm to adolescent mental health and urged families to freak out. No, to set limits. Okay. And governments to, to set tougher standards for use. Is it, I mean, he is the surge in general, Owen, and I guess he's supposed to warn us about things like this seems like there are worse things in this world.

Owen Thomas (01:46:25):
Well, it, this also seems like 10 years too late. I mean, it was years ago that a leak revealed that Facebook's own researchers had found that social media harmed adolescents. I mean, Facebook knew this and did nothing. So I think there's a lot to be said for it's time for government to step in because the companies are, you know, it's not like they're ignorant of the problem. They're fully aware of the problem. They don't seem to care. They don't seem to be willing to take the steps that might, you know, cost them money.

Leo Laporte (01:46:56):
Well, they shouldn't. I mean, this is their business. Address it. They're never gonna admit. Oh, yeah, we're, we're the, it's like Fentanyl saying, yeah, I'm the problem. It's not gonna admit it. <Laugh>

Owen Thomas (01:47:06):
A Taylor Swift friends. Yes. <laugh> Facebook clicking the mirror and saying, it's me. I'm the problem. It's me. I'm the problem

Leo Laporte (01:47:13):
<Laugh>. But, but I don't know if government is the right tool to fix this. Montana has signed a bill banning TikTok starting on January 1st. Is that the way to answer?

Owen Thomas (01:47:25):
Well, that's, you know, that has nothing to do with protecting children. Right? It's, it's about performative geopolitics at a, at a state level scale, which makes no sense given that our union you know, seeds, foreign relations to Right. To the federal government. Right. Montana should not be in this business. And it almost certainly violates the free speech rights of TikTok users in Montana.

Leo Laporte (01:47:50):
Tiktok creators and TikTok itself have both sued. It will, I mean, on, on First Amendment grounds alone, it'll probably get thrown out. Nevertheless, there are a lot of people that think TikTok is a, you know, a

Owen Thomas (01:48:01):
Yeah, no, I mean, I, I I think that, you know, the question of whether a state should have the authority to ban an app is right. Separate from is that app actively harming people? I also, you know, I also think that like humans are our own worst enemies. We will find ways to hurt each other whether it's in, in person, in writing, you know, the founding fathers used to like rip each other you know, rip, rip each other apart in like pamphlets that they circulated. If you've, you know, if you've watched Hamilton, you know, a little bit about this <laugh>. But to the point, I mean, aside from being a decade too late, I think, I think there is a lot of value here to the surgeon General, even in, you know, even if it doesn't lead to regulation, just saying this is a real problem.

(01:48:49):
The, the flip side is that this horse is out of the barn, teens are going to use social media. What we need to do is adapt to it and fig, you know, like figure out how to, you know, figure out how to get our, you know, our youth, like in a healthy place where they can be on social media, know how to take it with a certain lightness and grace. And, you know, and, and also behave appropriately because, you know, a lot of the problem is, you know, teens bullying other teens. You know, it's, it's mean girls on, on you know, internet you know, on internet steroids.

Leo Laporte (01:49:32):
I have to point out the number one cause of death among children and adolescents in the United States is firearms. They're not doing anything about that. I, I think the Surgeon General put out a warning about

Glenn Fleishman (01:49:45):
Firearms a long time

Leo Laporte (01:49:46):
Ago ago. A firearms will kill you. So be careful. Listen. Yeah,

Glenn Fleishman (01:49:50):
Yeah. No, I mean, sir, about chil, this is actually, it's, you know, it's a public health disaster. Firearms fall into that category. There's been a lot of efforts at state and federal level to declare it that. And it's funny, the social media thing might fall in the same category where it's like, yeah, we all think, or no, we all, but a significant number of people think it's a, it's a bad idea to allow unfettered access. But firearms a different issue in that regard. But how do you provide it with the, how do you not do it? Something that violates the constitution, whether or not you support the full range of it in the Second Amendment, or really want in the First Amendment.

Leo Laporte (01:50:25):
Well, I guess you're right. Tiktok is not protected account by the Constitution, but the First Amendment does kind of cover this whole thing. Yeah. Well,

Glenn Fleishman (01:50:32):
Corporate, I mean, this is the double-edged sword of giving corporate free speech rights, is that corporations have expanded rights and they they would've 20 to 50 years ago, that allow them to defend themselves. In that way it's when, it's ironic when Free Speech Maximalist are saying that corporations like Disney can't say what they wanna say. It's sort of a funny thing. I mean, I guess it's always a don't need, the freedom from repercussion is not what the First Amendment says. The freedom from government intervention and your right to have speech. But I don't know. I mean, we kept our kids off devices. We, they did not have full-time devices until they were in their early teens, and they didn't have access to devices overnight. And we set limits, and

Leo Laporte (01:51:10):
That's probably a good idea. Yeah.

Glenn Fleishman (01:51:12):
Contract with them that schools had suggested about usage. And a couple times they're like, oh, I set up an Instagram account. What, without telling you, we're like, all right, well, we're gonna disable it, but we're not gonna, you know, you talk what's fine. And so now of course, when I got off Twitter, they both got on both my, my 16 year old, 18 year old just for the lulls. But they're mature enough to deal with it because they laugh at Twitter as opposed to our sucked into its dramas.

Owen Thomas (01:51:35):
I, I, I would point out that there's an interesting new legal avenue for kind of you know, like holding social media companies accountable for their actions. And it's, you know, it's not free speech because that's pretty cut and dried, and it's not section two 30, because that's proved very tricky to kind of take apart. And Mike Masnick among others has, has laid out a very strong argument for why you don't want to, you know, take apart Section two 30 to just to go after social media companies. But product liability has been an interesting kind of legal tactic. Snap, for example, was found liable. I believe I, I'm not sure if this was a final ruling, but in some early rulings in a Case, snap was found liable for creating a filter that lets you post how fast you were going when you took a Oh yeah. Took

Leo Laporte (01:52:28):
A snap. Yeah. Yeah. Oh, yeah. I used that on the bullet train in Japan. But also a number of people used it while they were driving.

Owen Thomas (01:52:35):
Yes. And it, you know, and and the plaintiff's lawyers argued that this encouraged people to speed to kind of show off. And I think that, you know, I think that this could be, you know, it could actually be a valuable corrective for people to pursue product liability claims against these social, social media companies to say, snap,

Leo Laporte (01:52:55):
By the way, design of that filter, and probably would have, even with abstinent court case. I mean, I think these companies try, I re snap does I know they try to be good citizens.

Owen Thomas (01:53:08):
Well, I think should it have existed in the first place? Should, you know, I think that well,

Leo Laporte (01:53:13):
These mistakes happen.

Owen Thomas (01:53:15):
Well, Kara Swisher talks about how Facebook showed them, you know, their live video broadcasting features and, you know, she said, you know, someone is going to shoot someone live on this. And they said, what are you talking about? That's, you know, like, they didn't, it wasn't even conceivable

Leo Laporte (01:53:32):
To them like she was, right. Yeah,

Owen Thomas (01:53:34):
Yeah. She was absolutely right.

Leo Laporte (01:53:35):
Unfortunately, yeah. Jeff, Jeff Jarvis, I'm gonna channel our our host from this week in Google posted this text from the Surgeon General 51 years ago. All the available statistics confirm the pervasive role television plays in the United States, if not throughout the world, more people on television sets and more people watch television than make use of any other single form of mass communication. The question of violence on television has been one issue that was raised almost immediately after television became a major contender you know, warning about television bef after that, they warned about video games, violent video games. There's always been a new cultural phenomenon that we old people look at and say, you kids in your rock and roll, it's running your mind.

Daniel Rubino (01:54:26):
It reminds me the the pessimist archive, that's all they do. They, they show a lot of that where they they'll show bio books, <laugh>, we're gonna take down society. Yeah. Then it was radio, then it was the bicycle. Yeah. So there, there is there is a certain irony here that, and we need to be careful that, yeah, we're just now overreacting to what's new. But I think social media is clearly different in some ways, right? It, the, the fact that it instantly accesses millions of people and its ability to spread disinformation or violence. You know, there was, I even hate talking about this story a couple weeks ago with the torture vid animal torture videos on Twitter. Oh, ugh. They took down their, you know, like, yeah. And no

Leo Laporte (01:55:08):
Kid should ever see that. That's horrific.

Daniel Rubino (01:55:11):
Right. I mean, that's worse.

Leo Laporte (01:55:12):
I

Daniel Rubino (01:55:13):
Just reading about it, I, I had like PTSD for days. Yeah. I didn't see it.

Leo Laporte (01:55:17):
Yeah, yeah,

Daniel Rubino (01:55:18):
Yeah. It just, reading about it wasn't so disturbing. Yes. I can't imagine children who saw that stuff. And it's like, what do you do about that? Especially since currently, you know, with Elon, you know, I hate that we keep picking on 'em, but Yeah, there's no reaction to this right there. It doesn't seem to be a concern. And then you have, like with the Republicans, it's like you have free speech and you can do whatever you want, but then TikTok is Chinese, so maybe we should ban <laugh>. Yeah. It's just like, you know, it, it's all over the place. And I don't, and meanwhile though, and to be fair, in China, what do they do? Right? They ban social media <laugh>, they don't want the kids on stuff. They're limited to gaming just for a couple hours a week. And they force the kids to learn stuff. And so their, their children are all like, you know, focus, education, engineering, <laugh>. And our kids are just like on social media all the time. I mean, TikTok is not even in China. I always thought that was interesting. Well, they have their

Leo Laporte (01:56:09):
Own, they their own version of it. Yeah.

Daniel Rubino (01:56:12):
Yeah. But TikTok is for the rest of the world. Right, right. So it's sort of like it's a fascinating thing. And I don't, you know, yeah. We had the first amendment, so you can't just ban stuff, but, well, we

Leo Laporte (01:56:23):
Should ban candy cuz it rots your teeth, right? Hell yeah. Hell yeah. No candy allow, I mean, you know,

Daniel Rubino (01:56:30):
Like

Leo Laporte (01:56:31):
As as Mayor of New York, Mike Bloomberg pushed the oh yeah. He didn't want bottles

Glenn Fleishman (01:56:35):
Sugar sugary beverage chapter

Leo Laporte (01:56:37):
Pop. Yeah. Here, by the way, thank you for referencing for name checking the pessimist archive. Daniel, this is fantastic. Oh, yeah. Oh. Yeah. Yeah. Look at this one. Teddy Bears Doom race. The teddy bear is a, not kidding. 1907. The Teddy bears a social menace, an aer and a better of race suicide. I mean, we've always had this moral panic. I'm gonna send this site to Jeff Jarvis. This is fantastic. Everything here's, you know, comic books, oh my god, comic good books called The Answer to Comic Book Addiction Kids Today, evil communications, comic books give Tots Bad Habits from 1949 and on and on and on. This is a great question. I don't think there's any bad,

Glenn Fleishman (01:57:27):
Bad media. I think it's just bad parenting. There's my

Daniel Rubino (01:57:30):
Statement. That's true too, right? I mean, I'm being, but you know,

Glenn Fleishman (01:57:32):
I'm being evil by saying that. But,

Leo Laporte (01:57:34):
But it's reasonable. I'm a good parent. But if there's a hazard to all children that parents, you know, may be having trouble dealing with that the government could step in. You know, you don't, you, you wanna make sure you la label Tylenol not appropriate for children under a certain age. I don't know, Tylenol, I, I've, I literally have a friend who lost her daughter to, to Tylenol over overdose. Oh wow. I mean, it's a deadly, it's a terrible deadly medicine. So

Daniel Rubino (01:58:03):
The parenting thing is true, but the, but it's such a complicated issue, right? Because now, like when I grew up in the eighties, like I had a mom at home for a while. Like, parents were allowed to parent their kids. Now everybody's so busy and, you know, there's so much going on. Like, even like everybody, both parents have to often work. It's like, it's very hard for parents to be true parents in that sense. So like, I, I even that I don't know what to, they can do that. They can handle all this.

Leo Laporte (01:58:30):
I don't know how, how women have survived the telephone. This is from nineteen seventeen two talking machines, a woman at a telephone are to blame for Sal's husbands and $4 potatoes. That, that's the explanation for wrecked Holmes in the high cost of living being given out by Mrs. John c Bly, president of the Housewives League. So there you go. There's the problem. The telephone, I don't know, I make light of this. This is a great sight. I'm, I just li I could go on and <laugh>, I make light of this. I just, it bothers me. The, the attorney General, I'm sorry, the surgeon, general <laugh>, thank God he's not, the Attorney General is focusing on this when there are so many other hazards to young people. Mm-Hmm. <affirmative>, I'm not sure I Well,

Owen Thomas (01:59:18):
We can, you know, like we can do more than, you know,

Leo Laporte (01:59:23):
More than we can do more than one. I agree. Yeah. I agree. But I, I think it's overstating the case to say, for instance, the mental health issues in young women in America today are due to Instagram and Facebook. I mean, that seems a stretch to me. Well,

Glenn Fleishman (01:59:39):
It's like, it's like accelerants are always a problem and they're hard to measure. Yeah. Because for some people it has no effect. And other people you know, it can be a dev devastating thing that they wouldn't otherwise have encountered. And so the question is, is the greater exposure to things that otherwise someone might not have seen, is that worse than real life? If the internet didn't exist, would they be facing the same hazards at this time? You know, you can't really do an ab test on that. I mean, we talked to our children, frankly about drugs and it's like, we're like, look, fentanyl is a different category. So, you know, Instagram and Facebook didn't exist when we were children. Neith did fentanyl in this variety. True. And it's like, true fentanyl is destroying people and killing people. Sure. And it's, it's not like pot or even, it's not even like heroin. It's not even like other of these super hard drugs when we were kids. You know, some of which were circulating around and new people who were engaged with. It's like fentanyl is a whole new thing. So how do you respond to fentanyl? I think there is a little bit, not that Facebook is fentanyl, but it's like, there is a little bit of that. Like, well, it's a million times more powerful than the telephone than

Leo Laporte (02:00:42):
A comic book.

Glenn Fleishman (02:00:44):
And it's free and it's free and your parents don't want you to do it. So

Leo Laporte (02:00:48):
I think you did the very sensible thing. And we used to do that. We turn off the router at 10:00 PM cuz kids will Oh that good. If they you know, it's addictive. I'll stay up all night if I if I start looking at Instagram, I can see doom scrolling and all that. And you know, I mean, there's certain things you can do. I've fortunately, my kids are a little bit older. I didn't have the same problems. Dr. Murti has also decried the public health crisis of loneliness. Isolation and the lack of connection in our country. I, yeah, I think that's true. I don't know if it's government's job to fix it, I guess is my question. Yeah.

Daniel Rubino (02:01:23):
Well, I think just highlighting it is important. Profess Galloway, you know, Scott, he talks about this all the time. And I think it's, it's true. You know, like be between the pandemic and the, the changing social structures and you know, how men is being, is being perceived. Like there's a lot of moving parts here. And I take a, I love working from home. I, I love being an introvert and being a shut in, but I'm also in my forties, I've done my time <laugh>. Right, right. I wasn't like that in my teens and twenties. But if you're in your teens and twenties, those are formative years to build these skills out. And if you're not getting them, it's like, it can cause a lot of problems. And there's that issue with and I forgot the name already, but they have it in Japan with men pillow,

Leo Laporte (02:02:04):
Men pillow prides, <laugh>.

Daniel Rubino (02:02:06):
Yeah. Well, it's, it's called like, ah, I've got the name of it, but there's a name for it for Oh,

Leo Laporte (02:02:10):
They, yeah. Yeah. They don't leave the house. Oh, yeah, yeah,

Daniel Rubino (02:02:13):
Yeah. They pull out of society basically. And that's starting to happen here for different reasons. You know, there's changing, you know, and now you have steroid use going up and, and young men because of now it's like body image stuff going on with men too, because everybody wants to, you know, they go on Instagram and they see all these guys who are jacked and it's just like, there's a lot of stuff that's going on. Which is why I'm not a parent, by the way, because even I'm like, well, Nick, I, I, I don't know how I would handle this stuff either. I think it's very complicated.

Leo Laporte (02:02:40):
It's called, and by the way, by the way, AI is making, go ahead. Oh, sorry. AI is making all of that worse. Right. All of the, all the filters. A AI makes everything worse. Right. Everything <laugh>, it's, it's called and everything better too, at the same time. It's so amazing. It's like blockchain.

Glenn Fleishman (02:02:57):
If I wanna go on Instagram and see somebody who's jacked, I just go on and look at Owen Wilson. I mean, come on. I mean Owen Thomas. Sorry,

Leo Laporte (02:03:03):
Owen Wilson.

Glenn Fleishman (02:03:04):
I, I had that joke all prepped that I still miss

Leo Laporte (02:03:07):
<Laugh>. It's called Hiko Morrie. He, Japanese Culture Bound Syndrome of Social Withdrawal. This is an article from the National Institutes of Health Health recommending it be added to the D sm mm-hmm. <Laugh>, the diagnostic manual for a mental illness. I don't know if that's, I don't know. Is it a new psychiatric disorder or I don't know. Next week we'll talk. We didn't, I meant to. I thought I would with you Glen, talk about what Apple's gonna do a week from Monday. But I guess we'll do that next Sunday. That's gonna be

Glenn Fleishman (02:03:39):
I might, I might lose my bet. I may have to eat, eat a hat of some,

Leo Laporte (02:03:42):
You think that Apple is not gonna release a vr helmet or what?

Glenn Fleishman (02:03:46):
Well, my, my argument was I didn't think they were gonna lease one as a product that would ship this year for a reasonable sum of money. But I thought it's possible they'd do a developer release that would be anchored to a computer and only available to developers and in limited quantities. And so I think I'm right about that second part. But all the, all the word that's coming out from people who seem to know birdie's inside the company which only happens when things are getting actually close to being released, that these, this information comes out. Seems like Apple will be ready to ship it this year as a consumer product. I still think it's ridiculous. I I don't see how they're gonna make it work, but I could very well be

Leo Laporte (02:04:24):
Hung. So let's just to get this on record, <laugh> exactly right. You're not saying they won't announce something, you're not saying it won't be a developer product. You just don't think it will ship as a consumer product in 2023.

Glenn Fleishman (02:04:37):
Yeah, I don't think it'll be out this year. It just doesn't seem right from like a, like,

Leo Laporte (02:04:42):
Like who would buy it is what you're saying. Yeah.

Glenn Fleishman (02:04:44):
Well that and physics and hardware and what's developed for it. I don't think that developers get the thing if Apple announces in June, even if they're ready to ship a developer or test version, which they, which they've done for multiple previous products in, in major transitions over time. Then it's one thing to port software, you know, across generations of Intel hardware, intel to M one or whatever. Those things have happened before. It's another to develop stuff that's entirely new and compelling in an ar vr environment. And I just don't think there's time to make it happen. Yeah. But I also just don't think the hardware could exist the way it's described. It's either ridiculous if it, if they're gonna ship what's being described widely sounds like an absurd Homer Simpsons car to me. <Laugh>. Or if it's not that I don't, I don't think the technology exists to make a thing that won't be like that. So that's,

Leo Laporte (02:05:33):
That's my, I'm I'm not disagreeing, I'm not betting against you ready to be wrong. I am not betting against you. I think you might be wrong, but not because you're not Right. <Laugh>, <laugh> you're wrong because apple's executives have put so much wood in those arrows that they can't pull back at this point. So even though you're absolutely right, apple should not do this. They're gonna do it. And I think it's gonna be the worst product mistake Apple has ever made in its history. I mean, it's gonna be bad.

Glenn Fleishman (02:06:03):
Wow. Bigger than the Newton.

Leo Laporte (02:06:05):
The Newton wasn't a mistake. I got four of 'em right there. And the Newton was the head of its tongues. Tongue.

Daniel Rubino (02:06:11):
Yeah. <Laugh>.

Leo Laporte (02:06:13):
No, the Newton, you could, you could reasonably say we wouldn't have the iPhone except for the Newton. Like the Newton was a precursor. Maybe this will be the Newton for, for what is to come in. Ar it's,

Glenn Fleishman (02:06:23):
It's, that's my, my reaction keeps being, if what I keep reading about being described is what were to ship, it seems like a complete failure. So I can't believe what they're gonna ship will be what I keep reading about, in which case they have to violate the laws of physics, or they just have some really clever hardware innovation we haven't seen before that it's going to wow people that I'm willing to be wrong about that. Cuz I don't know, I'm not, you know, I'm not in chip factories in, in China and across the US and in Germany or something, looking at what's being made. So it's, it's plausible that it's gonna be something that's gonna be extraordinary and I'm just not ready for it. The future shock will hit me there.

Leo Laporte (02:06:59):
No, I think you're absolutely right. I'm just trying to find the sales figures for the Newton will. So that's the only question. Will it be a bigger flop than the Newton? I like that.

Glenn Fleishman (02:07:10):
Now I own an Apple cube. I own a g G4 cube

Leo Laporte (02:07:12):
On, oh, I bought the cube on the last day. I got it home. And the news came out. They were discontinuing it <laugh>

Daniel Rubino (02:07:21):
The question is, do you think they're gonna talk about AI because they've been so quiet about any of that. Yeah. Like hope so. And to me that's the weirdest part of this. It's like if they're doing ar vr, it's like, alright, but everybody else is just like we're switching positions now on that. That's last year's fad. We're moving on where AI is the new thing and here's Apple. Like we're gonna do a r vr r even their geniuses and really figure this out. Like, you know what, Glen is kind of alluding to <laugh> maybe, or yeah, this could be just a really big error in, in terms of guidance.

Leo Laporte (02:07:54):
I think it's gonna be huge

Glenn Fleishman (02:07:55):
Error.

Leo Laporte (02:07:56):
I really do.

Glenn Fleishman (02:07:56):
They have the neural cores. They've got, they've been making chips with multiple cores. I mean I, you know, I like Apple's approach. They've done a lot of they've done a lot of machine learning stuff that is not very flashy, but works actually quite well. That's the device to device and, and end to end encrypted and relies on in some cases dedicated cores. Like Apple added, you know, OCR to photos library. So everything, every piece of text now I can search on it. That's extraordinary. It's an incredible improvement. And it's so quiet. Most people don't know they did it. Because they want it just to be something that's emergent when you search and you just, oh, there's more things

Leo Laporte (02:08:32):
Than that. That will be maybe Tim Cook. That would be an interesting attack to take. Tim Cook comes out and says, we don't think AI is ready. We don't think it's safe. We think it should be incorporated into existing products in a way that's safe, reliable, and private. And that's what we do as the privacy company. Meanwhile, casting shade on Google and Microsoft as not caring one bit about your privacy or accuracy. Yeah. the Newton, according to ours, Technica sold an estimated 200,000 units. Now it's interesting, it's not terrible because Apple's saying they expect to sell a hundred thousand of these, of these nerd helmets. So maybe Apple really is in their mind thinking this is our Newton and we will How long was it after the Newton came out that came out with the iPhone? The Newton, how long was it? Like

Glenn Fleishman (02:09:22):
12, 15 years? Yeah, quite

Leo Laporte (02:09:24):
A while. 1997 to 2007. Yeah. So it was exactly 10 years. That's short period. Yeah. Interesting. So it might be, that might be accurate. They might be saying in 10 years we'll have the AR device, the equivalent of the iPhone to the Newton. Yeah, that'd be interesting, wouldn't it? I actually have a little soft spot in my heart for the Newton. If this is a Newton, that wouldn't be so bad. Yeah, that wouldn't be so bad. But, and and by the way, let's not forget that one of the outcomes of Newton was the arm processor <laugh>,

Glenn Fleishman (02:10:01):
Ah,

Leo Laporte (02:10:02):
Which was the processor used in the Newton Apple bought, or actually founded arm based on the acorn risk machine, a r m. And eventually sold off their steak, but there would be no arm without Newton. Wow.

Glenn Fleishman (02:10:17):
Forgot about that.

Leo Laporte (02:10:18):
That's good. So these go these early, you know, and Apple's n Apple, I think Tim Cook is very aware of Apple history and the, and the lessons learned, it might very well be that's how they think of this. You've given me a new perspective on this. So it's so, it's not that you're wrong, you're just not right.

Glenn Fleishman (02:10:34):
I, I appreciate it's very Stephen Colbert <laugh>. You're so wrong. Why are you Right? <Laugh>

Leo Laporte (02:10:41):
<Laugh>. You're not, not

Glenn Fleishman (02:10:43):
Right. My, my analysis is correct and the patient is dead, but the patient should be alive according to my calculations.

Leo Laporte (02:10:47):
According to my careful calculations. I'm glad that y'all decided to come out and play today. That's what I'm gonna say. Even though you really didn't have to leave your homes. Owen Thomas, San Francisco examiner, bring back Dati. Love you. It's great to see you. And please give your favorite furry Alex a hug. No, I'm sorry. You're his daddy. I'm sorry. Let's get that right. No, I'm just teasing. I like to tease Alex, you know, he lives in my boyhood home. Have I mentioned that? How many many times have I mentioned that? That's hilarious. You didn't know that.

Glenn Fleishman (02:11:21):
Is there a plaque in Providence,

Leo Laporte (02:11:23):
In Providence, Rhode Island that, that house, that same house? Yeah. That is freaky. We found out it was complete coincidence. His, his fiance at the time, now wife Liza was out visiting. She, and she's was doing her residency at Brown. I said, oh, I used to live in Providence. She said, oh yeah, I I grew up there. Where'd you live? I said, well, I told her the street. She said, oh yeah, what number? And I told her the number. She said, yeah, that's my house.

Glenn Fleishman (02:11:49):
Wow. Oh my gosh.

Leo Laporte (02:11:51):
Yeah. Crazy. They gave me a tour last time I was out there. Anyway. That's amazing. That's amazing. So

Glenn Fleishman (02:11:55):
Where, but where, so where's the, is there a, like a, a velvet rope in front of your bedroom?

Leo Laporte (02:12:00):
There is a plaque and disturb it. Yes, there is a plaque. Leola port slept here. They, it's actually, they've done a nice job. They've fixed it up. It's beautiful. Yeah. But I, I did visit my old bedroom. That's great. Yeah. And it hasn't changed that much. It's kind of, kind of a strange experience. That is Glen Fleischman. We've gotta give a plug to shift happens. Is it too late to get a copy?

Glenn Fleishman (02:12:21):
No, you can still get it. You get a shift happens.site. There's a, a site that lets you pre-order. We we have set the order of the number of books that will ever be printed. Probably. cuz it's a big project. And then the printer is the printer said, gotta order paper cuz there's still a little bit of paper shortage. Oh dear. So you know, and this is a, it's a ultimately 1300 and se I lost track. 1,376 pages across three volumes. It's a lot of paper. It's a lot of paper. So they gotta go and, you know, talk to the mills. And the mills are still doing literally rationing. Like the printers can't order more than a certain amount each month. There's certain kinds of stock. So, wow. It's,

Leo Laporte (02:13:01):
And this is gonna be a full color. Beautiful full color. You're probably using really good paper, I would imagine. I

Glenn Fleishman (02:13:08):
See hard cover. Yeah. Two, yeah, the paper's. Very nice. Very nice paper. Spent a lot of time looking at papers. Let me tell you. <Laugh>. a lot of work. But yeah, so we're going on. We gotta go out in go on press. We'll be actually at the printers in Lewiston, Maine in July, watching sheets come off and every hour, hour and a half for at least a week, maybe two weeks, will be there 16 hours a day as they bring sheets off and go, how do you like this sheet? And we go, oh, looks okay. Or maybe you need to fix that color patch isn't perfect.

Leo Laporte (02:13:36):
Whatever. That's really awesome that you're doing that. I think that's really

Glenn Fleishman (02:13:39):
Fantastic. It's gonna be a lot of work, but it's gonna be, it's gonna be perfect. And if you give it, I mean the printers are good. If you let them do it, they'll do it. Right. It's very digital end to end process for analog things. Lots of color checking and so forth that happens all in line with the equipment. So they keep it on target. But still, this is a high touch book as they say. So we're gonna,

Leo Laporte (02:14:00):
And if touch it a lot, if you are in new England, you should go visit the Museum of Printing on Saturday, July 8th. Glen will be there along with Martian Doug Wilson, the director of a film about Linotype and some guy named Jeff Jarvis. To talk about

Glenn Fleishman (02:14:15):
Jeff, Jeff Jarvis

Leo Laporte (02:14:16):
To talk about keyboards, typewriters, and the Linotype. This will be, yes. I want to just go and see the nerds that show up, to be honest with you, <laugh>,

Glenn Fleishman (02:14:25):
It'll be great. It'll be great. Look, I sometimes go to conventions Waze Goose. They're called Waze Geese Old term. And you know, hundreds of people go to like an old printing factory or a museum. And we all talk about printing or letter press or type or whatever. It's a lot of fun's. Great. It's your goal every field.

Leo Laporte (02:14:42):
It's your goal to devote your life to this subject.

Glenn Fleishman (02:14:46):
Yeah. It's kind of accidental, but I, I love it. And you know, I, I joked at one point I was collecting obsolete professions. Cause I was trained as a type setter in the eighties and then I went into journalism. So, you know, maybe historian might be, I don't know how a little more staying power will find out. I don't Sure. So

Leo Laporte (02:15:02):
I didn't realize that. So you, because you'd been a type setter, you kind of came to this honestly.

Glenn Fleishman (02:15:06):
Yeah, I was a high school job and then in, when I was in college, I worked as a type setter and always had kind of a interest in, it's pretty history is the history of how we communicate. Right. And there's this whole, the 18 hundreds, we went from being a, a low information society in some ways to a high because of improvements in printing allowed the rapid transformation of, you know, news from, from what happened into newspapers. And it's an incredible thing. So I'm kind of studying aspects of that, that are very obscure.

Leo Laporte (02:15:33):
Awesome. Shift happens.site. I think the time to buy, you know, get this before it's outta print because it will be the one and only and you will have something that's pretty laser.

Glenn Fleishman (02:15:46):
What was that love that shot? That's a laser is a guy, one of the people who is responsible for the invention of the laser. He he co won a Nobel Prize for it. Also decided there should be a laser typo eraser. And he actually made a model, brought it on the air. It's alow, I'm liking his name. He went on the air with Walter k Cronkite and demoed it and you just zap it because the ink would vaporize and the paper wouldn't get as hot <laugh>. And he thought this could be miniaturized and just built into typewriters. And you just have, you hit it like, you know. Wow. Sadly the correcting electric was invented instead. Yes. And

Leo Laporte (02:16:21):
Mike Nemo's mom came along with Whiteouts. Yes. Yes. Very good. And and then eventually computers and you didn't use typewriter. Yeah. It's amazing. How can you imagine there's a technology that lasted a hundred years and then just disappeared. It's over

Glenn Fleishman (02:16:35):
Like, like buggy whips.

Leo Laporte (02:16:36):
Yeah. Like buggy whips. Thank you Glen. Always great to see you. Thank you.

Glenn Fleishman (02:16:40):
Pleasure

Leo Laporte (02:16:40):
To be here. Glen.Fun is two Ns in Glen. G l e n n dot funds. Thank you. Daniel Rubino, editor-in-chief of Windows Central. Great coverage on build. There was a lot more we didn't get to. It's build was I think really interesting this year. And Microsoft's focus. Yeah. First

Daniel Rubino (02:16:58):
Time in a long time.

Leo Laporte (02:16:59):
Yeah. Just fascinating. So I really appreciate you coming and filling us in on the best stuff. Thank you Daniel for being, thanks for having me. I appreciate

Daniel Rubino (02:17:08):
That. And now I'm excited about that book. That's a, that book looks amazing. <Laugh>.

Leo Laporte (02:17:11):
There's no, isn't it beautiful? What a work of art you've created. Glen, I just incredible.

Glenn Fleishman (02:17:17):
I'm just a project manager. Give all credit to Marchine, who is the designer. He took a lot of the pictures in it. He found the photos. He, you know, contacted people all over the world to take pictures of obscure typewriters and terminals. But yeah, it's about half photos. Also one of the selling points.

Leo Laporte (02:17:33):
I'm gonna, I'm gonna pick one up right now. I should have ordered it before. We've been talking about it all this time and I just like put it off. Put it off. I'm ready. Someday.

Glenn Fleishman (02:17:42):
A collector's item possibly. Yeah. Cause we're not gonna print the thing again. It's a

Leo Laporte (02:17:45):
Huge No, it's a huge taking. Can get so much work. Yeah, I can see that. Yeah.

Glenn Fleishman (02:17:50):
Might not ever

Leo Laporte (02:17:50):
Go back and Well, and I think we should have one on our, in our studio. Honestly. I think that's a, that's a great piece of history. So thank you for that. Lazy Gentlemen, we do twit every Sunday afternoon. I'm glad you were here for this episode. You can watch us live if you want. The livestream is at live dot twit tv. There's audio and video. We get started around 2:00 PM Pacific right after Ask the tech guys and we end usually around five or 6:00 PM we ending a little early tonight because this is the last episode of succession. And I have to put on my, my dark suit, my funeral suit and, and watch it. I'm very sad about that. <Laugh> don't, don't get me started. Anyway, we're glad you're here. Two to 5:00 PM Pacific. That's five to 8:00 PM Eastern Time.

(02:18:40):
2100 UTC after the fact. You can get on demand versions of the show ad supported at twit tv, our website. And when you're there, you'll see a link to YouTube. There's a YouTube channel devoted to this week in tech. You'll also see links to podcast players, so you can subscribe in them and in RSS feed because you could basically any podcast player search for twit, you'll find it right away. Subscribe to all our shows. We have a lot of great shows. Now, if you are a Club Twit member, don't forget you have your own feeds. Those are the ad free versions of the feeds. When you subscribe to Club Twit, we send you an email with links to all the shows, the ad ad-free versions. We send you a Discord invite so you can join us in our Discord chat room. We send you information about the TWIT plus feed, which has shows we don't put out anywhere else.

(02:19:25):
Like Hands on Macintosh with Michael Sergeant Paul Theros, hands on windows. We brought back Scott Wilkinson's Home Theater Geeks. That's a wonderful show this week in space and on and on and on. So club Trip members, seven bucks a month. It supports us. We appreciate it. If you can just take a little bit of the money out of your Shift Happens funds and send that to us we will be very glad to welcome you into the Club twit TV slash club twit. And we are very grateful to all of our Club TWI members for making this possible. That's it for this week in tech. Thank you everybody for joining us. We'll see you next time. Another twit is in the club. Amazing.

All Transcripts posts