This Week in Tech Episode 946 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 is time for TWI this week in tech. Great panel coming up for you. Ben Parr is here, the AI analyst, the Techn, Harry McCracken from Fast Company and our friend from Hawaii Aloha to Dock Rock. We're going to talk a lot about ai. The governor of California Vetos a bill that were required safety drivers in automated big rigs. Was that a good idea? Unity Blinks and says Nevermind. Plus [00:00:30] Apple and Microsoft and Amazon. All with big events. Big announcements. We've got the iPhone 15. Should you get a fine woven case? You want a closeup? I Fixit has it. We'll talk about that and a whole lot more next on twit
TWIT Intro (00:00:49):
Leo Laporte (00:00:50):
Doc Rock (00:00:50):
TWIT Intro (00:00:51):
From people you trust. This
Leo Laporte (00:00:54):
Is twit. [00:01:00] This is twit this week in Tech episode 946 recorded Sunday, September 24th, 2023. AI is number two. This episode of this week in Tech is brought to you by our friends at IT pro tv. Now ACI learning acis new solution insights assists in identifying and fixing skill gaps in your IT teams. Visit. Go dot [00:01:30] ACI learning.com/twit to listeners Receive at least 20% off or as much as 65% off an IT Pro enterprise solution plan. After completing their form, you'll receive a proper quote based on the size of your team. It's time for twit this week at Tech, the show. We cover the week's tech news every week. It's a different show. I kind of like that every other show. We do same panel [00:02:00] every week on this weekend Tech, I change the panel up. Of course, a lot of times it's people we know pretty well. In fact, in this case, three of my greatest friends starting with Doc Rock all the way from the Aloha state. Not Maui though. Oahu. Hi Doc Rock.
Doc Rock (00:02:16):
Good. How you been?
Leo Laporte (00:02:17):
I've been great. I heard you had a good time with my wife. Wait a minute. That came out wrong. You guys were out in Denver at podcast movement a couple of months ago.
Doc Rock (00:02:28):
Man, it was phenomenal. We [00:02:30] had a blast and your whole crew is just amazing. And right after that I left. I went to Japan and in my hotel, this is so hilarious. There is a JI DSI Rice restaurant where they basically highlight their $4,000 rice cookers and they serve like, yo Leo needs to be here. This is so epic.
Leo Laporte (00:02:53):
You and her were because you had told me to get a JI Rusi rice cooker, which I got. And that rice from [00:03:00] the mainland, it's the sushi rice, right? Yes, yes. And it's been very good, by the way. 20 pound bag. We're still working on it. It's not done yet, but we've been very happy with it. And she says you were trading recipes in the rice cooker. She has a hanker. I said, did he tell you about the Kentucky Fried Chicken leftover recipe? Says he didn't mention that. Yes.
Doc Rock (00:03:21):
Oh, that's the one. That's the one that you have to try. It is so good. I
Leo Laporte (00:03:25):
Have It seems have. It is great. It's so
Doc Rock (00:03:28):
Freaking good. It's the best thing to do [00:03:30] with chicken.
Leo Laporte (00:03:32):
Anyway, welcome. It's good to see you. My friend, doc Rock in the House also with us. Good friend Harry McCracken, the Techlog Global Technology editor at Fast Company now on Macedon Social. Hi Harry.
Harry McCracken (00:03:47):
Hi Leo. Good to see you. Even if I'm not sitting next to you this time,
Leo Laporte (00:03:51):
You are traveling to the web summit soon? Yes,
Harry McCracken (00:03:56):
Six weeks from now I'll be in Lisbon for the third year.
Leo Laporte (00:03:59):
Wow. I love [00:04:00] Lisbon.
Harry McCracken (00:04:01):
It's great. Great city integrate event
Leo Laporte (00:04:03):
And you just got back from another event, you're back on the road.
Harry McCracken (00:04:07):
Yeah, Fest Company had its innovation festival in New York last week, so I went to that and there happened to be a Microsoft event in town, so I went to that too. Went to that. Oh
Leo Laporte (00:04:16):
Harry McCracken (00:04:16):
I went to Iffa and Berlin less than a month ago I think.
Leo Laporte (00:04:20):
Yeah. Well good. I'm going to get an update on all of that in just a little bit, but first let's say hi to Ben Par, old friend Ben Par, the AI email@example.com. [00:04:30] Hi Ben.
Ben Parr (00:04:33):
Oh, hello. Quick notes doc. You made me hungry. I really want some rice. It's good, Harry, I'm jealous of the traveling.
Leo Laporte (00:04:40):
Yeah, what? You're not traveling man. Ben, you go all over the
Ben Parr (00:04:45):
Place. I don't get to go to Lisbon this year. Oh, I want to go to Lisbon in
Harry McCracken (00:04:49):
Particular. Get them to bite you. I might not be too late.
Leo Laporte (00:04:52):
I think I met Ben in Paris some years ago at Lou Webb, I think is Lu
Harry McCracken (00:04:58):
Ben Parr (00:04:58):
Leo Laporte (00:04:59):
Right? Yeah. [00:05:00] Yeah.
Ben Parr (00:05:00):
Oh my goodness. It's
Leo Laporte (00:05:02):
Been a long time. So before we get into the week's news, how is AI going, Ben? Is it still the next big thing?
Ben Parr (00:05:16):
So I would say it is, but you got to go through the hype cycles, right?
Leo Laporte (00:05:22):
We're in a hype cycle. I think that's a good point. Yes,
Ben Parr (00:05:25):
It's for sure you have a hype cycle, but the question that everyone asks is, [00:05:30] is it more like let's say tokens in crypto where you just have a big crash or is it more like the internet or mobile phones where you don't say, hi, I'm starting an internet company today, but every company is an internet company. I think most investors and most people in the valley think AI is number two, and at a certain point it'd be great when there's not all of the hype around things, but in the interim there's going to be a bunch of events happening this fall in [00:06:00] the San Francisco area. Ted's doing an AI conference next month. There's still a lot of stuff happening. Wait,
Leo Laporte (00:06:06):
It's where we get a title so early in this show, but I think we're going to call this show AI is number two. Is that okay with out of context? It has a very different meaning. Here's the famous Gartner hype cycle where you have a technology trigger, then you go to the peak of inflated expectations and then the trough of disillusionment as reality sets in. [00:06:30] But then if you're lucky, I think Bitcoin never got out of the trough of disillusion, but if you're lucky, you get the slope of enlightenment and the plateau of productivity. That's number two and you expect that. We'll where are we now though? In the AI hype cycle on the Gartner
Ben Parr (00:06:45):
Chart? It is really hard to know and it could be on the, it's definitely not as high as it was before
Leo Laporte (00:06:53):
We passed the peak of inflated in expectations, but I think we're still pretty high up. We
Ben Parr (00:06:58):
We're probably still pretty high up. [00:07:00] Yeah,
Leo Laporte (00:07:00):
The trough of disillusionment awaits.
Ben Parr (00:07:04):
I don't know what it'll look like though. That's the hard part. What does the trough actually look like?
Leo Laporte (00:07:09):
See, I think one of the problems is that at the peak of inflated expectations, you get all the AI bros saying this stuff is great, but watch out. It's an existential threat to humanity, which it's clearly not, but okay, to me that's absolute hype. That's [00:07:30] them saying, don't worry about the fact that black people face recognition is terrible and they get thrown in jail in a higher proportion because of it. Don't overlook the AI hallucinations and all that stuff. It's a threat to humanity and it's kind of hyping up this potential of it. That to me is at the peak. I think the trough of disillusionment is going to come when people say, oh, it's just a toy. It's just a party trick. It's not doing nothing. It's just like a three-year-old could do [00:08:00] and it's so much wrong stuff that it's dopey. I feel like I'm there. I'm in that trough already right now.
Doc Rock (00:08:10):
What people don't realize is the AI that we use every single day without even thinking about it. Okay, so I don't know Leo, I know you got iPhone 15. You I
Leo Laporte (00:08:21):
Sure did. I arrived on Friday in my hand here, right here. You know
Doc Rock (00:08:25):
What I'm saying? Like me too. Is that
Leo Laporte (00:08:27):
AI though? Really?
Doc Rock (00:08:29):
Yes. Some [00:08:30] of it is hundred percent. When you go, we were back, I think it was like iPhone 10 or whatever when we first got the picture. You take the picture and then the second layer, the picture goes, boom. Hey, look at me. I'm way queued and you thought that's it.
Leo Laporte (00:08:45):
Wait, wait, wait, wait, wait. We used to call it computational photography and this is my problem is that, and this is the same thing happened with and everything is you start talking about computation [00:09:00] as if it's artificial intelligence and it's just more computation.
Doc Rock (00:09:05):
The machine learning part of it is ai, it has the word learning directly in it. The neural part of it is ai because what they do is they take the notes and they take the information from all the other pictures available to come up with the best picture. So from way back to Microsoft Snapdragon, which is super old, you were taking all the other pictures and then coming up with information to make the next picture better. [00:09:30] Ai, I remember that.
Leo Laporte (00:09:31):
Yeah, Snapdragon, I remember that
Doc Rock (00:09:33):
Mean. People are talking about chatbots. Chatbot is the least of it and if you dig into the medical community for just like a half a toe, you even to the whole toe just to nail in and you see what they're doing with patient outcomes because of the amount of data that they can run through real quick and come up with better health plan scenarios. That's all AI too.
Leo Laporte (00:09:54):
Oh, come on. Wait a minute. I remember I B M said Watson's going to be the best diagnostician ever [00:10:00] and it was a flop and they had to stop doing it.
Doc Rock (00:10:05):
Leo Laporte (00:10:06):
M totally oversold that and they had ads on TV about it. Yeah, five or six years ago, way before it was. So maybe it's better now. Is that what
Doc Rock (00:10:14):
A hundred percent It
Ben Parr (00:10:15):
Is for sure. Better now, okay, I do this when I do these presentations and I always show where generative ai, which is all the chatbots, the chat GPTs, and it's a very tiny portion of that AI universe within neural networks, which is within deep learning, which is within [00:10:30] machine learning. It is a big umbrella term, but I think one thing we keep forgetting is there are some really cool applications to change the world. I was speaking at Climate week this week in New York or last week, and there's these entrepreneurs and all they're doing is building really advanced machine learning models to predict what ocean is happening in the ocean. So you could figure out what things need to be done to save it, figure out how to reduce carbon, and these are not the things that we're talking about every day in the news, the [00:11:00] chatbots and all that sort of thing. But they are real AI and they are really, really cool and they're really advancing really quickly and I like those things a lot the most
Leo Laporte (00:11:10):
Because easy, that's where you're going to get to the plateau of productivity with applications like that. So let's be clear though. I think that especially in popular media, AI gets confused with programming with computers. So let's be clear, what makes something ai, you said Doc Rocket's [00:11:30] machine learning.
Doc Rock (00:11:32):
Well, anytime you can take a bunch of data and have the data start to model itself based off of its previous data, that would be some artificially intelligent,
Leo Laporte (00:11:43):
I think of Alpha Zero, which was Google's program that played chess and go and gave it the rules to play Go. That's all they gave it and it played itself in four hours, a billion games [00:12:00] and learned how to play go from that. It ended up beating a very high level pro go master and to me that that's a machine that you just tell it what the rules are and it will figure out how to play and it did that. I think that's pure ai. I would agree that that's ai. No human wrote an algorithm for that.
Harry McCracken (00:12:22):
There's ai, there's machine learning which is part of ai. There's generative ai, which is a form of machine learning. So it's not one thing. I [00:12:30] think there'll be all kinds of different cycles,
Leo Laporte (00:12:31):
But is there some simple definition that we can use so that we can distinguish it from coding?
Ben Parr (00:12:38):
I give my personal definition, but there's no consensus. I usually say it is any type of computer technology that is specifically trying to mimic human behavior. So specifically trying to mimic the human ability to process large amounts of information and make new understandings from it. I don't
Leo Laporte (00:12:57):
Think that's AI necessarily though Ben, because Eliza, [00:13:00] which everybody agrees was not AI was a decision tree. The early psychiatrist that's built into emax, that was a program pure and simple, but that
Ben Parr (00:13:10):
Was not agreed, but that was not ai.
Leo Laporte (00:13:13):
It was trying to mimic humans though
Harry McCracken (00:13:15):
It was fake ai,
Leo Laporte (00:13:16):
Ben Parr (00:13:17):
Ai, it was fake ai. But for example, the chat g PT stuff, it is just writing, its kind of own coded itself. Basically all it's really doing is predicting, using [00:13:30] this constant learning that it's done. The next most likely word to say that is definitely some ai. It's just like,
Leo Laporte (00:13:37):
But that's where I think the machine learning or the training itself stuff is important. So chat G P T is a large language model and it learned from a large corpus, mostly from the internet of content how to do very good autocorrect actually what it does, it's very good autocorrect, but it learned that from a corpus of information [00:14:00] on its own. Nobody said whenever you see the, you should follow it with a noun. Nobody told it that it figured that out. So I'd agree with you. That's ai. The problem is, and Harry, you've lived through some of these, but I've lived through 'em all the AI winters, we've seen this hype cycle before. It's not the first time.
Harry McCracken (00:14:16):
Yeah. I wrote a story a few years ago about this actually really good documentary that was on C B S in 1960 about the future of AI and a lot of actually was on target except for the fact they thought that they would make enormous progress [00:14:30] by about 1964. They really envisioned getting to the stuff we're only figuring out now in just a few years. So the expectations are out of off and out of whack with the reality, but I do feel like in the long run it may not matter all that much
That we're confused about what AI is and isn't because you really know this stuff matters when we stop talking about it in the same way that people don't talk about PCs as a thing that much anymore. The internet, like Ben said, is [00:15:00] not a specific category of business. It's all businesses. 20 years ago people talked a lot about mobile. Now everything is mobile, so we don't talk about it and I really think that AI in the long run is just really sophisticated software and I would not be surprised if five to 10 years from now you just don't see that term used all that much anymore and that's when we know that it really matters.
Ben Parr (00:15:24):
Leo Laporte (00:15:25):
Yeah. That's kind of what we were saying that's that's the plateau right there in the Gartner [00:15:30] hype circle.
Harry McCracken (00:15:31):
Once it starts to plateau, it's just the air that we breathe and that's when it's most relevant. Yeah,
Leo Laporte (00:15:38):
I think it's reasonable though to try to define it and to pinpoint it because when you are in the peak of inflated expectations, there's a lot of conflation between what a computer does and what AI is and what's genuine AI and is.
Harry McCracken (00:15:55):
I mean some peaks have already come and gone. When Microsoft launched the new Bing [00:16:00] in February, I think there were people talking about how maybe Google would be a goner and this was great news for binging and they could make enormous inroads in market share very quickly. As far as I can tell, none of that has actually happened and Google has even has somewhat higher market share and bing slightly lower than it did. So that peak has already come and gone and I'm sure there are other ones that are still to come. For instance, with some of this productivity stuff such as the tools which Google [00:16:30] and Microsoft have announced and shown but are only now getting around to launching. I expect that those will not be quite as transformative as people think they are right now. But nobody knows for sure because they aren't quite here yet.
Leo Laporte (00:16:43):
The next generation that we're going to see of chat, G P T and of Google's Bard is what they call multimodal, right? Google calls it Gemini and they've started showing it to people. OpenAI is chat, G P T five, multimodal, I guess [00:17:00] even four has some multimodal. What is multimodal? What does that mean?
Harry McCracken (00:17:06):
Multimodal means that it's not just you typing stuff and getting text back. Images might start to be part of it. It's a little bit like in how in the early days of Google search you search and you've just got plain links back. You still get those links back, but Google gives you all these different things such as images and videos and so forth, and [00:17:30] little by little we're seeing AI move in a general direction where it still is kind of a chat experience, but it's not just about text conversations.
Leo Laporte (00:17:39):
But isn't hallucination still a problem?
Harry McCracken (00:17:42):
It's a huge problem. Yeah,
Ben Parr (00:17:44):
Huge. It's getting better, but it's still a huge problem. So again, it really depends on the, you use this stuff, right? If you're relying on it to write you a complex essay or to write you legal documentation, no, it's not going to work. But if you're relying on it to give you ideas, so [00:18:00] as an example, lots of students are now using it to ask questions, how do I solve this math problem? Which they used to talk to their teachers about. It's interesting actually looking at the data, but I've
Leo Laporte (00:18:12):
Been using Wolf from Alpha to do that for seven years. Is that AI when Wolf from Alpha does it? I guess it is,
Ben Parr (00:18:18):
It has ai, but it's more about the interface. Some of this stuff already existed, it's just chat G PT gave a really nice interface that everyone could understand.
Leo Laporte (00:18:25):
You might well get the wrong answer from chat G P T because [00:18:30] what I understand people like Sam Altman's saying is it's not written to be for factual accuracy. It's not when you have an auto-correct, it's not checking for fact, it doesn't even understand what it's saying. So can't
Doc Rock (00:18:43):
That's the wrong question. You get the wrong answer, but if you ask the question, right, nine times out of 10 you get the right answer,
Leo Laporte (00:18:49):
But the 10th time is a problem.
Ben Parr (00:18:52):
The thing is
Leo Laporte (00:18:52):
It's not a good search tool because
Ben Parr (00:18:55):
Humans get this stuff wrong just as much too. I'll say this, the traffic [00:19:00] for Jet G P T has increased significantly since the beginning of September, end of August. Guess why I thought it was dying. Students all came back,
Leo Laporte (00:19:07):
All the students came back of the school.
Ben Parr (00:19:11):
Doc Rock (00:19:12):
With people, right? If I ask you the wrong question or I formulate the question wrong, you're going to get a wrong answer. So here's a prime example. I used to do something really dumb. I used to be like, Hey Karen, what do you want to eat? And she's like, I don't know. Well, you can decide, okay, let's go get pizza. I don't want to eat pizza now I say, here's your choices, [00:19:30] pizza or burger, which one do you want? And she'll go Pasta, boom. See, it was not ai. I learned, I machine learned. Don't ask her an open-ended question, you're never going to get an answer. But if I give her options, she'll pick one. It's the same thing, Jack g b t works this I'm God do not tell her I said that. I'm not trying to sleep at Leo's house, but you know what
Leo Laporte (00:19:51):
I mean. It's a long flight. Let me tell you, you don't want to sleep here up straight. You set the,
Doc Rock (00:19:57):
You'll get the answer that you're looking for and [00:20:00] it's kind of the same thing. And I think the difference between Wolf from Alpha and type G B T when you're talking about a math thing is it will tell you how it came to those and you can learn from that. There is a possibility to learn from it.
Harry McCracken (00:20:12):
The other day I was trying to look up when a well-known person went to college and where he went to college and I asked Chad g p t and every time I asked that they would give me something that was wrong but close enough that you wouldn't know unless you already knew what the answer was. And I just sat there and I had it generated [00:20:30] over and over again and every single time it was wrong, but in a way that kind of made sense and that's
Leo Laporte (00:20:35):
Harry McCracken (00:20:37):
It's incredibly dangerous because it not only gets stuff wrong, it often gets it wrong in a way that sounds like it could be true. If you already know the answer then what are you doing asking chat G p t in the first place?
Ben Parr (00:20:48):
It really does depend on the kind of question you ask. As Doc said, for example, if you're asking questions about the stuff that most students would be asking about how you solve certain kinds of math problems, [00:21:00] it'll get it right pretty much every time because there's so much information about those things. It's always going to almost pull the right answer for those kinds of questions. If you're asking more specific questions where there's very little data on the internet or it's more new details about an individual person who may not be super famous, it's more likely to get those wrong. And the thing is, I think what's happening right now is that certain, especially students and certain people are working around where its limitations are and are getting more [00:21:30] out of it, and it is nowhere near perfect, but depending on the kind of work you do, the person that you are, you can still get a lot out of it. I just always marveled the fact that a year ago this thing didn't exist and now this thing exists and you could just go and brainstorm with it about ideas for your next show or for your newsletter, for whatever, and it'll be a week 24 7 and it has access to all the knowledge of human race and of course it'll get it wrong sometimes, but is [00:22:00] it
Leo Laporte (00:22:00):
Ending in 2021? Wait a minute though. But in September, 2021,
Doc Rock (00:22:04):
It stops actually the plugins that's no longer
Ben Parr (00:22:07):
Doc Rock (00:22:08):
I just did it
Where I had to do brainstorming and I did it for a class. I was teaching my people on my live stream how to do brainstorming using chat G B T, and I was telling it to use different brainstorming models like I had to do scamper model, I had to do six hat thinking model and that kind of stuff. It nails, it nails every time because you're asking a question that's like [00:22:30] that. But I will also admit to Harry's point, I was showing someone the built-in biases people say of Chad G B T, and I had to correct that. So I did an example to prove the built-in bias. I was helping a friend of mine who is a black history major and we were like, let's look up some of the top black female poets in history and it couldn't pull that list together. And she's like, see Chad g b t is filed.
And I was like, no, the internet is missing Majority of that information. You and [00:23:00] I know because we have parents and books and things like that, but the internet doesn't actually know because that shows the bias that's built into the internet. That's not Chad g bt fault. However, we can give it these documents. So we have these PDFs, let's throw these PDFs at it and tell it to re-ask the question and now that we've seeded it with the proper information, it's there. So what it does help highlight is some of the biases that are just built into the internet. So yeah, maybe Harry, the reason why [00:23:30] it couldn't find the direct answer to the person that you're looking for is the information about that person isn't well spread or whatever. This is
Harry McCracken (00:23:38):
True, this is a well-known person, but not so well that this information was all over the internet, which is why I went to chat chief PT in the first place. But I did it kind of as a test because I already knew that G P T is actually not all that great for getting information, which is difficult to find with Google. If Google can find it, then chat G P T can also find it. And if Google struggles then there's [00:24:00] nothing magical about chat G P T for the most part that makes it better at
Leo Laporte (00:24:04):
That. I think you're missing the main point though, which is that it will blindly and happily invent stuff. Google does
Harry McCracken (00:24:14):
Not links to
Leo Laporte (00:24:14):
Articles exist. People are dead when they're not schools they went to, they doesn't. This is why I think when you're generating images, you're safe because there's no correct or incorrect image. [00:24:30] As soon as you're asking it for factual information, I think you're making a huge mistake.
Ben Parr (00:24:37):
I think we're going slightly off on the direction. I'll give you an example of how I use ai and actually this came from my co-founder match at Matt prd. For anyone who wants to follow him, he and I will talk into Discord to get an audio transcript of just what happened to the week, just random notes in our head and then we'll copy and paste that entire transcript, put it to chat G p T, just ask it to organize it and give us the to-dos. [00:25:00] And then bam, I suddenly have all my to-dos, I have organized notes from the week I have organized information. I'm suddenly way more productive. I have an AI assistant that follows me onto Zoom, it's called Otter, and it takes a transcript of the calls that I have and then I can ask you a question like who said X or what were the to-dos that someone gave? And they're pretty much right because it's coming from a short piece of data and the end result is I don't have to take notes in the meetings anymore. I can just ask the auditors, like having the AI assistant there. [00:25:30] It is a really powerful tool that requires a lot of work to really learn. If you relied on it for facts and for Googling, you will fail. But if you rely on it for productivity kind of things, you could do really, really good things with it.
Leo Laporte (00:25:43):
So then let me reframe it. You were saying you just have to ask the question, right? That's actually not the answer. You just have to know what it can do and what it can't do. You shouldn't be asking it for issues of fact because it doesn't know what it's saying. But if you're asking it to summarize, one of our staffers [00:26:00] is in our discord. He says his mom is under treatment and he says, when I go to the doctors, I record the doctor's talk. I use something like Otter AI to transcribe it. I use another tool to interpret it and synopsize it and then deliver it to the family. It's very helpful. That seems like because you're giving it a corpus and you're saying just summarize this, that seems so it's not merely asking it the question in the right way. It's being, [00:26:30] I think, and this is where the hype cycle has been very harmful. It's understanding what it can and cannot do.
Harry McCracken (00:26:36):
It's really good at summarizing things for sure. It's
Leo Laporte (00:26:38):
Great at that. That's what I use it for. Yeah, that's fine. Or brainstorming because it's hallucinating, so you don't know what you're going to get, but you are a human with agency and you can reject the stuff that makes no sense, but you're not asking it as a reporter writing a story about somebody just, can you give me this guy's cv? That would be a mistake. Now [00:27:00] mistakes happen. You can go to Wikipedia and get the same and get errors as well. Humans, as you pointed out, make mistakes too. But I think because it's ai, there is at least in the culture this sense of, oh, well this is just in the early days of computers, we the same thing. Oh, it's from a computer. It must be true. In the early days of the internet, we said the same thing. Oh, I read it on the internet. It must be true. We've learned the mistake that is, have
Ben Parr (00:27:27):
We? Well, I wish we
Leo Laporte (00:27:28):
Would. We really? Well we [00:27:30] should. Some of us have hopefully. Yeah,
Ben Parr (00:27:32):
The only some of us have Leo.
Leo Laporte (00:27:34):
I know that's a good point. So that's why, and that's why I keep harping on this is there's so much hype for what AI can do that it's really important to understand as much to understand what it can't do and it really isn't serving it. The conversation when Sam Altman says it's an existential or Elon Musk is an existential threat to humans' existence that is really [00:28:00] over-hyping it. It's really important that we understand what we can do. If you're going to synopsize some PDFs, that's great, that's great, but the problem is here's the problem. That doesn't sound as cool.
Ben Parr (00:28:15):
There are lots of interesting, I think cool use cases being built. I was explaining before I saw, I was speaking, like I said before, a week ago at the Climate Tech Summit Climate week in New York, [00:28:30] and there's just these amazing companies leveraging ai, some using generative, some using traditional ai, what's called machine learning to just to do amazing things. And I think what's happening right now is that a lot of developers are trying to figure out how can I use this new technology, the large language models in particular and push 'em to their limits in specific ways. Some are like summarization, it was a common one, but I'm starting to see more that are very vertical specific and they train it really specifically to reduce the errors [00:29:00] by a lot and they give it specific information like legal cases for example, or to write really the better first version of an email.
And so I think there will be a lot more things like that and I do think that the error rate will reduce, but also is it going to replace people Googling? Not completely, no, but there is something happening. We are definitely in an over hype cycle. It always happens in that beginning. I have no idea where it will land though. And the big question is [00:29:30] always, does it land more like we can name a bunch of thundering technologies, crypto and all that recently, or is it more like we've talked about the skin, the internet or just things that are part of the basis. They're not perfect technologies, but they are a fabric of our lives and my guess is what I always do is I always look at what the gen alpha this GenOn in school is doing and their behavior is to go to BT to [00:30:00] ask questions of how to solve problems, not to write my essay. It's like how do I solve this problem? What is the method for doing it? And they do it over and over again and they're going to be doing this when they come out of college and when they come out of high school in the workplace, and it's going to be the same way that they use a calculator. They're just trained on this stuff and it's already happening.
Leo Laporte (00:30:19):
I'll give you a very timely real world example. As you know, Tesla and others are creating autonomous heavy duty trucks [00:30:30] for, I mean it's a perfect use for it. They stay on the highways. Self-driving works pretty well on freeways. 10,000 plus pound trucks. California assembly passed a bill, which requires a trained human driver to be present in those vehicles. And you know what? I'm not sure the assembly was wrong on that. It seems like you got a 10,000 pound plus vehicle hurdling down the highway at 80 miles an hour. Maybe there should be a safety driver. [00:31:00] The governor on Friday vetoed it and said, no, you don't need safety drivers. Let 'em go. Of course as California so goes the nation, if you create these trucks and they can't drive in California, that's going to be a big problem. And I don't know why Gavin Newsom decided it was a bad idea to have safety drivers. Is he succumbing to AI hype or is he right?
Harry McCracken (00:31:27):
Well, we've seen with robo taxis, [00:31:30] which are everywhere in San Francisco that the people who build the T don't necessarily anticipate every issue that's going to happen in the real world, which is why activists can do things like put a safety cone on the hood and totally flummox accrue robo taxi. So in some ways trucking's a little bit easier of a challenge just because they're spending most of the time on a highway and don't need to stop and there aren't people running back and forth, but well, I'm not an expert on this technology. It does kind of feel like [00:32:00] we'll only truly know what could go wrong when it's a little bit more commonplace than it can ever be, the more of the controlled testing that they're doing these days.
Leo Laporte (00:32:10):
The c e O of General Motors, Cruz self-driving vehicles was pissed off because the Department of Motor Vehicles forced them to slash their fleet in half after a number of collisions. He said there'd been far more collisions with humans driving these cars in the same period of time.
Harry McCracken (00:32:29):
That's true. I mean if we held humans [00:32:30] to a high standard, we wouldn't be allowed to drive either.
Ben Parr (00:32:34):
It is an interesting way we perceive things because by all measurements, most of the self-driving vehicles do perform better than a human, but we hold the machines to a higher standard for who actually, I don't really fully know the race, it's just psychological, but really humans are dangerous creatures in front of these giant death machines that we go and drive around.
Leo Laporte (00:32:59):
Does [00:33:00] a safety driver make an autonomous truck safer or less safe? Maybe he makes it less safe.
Ben Parr (00:33:07):
Oh, look, for sure. The reason he's doing it, if I had to suspect is because he just wants to make sure that self-driving trucks advance and the technology and the building and the companies there. He
Leo Laporte (00:33:18):
Probably also has donors at Waymo and Tesla. I mean it's sad, but we can no longer trust politicians to do the right thing. We know too much.
Ben Parr (00:33:30):
[00:33:30] The more interesting question is will they the house, the Senate, the legislative branch of California overturned the veto because that actually hasn't happened apparently. I looked it up in decades, but they could with two thirds majority of both
Leo Laporte (00:33:48):
Houses and it was the passed for the big majority.
Ben Parr (00:33:51):
Yeah. Well, will they let it lie? Will they not? I don't know. The
Leo Laporte (00:33:54):
Teamsters, international Brotherhood of Teamsters wanted Newsom [00:34:00] to sign the bill. They say autonomous trucks, some of which weigh over 80,000 pounds are unsafe. And by the way, there's the issue of job losses, right? Teamsters represent drivers, so obviously they're
Harry McCracken (00:34:13):
Against, although part of the reason that I think this technology does ultimately make sense is that driving a truck and when all is said and done is not something that appeals to all that many people and it's really tough to get enough truck drivers. There's a shortage, all these trucks,
Leo Laporte (00:34:28):
There's a shortage.
Harry McCracken (00:34:30):
[00:34:30] There's a shortage, and we're more and more dependent on these trucks delivering all the stuff we're ordering over the internet. So it's not like we can find enough humans to drive these trucks and I think there's a lot to be said for AI being applied to the kinds of jobs that not enough humans want to do in the first place.
Doc Rock (00:34:49):
Harry is nailing a very important point. We spend a lot of time talking about job loss in various categories and it tends to be the ones that are the loudest are ones that no young people [00:35:00] want to do. So we're not backfilling that. We're just trying to appease the crew that is still there in Japan,
Leo Laporte (00:35:08):
Unfortunately, the crew that's still there is the one writing all the articles. They're the ones at risk.
Doc Rock (00:35:16):
What they do in Japan for jobs that are becoming obsolete is the government actually steps in and it helps make sure that those people know that they're covered through the rest of their life. They just take care of it so that
Leo Laporte (00:35:28):
They don't, we don't do that here. Yeah.
Harry McCracken (00:35:30):
[00:35:30] Well, I've talked to Aurora, which is one of the big names in autonomous trucking, and basically their argument is that the people who are currently driving trucks don't have a lot to be concerned about. There will be enough jobs for them and the challenge is they're not being replenished with younger people and that is where the value comes in of autonomy.
Ben Parr (00:35:51):
It will almost certainly have to be that case. I think we're also not yet thinking we have to think more about that personal, what happens if people are [00:36:00] losing their jobs to some of these ai? I am going to give a plug. It's my fiance Deborah Chen. She's an award-winning player. She actually wrote an award-winning play called Drive. It's about a group of truck drivers who lose their jobs to self-driving trucks. Oh wow. Very relevant. Oh wow.
It won the Newcomb award over at Dartmouth. And so it goes through that personal turmoil and it asks questions about U B I and what role does that play if we do move towards it, to me it's inevitable. We're going to end up towards the self-driving [00:36:30] trucks. They do just, we are going to have a shortage of drivers. They don't have to sleep. You could drive 'em 24 hours. It makes sense from an economic perspective, so it'll happen, but we haven't thought at all about how we're going to shift our economy, not just for the self-driving trucks and truck drivers, but for AI in general as it will, at least if it destroys jobs, probably it will at least transform a lot of jobs really quickly. And people are, it's really hard to change jobs if you've been doing the same thing for two decades and then you [00:37:00] suddenly don't have that job anymore. Lots of questions to
Leo Laporte (00:37:02):
Answer. The American Journal of Transportation, the truck driver shortage continues in the us. 80,000 drivers are needed this year to make up the loss and they say by the year 2030, there'll be a shortage of 160,000 truck drivers. Now there is a difference between long haul drivers, that's the drivers that are replaced by these robo trucks and less than full loads [00:37:30] drivers, short haul drivers and less than full load. And I think that it is the case. I can't remember where the shortage is. Is it among L T L or is it long haul? I can't remember. I
Ben Parr (00:37:46):
Harry McCracken (00:37:48):
To look at, I mean long haul I think is a challenge because there aren't not all that many people who went to spend their entire life. The
Leo Laporte (00:37:53):
Road. You're on the road. Yeah, yeah.
Ben Parr (00:37:55):
It's a hard lifestyle. Yes
Leo Laporte (00:37:57):
It is. Grandpa did it. People age 55 [00:38:00] and older make up more than half of the current long haul truck drivers. So it is not a young man's profession. Now there are plenty of jobs that young people would like that might be taken over. All the good PR jobs are going to go away. I was talking to somebody who had just completed his highest level degree in actuarial science, an actuary, somebody who calculated statistics for insurance companies on what should the premium [00:38:30] be on an insurance policy so that we make money. And so there's a lot of statistics, and I didn't say this to him, he was very excited. He's getting a great high paying job because of this, but I wonder how long before AI does actuarial stuff and you don't even need, what do you need a human for? Those people
Ben Parr (00:38:49):
Should work maybe half the right, at least half the amount of humans potentially. Maybe not replace all but make
Leo Laporte (00:38:55):
What jobs are most at risk. Ben, you have an AI startup. This is something [00:39:00] you've been covering for years.
Ben Parr (00:39:03):
Yeah, I've been doing this since 2016, which is crazy when I think about it. I mean anything where there's repetition and going sifting through large of data legal is an example of a perfect one where AI will have a real impact.
Leo Laporte (00:39:16):
Although we did see that lawyer get censured by the court for submitting a brief that he used chat g BT to write, and it was factually wrong and made up precedence.
Ben Parr (00:39:26):
It's an example of using the technology wrong and asking the wrong question [00:39:30] versus there are some people building companies that specifically focus on legal and they train it on cases.
Leo Laporte (00:39:40):
If I going to, I mean college senior trying to figure out what my career should be, you say don't be a lawyer. Is that right?
Ben Parr (00:39:48):
It's more like the type, so for example, in journalism, I would be a copywriter. I already know a lot of copywriters who are losing clients and losing more jobs. It's absolutely happening because for a [00:40:00] random company in the middle for random companies, they don't actually care how accurate the thing is. They just want to have some content that ranks they don't care at all. However, if you, for example, investigative journalism, not going to be a problem. Why?
Leo Laporte (00:40:13):
Because humans need to do that. Yeah.
Ben Parr (00:40:15):
It's not about writing, it's about sourcing. It's about deal flow, it's about uncovering secrets, columnists, same kind of thing. Could
Leo Laporte (00:40:22):
AI get good enough to do that job too?
Ben Parr (00:40:26):
If so, it'll be a long time.
Harry McCracken (00:40:28):
Yeah. I mean to give a plug [00:40:30] to Fast Company, our current cover story is a big story I did on Microsoft and ai, so it's relevant to this conversation and I got to talk to Satya Nadella and all these other people, and I don't see AI successfully convincing a company like Microsoft to let it interview its c e o anytime soon, and I was only able to get that because of they had some interest in the brand and like an existing relationship with us. And I'd say that humans who are going [00:41:00] to be journalists need to focus on the stuff that humans do well, which is stuff like relationships with other human beings. That's going to be very difficult for AI to replicate.
Leo Laporte (00:41:10):
Honestly, I don't know. I don't know what a C E O does, but it feels like a CEO's job could be done pretty well by an ai.
Ben Parr (00:41:19):
Leo Laporte (00:41:21):
It'll never happen. Thousand percent CEOs are the ones that are going to keep that from happening. But honestly, middle
Harry McCracken (00:41:27):
Leo Laporte (00:41:28):
Middle management, do [00:41:30] you need middle management?
Harry McCracken (00:41:31):
Well, I think AI could do part of the CEO's job, thereby freeing the C E O to focus more on the stuff where human judgment does matter. But that's true for all kinds of jobs, probably including my job. There are certainly parts of being a journalist I could offload to a computer giving me more time to really emphasize the stuff where I do bring some unique value.
Leo Laporte (00:41:53):
Yeah, you were at the Microsoft event. I want to talk a little bit about that. Let's take a little [00:42:00] break When we come back, we've got a great panel here. I'm sorry. I often get into this a AI discussion. I'm trying to understand this and I think there is a lot of misinformation. I want to get the real information as best we can. You deserve that and when you get people who are smart like this, digging on it, it's great. I do want to talk about your interview with Satya and the Microsoft event, Harry McCrackens here from Fast Company, also Doc Rock from Aloha, [00:42:30] and my good friend Ben Parr who is the AI analyst. So I mean that's why we got you guys. We've got the good people in. So I should ask you these questions. Our show today brought to you by, well, you know the name it Pro tv, right?
We've been talking about them for 10 years since they started. Well, now they're a C I learning and that is really good news because they get the resources of a C I learning, but they keep all the things we love about it. Pro [00:43:00] tv, when you talk about talent shortages, it's not just truck drivers. 94% of CIOs and CISOs agree attracting and retaining talent is Job One with today's IT talent shortage. More important than ever that if you've got an IT team, you keep 'em up to date, you keep 'em happy, you keep 'em informed, and if you want to get into it, there's no better time to do it. 87% of companies say they have skill gaps in their [00:43:30] employees and that's where IT Pro TV and ACI I learning can come to help because ACI I learning now has something called Insights. It is a skills gap analysis tool.
You can use Insights to find out what your team knows, what they don't know, where the training needs to happen, and make sure that the training you're providing is working. This is a great tool in a quick one hour assessment, a C I learning's insights will allow you to not just see but understand [00:44:00] and fix the skills gap on your IT team. This is a solution every IT manager wants. With insights, you can identify specific skills gaps in your employees. You can see where your team's weaknesses are. You can empower your team then with personalized training, not redundant training, training for stuff they already know. What's the point stuff they need to know? Blanket training, wastes money and time and maybe even more importantly, bores your team. Keep the team happy. They love [00:44:30] it when they're learning something new. And that's what you can get from a c i.
Learning Insights offers detailed solutions, support, strategy, issues, recommendations. It'll actually work up a training plan for individuals or your whole team. You can compare your results against other organizations so you know how you're doing. You can test your skills and close the gaps with Practical Labs that allow trainees to focus on the skills they need the most and it's training they love because it's informative, but it's also engaging, it's entertaining, [00:45:00] it's training from people who with a passion for the subject and that passion communicates. A C I Learning helps you retain your team and entrust them to thrive while you're investing in the security of your business. It's a win all round 7,200 hours of content engaging content. New episodes are added all the time. They've got five, was it seven Studios running five days a week because it is always changing new tools, new certs, new questions.
A C [00:45:30] I Learning Stops its competitors with a 50% higher completion rate. And that tells you something. That tells you that these are videos that people really enjoy and helps them because nobody's going to do it just because they're entertaining, they're doing it because they're learning training solutions. Your business has been waiting for future-proof, your team and your company with insights from ACI Learning. And if you're an individual, there is no better time to get into it. And no better way than ACI learning visit go dot aci [00:46:00] learning.com/twit. If you're a company, your discount can range from 20 to 65% on that IT Pro enterprise solution plan. Here's the deal. You go there, fill out the form, get a quote tailored to your needs, it'll tell you how much you're going to save based on the size of your team. Go dot aci learning.com/twit, the best IT training, the best tools. You got to try it ACI firstname.lastname@example.org [00:46:30] slash Twitter. Please use that address by the way, so they know you saw it here. And we thank ACI Learning. They've been a great partner all year long, our studio sponsors and a great company with a great heritage. We appreciate their help. Satya Nadella is winning big techs AI War says Harry McCracken. That's a pretty good get. That's a big interview. Was this right after the Microsoft event this week?
Harry McCracken (00:46:57):
Well, this was for our current print issue. [00:47:00] So no, actually I've worked on this over the past few months and of course one of the challenges with writing for a print magazine is it kind of has to be stuff of at least a certain degree of lasting value, not going to capture the breaking news. Oh, but I'm so
Leo Laporte (00:47:15):
Glad that people like you still do these long depth profiles and you get so much value out of this. So that's exciting. How much time did you get to spend with Satya
Harry McCracken (00:47:26):
For this story? I spent, gosh, I guess [00:47:30] a little over an hour and two chunks, and I talked to a whole bunch of other people like Kevin Scott, the C T O and Mirati at OpenAI. And I actually seven years ago did a previous cover story on Nadella, and this was kind of a sequel because so much has happened in that time and I did sort of feel like just all the accumulated thinking I've done about Microsoft helped me tackle this one.
Leo Laporte (00:47:57):
How important is AI to Microsoft [00:48:00] going forward?
Harry McCracken (00:48:02):
I mean, right now I think if you ask them, they would say it's almost everything. And the thing that people latch onto is the steel they made with OpenAI to get access to G P T, which is pretty new and was a really smart prescient move. But my story talks about the fact that Nadella has been c e O for a little under a decade and much earlier in his tenure, [00:48:30] he started to reorganize the company around ai. They've been doing AI for 30 years and almost had a reputation for it, not really resulting anything, all that valuable in the real world. And he kind of put more of the people into one big team and he started a meeting of all his top executives. It was entirely about ai. And so when the open AI opportunity came up, I think they were much better poised to grab it than they might've been if he hadn't done a lot of other moving of the chess pieces [00:49:00] previous to that.
Leo Laporte (00:49:02):
I think it's safe to say, I mean you look at the Steve Ballmer era and then the transition to Satya Nadela, he has essentially pivoted the company. He's moved it into the cloud. He's done it very skillfully. He's made it very valuable. He's been a great C E O I would say. What is it almost? Is it eight or nine years now?
Harry McCracken (00:49:24):
It's going to be 10 years early next year.
Leo Laporte (00:49:26):
Yeah, he has done a brilliant job. What's interesting, this event [00:49:30] Microsoft had this week for the first six months of the year was going to be a surface event. It ended up not really, surface was almost an afterthought. They introduced Surface laptop, studio two, and they updated their surface laptop go. But really it was all about AI and it was all about co-pilot in Windows. The next version of Windows copilot of course is Microsoft's [00:50:00] ai. They said it's a single copilot across all services and apps. Dolly three, the new image generation tool from open AI is going to be part of Bing Chat. Personalized answer. I mean, this was all about ai. This was Transmogrified from a surface event.
Harry McCracken (00:50:20):
The hardware was kind of the one more thing at the end,
Leo Laporte (00:50:25):
Kind of a shock in a way. This shows how [00:50:30] first of all, surface was almost a hobby for Microsoft. Everybody wonders really why bother? You've got Azure, which is their cloud business, and incidentally smart move with OpenAI because what does OpenAI spend all that money? Microsoft gave it on Azure so it comes right back. In fact, I suspect a lot of the $10 billion was Azure credits. So AI is good for Microsoft as a business.
Ben Parr (00:51:00):
[00:51:00] It's a differentiator for
Leo Laporte (00:51:02):
Them. It's not just a product. Yeah, it's a product and it's a business.
Ben Parr (00:51:06):
It's product. It's a business. It's a differentiator. They sell to their enterprise customers. They can make it feel different than what you get on a Mac or something's. Look, I've said it here on this show before. I do think Satya Della Nadella is the best c e o public company c e o of the last decade. I agree. I'd give the word maybe the Nvidia [00:51:30] c e o for just this year. But the decade, Satya incredible turnaround, just insane,
Leo Laporte (00:51:36):
Been able to do you know who Benson Wong and Satya Nadella and Tim Cook at compared to is Sundar Pya Google who? If they're at the top, he's got to be at the bottom of the list. Yeah. Am I wrong?
Doc Rock (00:51:51):
Leo Laporte (00:51:52):
Google's Google, which should have owned this is been playing catch up on ai. It's interesting though. [00:52:00] Apple has no AI at all.
Doc Rock (00:52:03):
Don't believe that.
Ben Parr (00:52:04):
That's not true.
Leo Laporte (00:52:05):
That's not entirely true now. Well, Siri ain't the most intelligent beast in the kingdom. Where is Apple's AI
Doc Rock (00:52:13):
Secret squirrel? They're waiting for the exact right moment for it to pop out. And you know what they're very famous of? They're very famous of sitting back and watching everybody run to the dance floor and do all of this stuff. And they sit back and they watch and then they just stroll in [00:52:30] with the dopest outfit with the thick chains and the fresh pair of dunks and just takes over the dance floor. The J-Lo, they're very, very famous for that. So they're just literally just sitting back and watching. And when the Vision Pro comes out right on the heels of that, they're going to be like, oh yeah, and by the way, here's this AI thing we've been working on for 15 years. What will it
Leo Laporte (00:52:50):
Look like? What will it look like? But I mean, I guess you did argue, and I think you can argue that, that there was more and more AI in [00:53:00] the phones as in computational photography and
Doc Rock (00:53:04):
Yes, I mean, not the AI type of stuff that we're doing right now, sir. If as a Apple fan person, I won't let myself be called the boy at 57 years old, but as an Apple fan person, one thing that, one, you
Leo Laporte (00:53:22):
Can be a fan geezer at that age. You're the fan teaser is
Doc Rock (00:53:26):
Freaking Siri. Siri, you're hurting.
Leo Laporte (00:53:29):
Doc Rock (00:53:30):
[00:53:30] What is your problem? And I was like, okay, now with the new phone and with the watch, I can do all the computations writing on the watch. Let me ask this girl something. And she was like, huh? I was like,
Leo Laporte (00:53:41):
Listen, let me see what I found on the web about that. Let me ask my buddy chat, g p t about that. I don't know nothing. So they hired, apple, hired John Deandre from Google to run their AI like four or five years ago. So they have very good people. In fact, Wayne Ma, just a couple [00:54:00] of weeks ago in the information Apple Boosts spending to develop conversational ai, and there's Jan Andrea as the cover boy for all this, but it's all with everything with Apple because they're very tight. We don't know what they're doing. It's all in-house.
Harry McCracken (00:54:19):
The industry. Sorry, go ahead Ben. No, you
Ben Parr (00:54:21):
Harry McCracken (00:54:22):
I was going to say one of the interesting things at the Microsoft event is that with the Windows copilot, you are starting to see AI being built [00:54:30] into operating systems in a way that's meaningful. Instead of having to burrow through Windows trying to figure out how to do something, you can just tell Windows to do it. And I do feel like that general dental thing, the example they
Leo Laporte (00:54:41):
Showed are so dumb like, Hey, Chet, hey, binging Chet, turn on dark mode. Well, that's a dopey. Anybody could or Hey, Bing, Chet, Jennifer is Spotify playlist For me, those are trivial, useless
Harry McCracken (00:54:58):
Maybe. But if you use Windows, there [00:55:00] are like dozens to hundreds of those things you need to remember how to do. That's just a
Leo Laporte (00:55:04):
Testament to how complicated Windows has become. On the
Harry McCracken (00:55:06):
Other hand, think operating systems may evolve, but Apple pivot,
Leo Laporte (00:55:10):
Sorry, I just unplugged my headset. Go ahead, Harry.
Harry McCracken (00:55:14):
I just say I think Apple needs to sort of think about some of that stuff, and at some point it'll be way more sophisticated than what Microsoft is showing. Now, operative systems could evolve, like,
Leo Laporte (00:55:23):
Hey, make a pivot table in Excel to show me my yearly profits if I make this widget and don't make this widget, then
Harry McCracken (00:55:29):
That they're doing [00:55:30] that kind of stuff too.
Ben Parr (00:55:31):
Yeah, that's coming or it's already here in some cases, but you can even do G P T by using the code interpreter and uploading it in, and they're going to do the same kind of code interpreter stuff.
Leo Laporte (00:55:43):
I see a lot of programmers saying, oh, I can use GitHub copilot and it writes my code. That seems very dangerous to me. You
Ben Parr (00:55:53):
Code is in general really structured. It's actually in the same way that law [00:56:00] and accounting are really structured. Code is really structured. It is a great use case for some of the stuff that is just normally written that you just have to rewrite over and over again. It takes over those. You have to have the human for the more complex portions where, but you don't have to
Leo Laporte (00:56:14):
Write the login code again anymore because you didn't know do that.
Ben Parr (00:56:17):
Yeah, or just a long call to the database or whatever thing that you might have to just rewrite over and over again is
Leo Laporte (00:56:25):
One thing I've seen that's pretty cool. Copilot can look at a SQL query and tell you what it does or [00:56:30] can look at a complex regular expression and say that's what this does. That's a useful tool. That's every programmer. I see a lot of programmers swooning over GitHub copilot, so okay, what has Apple going to do with this?
Doc Rock (00:56:48):
I think some of the first things you're going to see is like you said, things that have to do with the oss and I think one of the brilliant ones that even Siri can't do right now is, listen, [00:57:00] I'm having a little bit of a weird issue. Can you contact the AL store and show me the next five available appointments? That'd be cool
Leo Laporte (00:57:07):
Doc Rock (00:57:08):
Pop right back up on the screen and be like, you can go to Al Moana, Kahala or Waikiki and you got times frames X, Y and Z. Which one would you like me to set up? You know what SS girl who right now currently is horrible. Can you give me two o'clock? Tuesday bet. All right, boo. Two o'clock Tuesday is booked up. You're going to see Harry McCracken check in 15 minutes early. Make sure you bring this pretty.
Leo Laporte (00:57:30):
[00:57:30] Wouldn't it even be better if it also knows about you? See, I want a personal agent that knows all about your schedule and your life.
Doc Rock (00:57:38):
Yeah, because it's going to already do that because your calendar is already on your computer, so we can already see
Leo Laporte (00:57:41):
That. How about if it says, I got it, this will be great Doc Rock, your laptop is about to fail. Why don't you go out and see if you can change it on the, and we'll book, we'll book an appointment for you and you'll just go get it fixed because it's going to fit a hundred percent
Doc Rock (00:57:59):
Right. [00:58:00] Read the smart stuff. Smart is kind,
Leo Laporte (00:58:02):
But whatever you do, don't let it open the pod bay doors. Okay. You have to have a key to the pod bay doors. I'm just saying
Doc Rock (00:58:10):
You can't do that.
Leo Laporte (00:58:11):
Can't do that.
Doc Rock (00:58:12):
I think it would definitely be do stuff like that.
Leo Laporte (00:58:14):
Amazon at their event this week, and I think there's rumors that Apple's going to do this. They're going to put more AI in Echo and Siri. Do we want that?
Ben Parr (00:58:28):
Yes we do. Do we want, because [00:58:30] it sounds like going to get
Leo Laporte (00:58:31):
Chatty. They're going to get chatty, they're going to start talking to you.
Ben Parr (00:58:37):
Any improvements in improvement? All I really use my echo for and my Alexa is just basic commands, weather, things like that.
Leo Laporte (00:58:47):
That's the problems why Amazon loses 10 billion a year on Echo. You're just using it as a timer and asking it for weather reports. You're not buying stuff,
Ben Parr (00:58:57):
But I should be able to ask it questions about my calendar. [00:59:00] I should be able to ask it questions about the world. I should be able to ask more complex questions and it should be able to deliver an answer. Same way we could do it in text with the new large language models. They're coming, it's all coming. Apple always waits until it's perfect and then you're like, holy crap, why did we ever use
Leo Laporte (00:59:18):
Something else? So you think that's what's going to happen, that they've been looking at this going, we can do it better. We're not going to release it until we can do it better.
Doc Rock (00:59:26):
Well, right now, guess what everybody else is doing? Everybody else is letting all the bad press [00:59:30] and all the weird stuff happening and taking all the heat and Apple just ignores all of that and then as soon as all of the bad press and now people are starting to feel a little bit good about it, apple comes. You thought that was good? Watch this new outfit. Wow.
Leo Laporte (00:59:44):
It does remind me of when we had iPod player, I mean MP three players that were this big and they were terrible interfaces and Apple comes along with the iPod and it's like, oh, that's not an innovation. We've had these before but you've made it better so maybe [01:00:00] can do that again. They haven't done that in a while. I do think
Harry McCracken (01:00:03):
That Apple will probably focus on actual useful stuff like the new Apple watch. It has this double tap where you can tell that you're doing this and if you're not trying to make that gesture, it knows you're not trained to do it and that's totally machine learning because it's based on all this data about stuff like the exact movement of your fingers and stuff like they
Leo Laporte (01:00:29):
Do that with and [01:00:30] so forth with fitness, with their workout stuff. If you get on an elliptical, it knows you're on an elliptical because they have done a lot of machine learning on how people move When you're on a rower, it knows the difference and so it will actually say, are you on elliptical right now?
Harry McCracken (01:00:46):
Look, I don't think Apple, actually Apple did not use the term AI once at its event the other week as far as I remember, but it was full of small useful things that they couldn't have done without one form of AI or another.
Doc Rock (01:00:57):
They have not uttered word AI and anything [01:01:00] at all for the last, however, they've completely never said the word and that's how I know they're working on something. Dope.
Ben Parr (01:01:07):
Leo Laporte (01:01:09):
Go ahead Ben.
Ben Parr (01:01:11):
No, it's is the Apple playbook to a T. It
Leo Laporte (01:01:14):
Just is. I love this. In the information article, big long article about Apple and working on AI and the one line an Apple spokesperson declined to comment. That's all Apple had to say about that. Yeah, they're smart. You know what? They're smart. Well, everybody else is hyping it up and taking [01:01:30] the arrows because they're the first over the hill. They're waiting and waiting until the smoke clears and they're going to release something. Maybe not even call it ai. Just more useful, just better products.
Ben Parr (01:01:42):
I mean, look, in the end, I think actually looking at the business models of these big companies really shows you the direction things will go. Microsoft needs to excite developers using Azure and other things. It makes sense for them to be early to these things because now more developers are using Azure and the new AI stuff with open AI [01:02:00] and things like that versus for Facebook for example, they don't want to sell a large language model to the public, so that's why they made it open source for them. If they could improve it so that they could use it within Messenger and things, that's their win. And then for Apple, they just need to sell more devices and for them they're already selling lots of devices. They're going to make a bigger leap to make it so that it feels like, oh, I need to have this new device because it can scan my brain and knows exactly [01:02:30] what I'm thinking and now I no longer need a phone that's a big square. It's like a little dot or whatever crazy thing they come up with. That is always what they do. It's just because their only goal is to sell more devices and it gives them the capability to be patient.
Doc Rock (01:02:44):
Yeah. Yes. Same with the headset, right? We're not going to call it a VR headset because it automatically aligns it with all the other stuff that's out there that's fun and queue and gimmicky, but not really do anything. So they're going to [01:03:00] use a whole different terminology and they will even take the AI word and come up with something sound. All the tech people are going to say it's dumb and we're all going to say it's silly, but then a whole bunch of people will use it and love it and then we'll continue to say it's dumb and silly, but people make fun of their names when they switch to parks and they went to cats, but you know what? You didn't have to remember what number you were sitting on and all the other operating systems of I'm on 5.3 0.1 point x and they're like, oh yeah, I'm on the one that's named after the place [01:03:30] in California. And you can remember that there was a funny bit that Kimmel used to do and it's like what version of Windows do you have? Internet Explorer,
Leo Laporte (01:03:46):
You talk in the Harry and your Fast company cover story with Satya, how Microsoft was somewhat surprised by how far OpenAI had come with chat G P T. Did it take 'em off guard?
Harry McCracken (01:03:59):
I [01:04:00] think so because Microsoft had also invested a huge amount of effort in building an L L M and they had infinitely more resources than OpenAI and they didn't make as much progress. I mean, we know that even OpenAI was surprised by how much progress OpenAI was able to make, so it's not surprising that Microsoft was as well.
Leo Laporte (01:04:17):
So Apple's focus, I would guess I would say, and this would be very appley, is not to focus on the technology but to focus on what it can do [01:04:30] and how it's going to make your life better. And I think that's smart because you also avoid the pitfalls that chat G B T has had a little bit where if you focus on the technology, then when it doesn't do exactly what you want it to do, it becomes all about the technology not being very good. So focus on simple outcomes that you can achieve. How is Microsoft approaching AI in its future? It's certainly spreading AI on every product that it makes. What is their [01:05:00] point of view on this?
Harry McCracken (01:05:02):
Well for now, I think they're focused on doing the types of AI that you would expect Microsoft to do. So operating systems are an obvious area. Productivity is a huge place for them in search. They're a distant number two, but at least they've been in the search game for a long time and we know it matters. So I think all of those, these are powerful tapestries for them. Azure is also huge for them. As everybody else in the world starts [01:05:30] to build ai, most of it will run on somebody else's cloud and Microsoft as the opportunity to catch up even more than it has with a W Ss. By having the best ai. There are places where logically they might do some stuff like gaming where we haven't heard as much, but I think it totally makes sense for them to put the most energy right now into the areas that most defined Microsoft and where there's the most opportunity to do transformative stuff and probably also the most opportunity [01:06:00] to make money.
Leo Laporte (01:06:02):
Well, and Microsoft's in a great position because really their biggest, most important product is cloud is Azure and all of these AI models use the cloud. Azure is going to become the premier platform thanks to open AI for AI development. So they
Harry McCracken (01:06:22):
Win. If they have the best flavor of G P T in the cloud, that's a huge win for them.
Leo Laporte (01:06:29):
There [01:06:30] is an advantage, and this has always been this way. We've talked about this with Siri and Google assistant and Amazon's echo. There is an advantage to operating in public. One of the advantages, the disadvantage obviously for OpenAI is that they're going to get a lot of arrows is they're the first over the hill, but they also get a lot more information by operating in secret, not being public. Apple doesn't get as much feedback on what it's developing.
Doc Rock (01:06:59):
Well, yes and no [01:07:00] because a lot of people are getting the feedback from what they're typing in and doing that. But like you said, where Apple is going to put his stuff on, they're going to put their chickens on creature comforts, almost every nook and cranny of iOS, of the built-in things that are in iOS and watchOS has machine learning, which Harry said is a subset of ai. So even now that I can do this to change things on my phone, I mean on my watch, I am using stuff that humans do [01:07:30] on a daily basis search. If I say Cat on my photo album, I can find a cat in my photo album, even though it's mostly dogs. Sorry, people. If I say, Hey, ramen from Osaka in 2023, I can get that exact feature.
Leo Laporte (01:07:44):
It's a wonderful feature. Yeah, yeah.
Doc Rock (01:07:46):
You know what I'm saying? Right. So that stuff, all of that machine learning stuff that's built into basically every nook and cranny of iOS and now that the processor which Apple is doing different from everybody else is they're focusing on processors, [01:08:00] what their sips can do. That's also going to make them a little bit different because they're in control of the hardware and the language, so it's going to take a while. Yes. And it looks like they're sleeping, but they're sleeping with one eye open.
Leo Laporte (01:08:14):
Historically, innovation in technology has come not from the big companies, from the incumbents, but from the little guys. Is AI going to be a different story where because it's so expensive to create these LLMs [01:08:30] and it takes a lot of people and a lot of technology and a lot of brains, is it going to be the big guys that are going to win here? Is it the Googles, the Microsofts, the Apples, the Amazons, or are we going to see innovations? Ben, you have a small AI startup. Are we going to see innovations from in innovators, from garages in this
Ben Parr (01:08:51):
A ton, but I would argue, so for example, the most innovation in this space has come from OpenAI in the last year and [01:09:00] they were a small company. They
Leo Laporte (01:09:01):
Were a startup.
Ben Parr (01:09:02):
They were a startup a year or two ago. It's just because of the interest in the space and the usefulness of some of this technology is probably the fastest growing valuation I've ever seen maybe in my lifetime just skyrocketing for a reason. Big companies always win in some way, like Nvidia is a huge winner out of all this. No matter what they win, they're the
Leo Laporte (01:09:26):
Levis great. They're selling the jeans to the gold miners, right? Yes. Yeah.
Ben Parr (01:09:30):
[01:09:30] But there's innovations. There's these other companies that we're small startups and then they became Anthropic
Leo Laporte (01:09:37):
Anthropics. Pretty impressive. Although that's a big Google investor investment.
Ben Parr (01:09:42):
I mean this is how Google makes sure that they have some plays in all these other areas. A lot of
Leo Laporte (01:09:47):
Irons in the fire. Yeah,
Ben Parr (01:09:48):
Yeah. Mid journey. There's these ones and Mid Journey's never taken any money from any mid
Leo Laporte (01:09:52):
Journey's. Fascinating, right? That's a garage startup. He didn't want any investment
Ben Parr (01:09:57):
And knowing him, I [01:10:00] will be shocked if he ever takes money, just watch tomorrow. They announced they took money. They won't. They won't. They're real. And look, they're beating out thousand person companies on the image front, so for sure. Lots of startups, lots of areas. Now what we're going to see is how could you use this new set of technologies for daily life and they're going to be a couple of startups that crack that nut and they're going to be billion dollar companies. Yeah.
Leo Laporte (01:10:23):
The creator of Mid Journey, David Holtz self-funded. He had sold a startup, I don't remember what it was, but [01:10:30] took his money and started
Ben Parr (01:10:31):
Leo Laporte (01:10:32):
Oh, that's right. Magic
Ben Parr (01:10:33):
Leap. I remember I went in to go see the first version of Magic Leap when I was at CNET and getting to see the whole thing and it's super cool, but he's, he's like
Leo Laporte (01:10:42):
This not magic Leap, not magically leap motion to the other one. Leap motion. Leap
Ben Parr (01:10:45):
Leo Laporte (01:10:46):
Had that thing where you make gestures in front of your computer and it didn't take off even though it was a very interesting idea.
Harry McCracken (01:10:54):
And open AI's, C T O Mirati is also a veteran of leap,
Leo Laporte (01:10:59):
So [01:11:00] Oh, interesting.
Harry McCracken (01:11:01):
Two of the most important people started out there, even though that technology didn't go anywhere, it apparently was a hotbed of interesting talent
Leo Laporte (01:11:08):
And I think Leap Motion must've been AI as well. It had to understand what figuring where your hand was. Yeah.
Ben Parr (01:11:15):
Interesting enough technology to acquire and money
Leo Laporte (01:11:18):
For David Story. He doesn't need a VC in his life. How Mid Gurney's founder built an AI winner while rejecting venture capital. He's getting now 200 million in revenue, 40 employees. That's pretty impressive. [01:11:30] Okay, so I answered my uncle. You're right. There's a lot of startups doing this. Very interesting. Let's take a little break. We got lots more to talk about Doc Rock, he's all in purple. People say, where did he get that amazing microphone Muff
Doc Rock (01:11:49):
Reporter store.com. It's in the Netherlands and it's only like 20 bucks for the cuff, but it's like 20 bucks for the shipping but
Leo Laporte (01:11:57):
It weighs two ounces. It's nothing.
Doc Rock (01:11:59):
Yeah, it weighs [01:12:00] not even that. So now Smart. I ordered them like five at a time so I
Leo Laporte (01:12:04):
Can And you're wearing a very attractive, I might add. Thank you.
Doc Rock (01:12:08):
Leo Laporte (01:12:09):
That by the way. AI generated. Really? Yeah. Well so for all my skepticism on AI and I am very skeptical, we have a guy, I've talked about him before, Anthony Nielsen, he's going to do on our club twit, he's going to do an A m a soon who has decided to become our AI guy and he's thrown himself into this and he's doing amazing [01:12:30] stuff with Mid Journey with Claude from philanthropic. He's tried 'em all. He's doing all sorts of interesting things. We have an AI Leo in our discord. He's a wizard and so he does a lot of our graphics. He's a very talented designer and I don't know how much of that Anthony was generated by ai and how much was you? I think it was a human machine collaboration, but I like it. He did a good job.
Doc Rock (01:12:58):
You're going to need a word for that. Not hybrid [01:13:00] that just goes straight to car. You need a word for the
Leo Laporte (01:13:02):
Human, the beast with two backs.
Ben Parr (01:13:06):
Yeah, we need a catchphrase for the human machine collaboration. Yeah. H M C doesn't cut it. No, it
Don't. The human made that
Leo Laporte (01:13:12):
A human made that help any more than you say. Oh, the computer helped. No, a human made it. Right. That's bonito. Boni Gonzalez is running our board who is also a genius at this stuff. Yeah, you're right. Bonito. When you make music a human made that with your synthesizers, then [01:13:30] nobody says, oh yeah, that was a machine that made that. No, you made that.
I played the guitar. The guitar didn't play itself.
Leo Laporte (01:13:37):
Guitar didn't make the music. Harry McCracken makes the music at Fast Company. He is the global technology editor there and I'm glad they're giving you time to do long stories as Steven Levy called it Slow journalism. That's good. It's a pretty cool gig. Yeah, we need that stuff. I'm so glad and you do a great job. So nice to see you. Of course, as always. And [01:14:00] Ben Parr, newly engaged. You're getting married you said in April.
Ben Parr (01:14:07):
Yes, I can.
Leo Laporte (01:14:09):
I met your fiance solo Planning. Planning. She's wonderful as you say that. Very talented playwright. That's neat. Congratulations.
Ben Parr (01:14:19):
Go ahead. I'll leave it at, I never knew there was so much to think about with things like cakes.
Leo Laporte (01:14:26):
Have you gone to the tasting yet?
Ben Parr (01:14:29):
That's next [01:14:30] week.
Doc Rock (01:14:31):
Are you honeymooning in Hawaii?
Leo Laporte (01:14:33):
Ben Parr (01:14:36):
I might need to come sooner and come visit you before end of year. I would like somewhere warm and where I can sleep on a beach
Leo Laporte (01:14:45):
Oahu's. Wonderful. Up at the North Shore. So how has Maui's fire affected you, doc
Doc Rock (01:14:54):
Personally? Not directly in the way you would think, but I mean I know so many people. [01:15:00] We are very connected individuals. A bunch of my friends had restaurants that had just completely gone, I'm sorry, one of my best friends here who is actually a chef here, him and his wife have been running a lot of the getting food and supplies to Maui from there. So it's a connected of Hawaii being a two degrees of separation. It doesn't really affect Oahu per se, but it definitely does have an emotional cost. And then knowing that our state, [01:15:30] it's going to take us well over 15 years to even attempt to rebuild Lana and it will never be the same. So I think there's that aspect of it and we are very heavily affected by tourism revenue, much to the chagrin of many people that live here. And so that's going to affect the economy of Hawaii for probably a good seven to eight years.
Leo Laporte (01:15:53):
We were talking about this with Johnny Jet on the one hand, the Maui Tourism Bureau saying, come back, we need [01:16:00] need you. We need the tourist dollars. But the locals are saying, don't come here. We're healing. I don't know. I love Maui. It's beautiful.
Doc Rock (01:16:13):
Beautiful. That is really a tough one, and I have so much love and respect for the community here, but I think a large portion don't understand that without that tourist revenue, you're not going to do well, right? So it kind of sucks literally as [01:16:30] tech people, Hawaii has a very good tech backbone and we have been saying Stop letting tourism be our number $1. We can use some of the tech stuff here because we're the perfect place between Asia and the US as far as timing and things like that. And we end up shipping most of that brilliance out and now more than ever, we could use the dollars that would've come from our tech sector. And so those of us that have been in tech sector here forever have been saying, we could help you stop leaning on the tourist dollar, but the tourist dollar is so good.
[01:17:00] It's hard. It's like, yeah, I was just making money to pay myself through college. Okay, now you're finished college, why are you still? Well, because it's good money. That's why. So it's one of that weird scenarios. So that was me. I was doing nightclub work because I just wanted to pay for college. I was making a grip and I ended up doing it well into my forties. I was making a grip so I knew I should leave and I should not be doing this, but I was making so much money I didn't want to. So yes, that's the exact
Leo Laporte (01:17:29):
Thing. [01:17:30] I mean you can certainly go to Yle or ha's one of my favorite places in the world on Moo Island. It's okay. And be respectful, right? That's the key. Be respectful.
Doc Rock (01:17:43):
Be respectful. Our biggest problem with turns, and I think the logo's biggest problem with tourists is that people like to come and do dumb stuff. When the volcanoes were growing off, you got people going up there and taking pictures of the moon in the volcano. It's a religious thing to people here it is a spiritual [01:18:00] situation. Madam PE is not to be toyed with. So going up and taking a picture from the gram with your butt sticking out with the lava glory, that's disrespectful.
Leo Laporte (01:18:10):
Oh, you mean that moon? I thought you meant the different moon, different moon. Oh my god, that's so stupid.
Doc Rock (01:18:17):
Some guy decided to relieve himself into the, oh, no trade. These are sacred lands. People don't do
Leo Laporte (01:18:23):
That. That's one of the things I love about Hawaii is the Polynesian culture [01:18:30] is so amazing, so beautiful, very much so, so warm and ohana and that whole spirit and the aloha spirit. If you respectful, go visit. It's wonderful. It's beautiful. Just respect the place in the people because it's not Disneyland folks. It's a real place with real people,
Ben Parr (01:18:51):
Tourists, you cannot put out a volcano with you.
Leo Laporte (01:18:54):
Pete, do not
Doc Rock (01:18:55):
Leo Laporte (01:18:56):
Percent AI may be number two, but not in the volcano. Please, [01:19:00] please. Alright, we are back and we have more news. I don't know. I'm so sorry. I got the new iPhone. I'm ashamed to admit it, but I do it for work. It is like incrementally better than the old iPhone. Titanium's pretty, although a lot of YouTubers are scratching and breaking and nicking. Some say that the color, [01:19:30] this is the natural titanium. Some say that the colored ones can be scraped. I don't care. You put it right in a case right away and then other people are pointing out, well, if you're going to get a case, don't get apple's. Fine woven case, their replacement for leather because it's easily marred by scratches. I fix it, had a interesting microscopic take. What it tear down of the believe it or not, of the fine woven case where they showed, actually let me show you because it's a pretty [01:20:00] wild picture.
The fibers in this are one 12th the size of a human hair. So let me show you the pictures using a microscope from evidence scientific. So there's what it looks like in your hand. There's what looks like under 52 magnification. You can see each strand is composed of many strands. It's kind of like rope filament. [01:20:30] That black thing is a human hair. And then you can see the nylon filaments. They're one 12th the size of human hair, six microns that are woven together to make kind of ropes that are about twice as big as a human hair and then they show what the damage happens to it. Let's see here. It's key scrape. So this is the normal untouched fiber sliced [01:21:00] of the knife you can see, but this is scratching and all it does is it doesn't break the little nylon fibers. It rearranges them so it reflects light differently. So there seems to be some evidence that you could actually somehow with rubbing cloth on it or something, rearrange these to make it look more normal. They tried damaging in a variety of ways with coffee and so forth. It's going to hold a stain. Do not get [01:21:30] hot sauce on your fine woven fabric. That's not a good idea. I have to say having I bought one of these and I wouldn't recommend it at $59. It's a lot of money and very expensive for something that is not an exactly
Doc Rock (01:21:47):
Loops. Don't be just because I personally don't like the feel, but I got to say people are stupid. It really irritates me because listen, I understand. I've been in the marketing department. [01:22:00] This stuff does get clicks. If you scratch leather, does it scratch? Absolutely no, but
Leo Laporte (01:22:06):
That's the beauty of leather is it takes on a patina of wear. I don't want a patina of wear. Am I fine woven case,
Doc Rock (01:22:14):
But that's what it's going to do. It's fabric fam. You don't grow scratching, but
Ben Parr (01:22:20):
It's not the worst thing character in your case.
Leo Laporte (01:22:23):
That's fine. I'm replacing it with leather. So
Ben Parr (01:22:28):
Steve, I'm going to take this bottle of hot sauce and [01:22:30] I'm just going to pour it on my phone. No,
Leo Laporte (01:22:31):
Look, it's a little tvm
Ben Parr (01:22:32):
Leo Laporte (01:22:33):
Oh my god.
Doc Rock (01:22:35):
Super resilient. And then let me scratch the color on a titanium. Will it come up? Yes. If you scratch the paint on your freaking car, it'll come up on people. The stuff that people do to try to make a mountain out of a molehill, it's so obnoxious right now and it creates a necessary, another argumentative, divisive [01:23:00] type thing that our culture doesn't need. Stop looking for fights. Legit. Stop looking for fights. This
Leo Laporte (01:23:07):
Is sad. This is, and I know you're a YouTuber, but this is the par problem with YouTube is it? Inc. Hundred percent. Itens called out sensationalism. I mean, if you want,
Doc Rock (01:23:17):
I called them out. I called them out a hundred percent Leo, because I think it's disingenuine and I can name you two channels right now. Every Apple product, this is the cycle. This is the greatest thing they've ever come. Oh, guess what? This is what Apple's not telling you. Oh, don't buy this [01:23:30] because Apple's screwing you. Oh, we bought all of them and these are the greatest products. Oh, you know what? 8 million
Leo Laporte (01:23:36):
Views later, they
Doc Rock (01:23:37):
Have eight. Have eight to 10 videos they make for every single Apple product, and it's the exact same video.
Leo Laporte (01:23:42):
Doc Rock (01:23:44):
It's obnoxious, and you know what the 15 is in? Every phone is supposed to be incremental because the 14 people, Leo and I aren't supposed to buy a 15. This 15 is for the 11 and 12 people. To them it's a massive upgrade. You're [01:24:00] not supposed to buy a phone every year, but somehow people like us in the tech cycle who buy a phone every year, and I'm literally sitting on the table 13, 14, and 15. We're a rare breed and we're not supposed to talk about it from our lens. We're supposed to talk about it from the lens, from the average person, because your audience is who you're supposed to be reporting to and people are like, oh, I'm not going to buy the Apple Watch two because it's not a significant upgrade over the Apple Watch Ultra. You weren't supposed to replace the $800 watch. You never know
Leo Laporte (01:24:30):
[01:24:30] There is. Well, who's really big on the link bait? The Wall Street Journal inside's Apple's spectacular failure to build a key part for its new iPhones. Oh yeah, this one. The point of this is they tried to replace the Qualcomm modem chip. You remember they went to Intel to do a modem chip a few generations ago and it was so slow they had to slow down the Qualcomm chips, so the phones were performed equally [01:25:00] for a long time. Apple's really wanted to make everything inside of all of their hardware. They don't want to go to Qualcomm to make a chip. In 2018, Tim Cook, according to the Wall Street Journal, gave marching orders to create a modem chip. They hired thousands of engineers so they could get Qualcomm out of the iPhone and have it all Apple. They had hoped to have it ready for the iPhone 15, but Tate tests late last year. [01:25:30] I'm quoting the journal, found the chip was too slow and prone to overheating. More and more importantly, its circuit board was so big it would take up half the iPhone making it completely unusable. They pay a lot of money about 7.2 billion last year to Qualcomm for, it's not just the hardware. They also pay for licenses because Qualcomm owns a lot of these technologies and licenses them out to Apple, [01:26:00] so there's still a Qualcomm chip in there.
Ben Parr (01:26:07):
I got to pull my best doc rock, which is like, this is why it's called r and d. You spent buddy to try to figure these things out. It literally, someone read it, pointed out, it literally just says in the article they're still working on it because they're still working on it
Leo Laporte (01:26:23):
Ben Parr (01:26:23):
Year. It's a really hard complex problem that Qualcomm has done a really good job of solving. [01:26:30] I expect one day Apple will go and solve it. This is not a crazy article. Apple has tons of money to spend on lots of r d in research and not everything will bear fruit and also not everything will bear fruit immediately. Yes. Click bait. Click bait. Yeah.
Doc Rock (01:26:48):
If you have $27,300, Leo and I asked you to give me seven bucks, would you give it to me? Yeah. Alright, so if you got 2.73 trillion valuation,
Leo Laporte (01:27:00):
[01:27:00] They spent the equivalent of $7 to do this. I get what you're saying.
Doc Rock (01:27:05):
Shut up. I hate when they do that
Leo Laporte (01:27:09):
Here. They say, the journal says that it is tricky to make these chips unlike A C P U where it's pretty straightforward thing. These guys got to work with all these outside environments, five G and two G, and three G
Ben Parr (01:27:25):
In every country
Leo Laporte (01:27:26):
Too. In every country. And it's complicated.
Harry McCracken (01:27:29):
I do [01:27:30] think Apple was not that naive. They knew what they were getting into. Even if it's taking longer, that probably doesn't come as a huge shock to them. And if the disaster would be, if they at some point killed this project and had to go with Qualcomm forever, but assuming they actually do make a modem at some point, it'll be worth it on a variety of fronts, including the fact that it'll save them money over time. And they really do not want to be dependent on Qualcomm, who they've had legal tussles with in the past and [01:28:00] with the chips they've made with their CPUs. We've seen how much value that has. And I mean, when the Mac switched over to an Apple chip, it totally re-energized those devices and Apple can afford to be really patient if it pays off eventually.
Doc Rock (01:28:15):
Yes. I was going to say, Harry, I wonder if Aaron and Yang mentioned how much they saved on not paying Intel if that's more than the 7 billion they spent. You know what I mean?
Leo Laporte (01:28:25):
Right. One thing Apple was looking at, according to C N C, they [01:28:30] were working with Goldman to do a stock trading feature for iPhones like a Robinhood built into your iPhone. Sure, good money, but I don't know, should Apple be in that business? They decided not to do it as the market slumped, so they thought the timing's not right, but they were prepared apparently to put a Robinhood in your iPhone. Not Robin Hoods, but Apples with Goldman's. Goldman has had [01:29:00] some trouble with, they claim they lost a lot of money offering the Apple card. They wanted to get out of that business, right, Ben?
Ben Parr (01:29:07):
Yep. The Apple card Goldman has been slowly moving away. For those who don't really understand Goldman, Goldman has always been for these high-end wealth, these large working with other enterprises, haven't not been a consumer bank. Banks like Citibank and Wells Fargo, their bread and butter is this business consumer, [01:29:30] this stuff that they do with consumers at Goldman Triad. There's a lot of complicated stuff with issuing credit cards, with chargebacks, with all these other things that add up to costs. And if you really want to do it, you've got to keep doing it for a very long time. And honestly, I think Goldman's gotten, they've evaluated the thing and they've just decided they should probably pull back some given with how the market's going, how deep the entrenched the competition is. [01:30:00] Yeah. I expect at some point that the Goldman card will, Amex will take over that business or someone else and it'll be fine and Apple will definitely remain in the card business with somebody. What that exactly looks like, I don't know. So I wonder if it's also tied to the fact that Goldman wants to do maybe a little bit less consumer, that they're also like maybe we shouldn't do the stock trading platform with Apple at the current moment.
Leo Laporte (01:30:25):
Harry McCracken (01:30:27):
The terms on the Apple card are actually quite fair to the consumer, [01:30:30] which is an unusual thing in credit cards and apparently one of the reasons it turned out not to be so attractive to Goldman.
Leo Laporte (01:30:36):
Yeah, I mean, I actually let Apple hold my Apple cash in a savings account that has 4% a p. I thought, well, that's better than I get from most banks and why not? It is just sitting there.
Doc Rock (01:30:49):
You know what happened right after that Leo? All the banks started offering, I
Leo Laporte (01:30:52):
Doc Rock (01:30:54):
It's like, oh my
Leo Laporte (01:30:55):
God, we better up those interest rates. Apple's eating our lunch.
Doc Rock (01:30:59):
I [01:31:00] always held a lot of money in my PayPal account because of Super chats and things that come from YouTube, all my YouTube affiliate stuff. I started routing it straight to the Apple apple savings right away, and hey, that joint kind of halfway paid for my trip to Japan. I know
Leo Laporte (01:31:15):
I could live on 4% in my old age that would,
Doc Rock (01:31:18):
I'm like, listen, I would bend and I really hope Amex takes it over just because I feel like they're better at what they do. But yes, I love my Apple card for right now,
Leo Laporte (01:31:30):
[01:31:30] According to C N B C, their source said the infrastructure for an investing feature is mostly built and ready to go. If Apple ever decides to move forward for it with it. I feel like that's a risky business for them to get into because if people lose money on meme stocks and stuff, they're going to kind of blame as they do. Robinhood blame Apple, they blame Robinhood for a lot of that. How easy it was and how cheap it was to trade
Ben Parr (01:31:57):
Also opens you up to lots of lawsuits. Imagine [01:32:00] if algorithmically the Apple stock is at the top and Google is seven down or Microsoft is eight down,
Leo Laporte (01:32:08):
But that's the way it is, Ben.
Ben Parr (01:32:12):
No, it needs to be Microsoft at that top iPhone.
Leo Laporte (01:32:15):
Yeah. Let's see what else is going on? Let's take a little break. We'll come back. We'll talk about the Amazon event. We didn't talk about that yet. And a big departure from Microsoft. Amazon Snags, Panos Pane. More to come with [01:32:30] an excellent panel. Ben Parr, it's great to see you as always. The Techn, Harry McCracken and the Doctor of Rock, Mr. Doc Rock, Amazon's hardware event. I did not get invited, but I never do and it's not a streamed event. So did anybody watch anybody up on what Amazon? Who Caress, right?
Ben Parr (01:32:57):
Leo Laporte (01:32:58):
Cares, I'm sorry, but [01:33:00] new the chat capabilities, they got wired, wired live, blogged it, so they're going to put AI in Amazon's echo. Not a surprise. We'll see how that works out. Everybody's going to do it. Apparently a new Echo show eight, which looks like the old Echo show eight, but it's new Echo. Does anybody own a pair of these echo frames? [01:33:30] Y
Ben Parr (01:33:31):
Someone out there just like tweet X at us if they have the echo frames. I am deeply curious of the answer to that question.
Doc Rock (01:33:39):
Why? I tried 'em and it's weird. Just believe it at that. I tried at podcast movement. It is very, very weird.
Leo Laporte (01:33:47):
Oh, were they there showing those off or did somebody have 'em?
Doc Rock (01:33:50):
Yeah, no, Amazon had a big booth. Yeah, because they're trying to get more people to do Amazon podcasts because well, everybody else is doing podcasts,
Leo Laporte (01:34:00):
[01:34:00] Right, right. So how does that work? Is Echo in the frame and you talk to her? Is that what's going on? Yes,
Doc Rock (01:34:10):
Is in the frame. That's it. And one thing that's kind of cool, you can answer the phone, you can listen to your podcast or your music or whatever while you walk around in the frame, but the listening experience isn't that dope because it's not a bone conductor. It's a little speaker. I don't know it. It's just weird. I don't think
Leo Laporte (01:34:30):
[01:34:30] It sounds like that, right? It's just a little Timmy
Ben Parr (01:34:34):
Has been doing this and they have a bone conductor, that bone conductor they've been doing for years.
Leo Laporte (01:34:40):
That works well. It works
Ben Parr (01:34:40):
Pretty well. Yeah, I have one of those don't, one of the issues with Apple stuff look is useful. A Kindle is useful. How often are you going to replace those things? You could have one of those for 5, 7, 10 years before you needed to replace them. There's new thing that's like, wow, I have to go and change out the [01:35:00] thing that listens to me for another thing that slightly listens to me better.
Doc Rock (01:35:04):
Right. Well, the only thing that makes sense, okay, the firesticks make sense because now they added 'em to wifi six helpful because your streams and your speeds and stuff. So look guys, all we did was made the processors better. That's fine. You don't have to make up stuff. Just tell us that you're just doing it for slight spec bumps and we would be cool with that, but that's never what they tell you. Can I
Ben Parr (01:35:26):
Give one? I agree, but one counter is that [01:35:30] almost every TV is already built in with all the smart streaming, so you don't need a Fire Stick or a Chromecast anymore to stick in. Yeah, all the Samsungs come with everything built in.
Leo Laporte (01:35:41):
Roku I think saw the writing on the wall. They started building Roku into some tv. A lot of TVs.
Doc Rock (01:35:48):
Amazon builds it and the TVs them and Toshiba make. They
Leo Laporte (01:35:51):
Had Amazon TVs there. Yeah.
Doc Rock (01:35:53):
Yeah. That's why it's funny. But I agree with Ben at this point in time. It's all built into the smart [01:36:00] stuff, so I haven't fired up my stick in forever.
Leo Laporte (01:36:03):
Exactly, exactly. They're
Harry McCracken (01:36:04):
Throwing a lot of stuff at the wall, but it does feel like if these products ever change in a big way, it might have more to do with the ai, which is based in the cloud rather than the hardware. And there's a good chance that your old Amazon hardware will support this new ai and so there's not a huge incentive to get the new stuff. I mean, I'm still using Amazon's original smart speaker [01:36:30] whose name I'm trying not to say because it's sitting right behind me and it still looks fine. Don't wake
Leo Laporte (01:36:35):
Doc Rock (01:36:39):
Aaron Soundbar though, only for just one TV that I have that I don't have any sound connected to it and buying one of these at one 20 versus buying two Apple baseballs at 200. That intrigues me. But I know for a fact that the two little Apple [01:37:00] baseballs are phenomenal for watching tv.
Leo Laporte (01:37:05):
Amazon is hiring, so this was kind of a one-two punch. First we heard that Panos Penne, who was a strange fellow who's been, he started working on mice at Microsoft and 20 years later he was running Windows and hardware and is very famous for his weird kind of channeling Steve Jobs keynotes in which [01:37:30] he was always pumped for whatever it was. He's leaving Microsoft probably by mutual agreement. I've been trying to find out was he pushed out? Did he want to get out? I think it was by mutual agreement. He's going to Amazon to replace the retiring. Dave Limp who ran Amazon devices. He was the guy who did the Amazon presentation. So a pane I guess next year. Do you think he will attempt to put his stamp [01:38:00] on what Amazon's doing or is Amazon such a big battleship that it's going to be hard to change? He did a lot of things at Microsoft that maybe some people did not agree with. I don't know.
Harry McCracken (01:38:12):
I think he does have a little bit of clarity of vision, which conceivably is helpful given the degree to which Amazon is usually about flicking spaghetti and seeing what works. Microsoft
Leo Laporte (01:38:22):
Vastly didn't do that. I mean Surface was a very clear product line he ran that.
Harry McCracken (01:38:27):
You can certainly criticize some of the stuff they did [01:38:30] well, like M the Surface Duo and the product quality was not always great out of the door, but they did fewer things and they were mainly interesting things and it's quite different from what Amazon has done. And it'll be interesting to see, assuming he does start there, whether they hired him to get some of that approach rather than expecting him to do things the way Amazon usually does them, which is to do a million things and see how it all plays out.
Ben Parr (01:38:59):
Agreed with Harry. [01:39:00] Amazon also just needs new blood in that specific division. There are useful, interesting things happening there, but it doesn't feel unified and it feels like at least they have a real shot if they have the resources and they have the distribution, if they have the right person pushing in new vision with some cutting edge stuff, there's a real shot. You should never count Amazon out of anything. This could go well. We'll see. It needs to blood. Yeah.
Leo Laporte (01:39:29):
Yeah. [01:39:30] I mean it can't be any worse. Let's put it that way. Unity, I did not much cover this because I knew that Unity would back down. I just didn't see any way around it. Unity, which makes a game engine that is widely used, there's really two big players in this space. There's the Unreal engine from Epic, which Apple of course doesn't use because they hate Epic. So Unity is widely used on the Apple platform. It's used for games on a lot of platforms. They [01:40:00] incurred a huge amount of heat from developers when they said, we're going to start charging you a fee for every install and it's going to be retroactive and it's going to be 20 cents a user and it's going to cost you a lot of money. At which point developers said, well, in that case we're not going to use Unity. And people started looking for Unity alternatives. Of course, it's pretty tough for a game that's already in Unity. You can't really back out of it. And that was part of the big problem was this is a retroactive fee. This is going to kill us. So Unity has backed down. They apologize. [01:40:30] They say the fee will only apply starting next year with the l t s version of Unity and it'll only be for new games, not existing games. They completely blinked. I think this is probably the inevitable right thing to do.
Any thoughts on that or
Ben Parr (01:40:51):
Don't piss off your developers? Yeah, that's it. Yeah,
Leo Laporte (01:40:53):
That's a good moral, don't screw with your customers. A [01:41:00] couple of big IPOs. I know there were a lot of people who had Instacart who were very, very happy. Instacart went public on Tuesday, then Klaviyo,
Ben Parr (01:41:15):
Leo Laporte (01:41:16):
What is Klaviyo?
Ben Parr (01:41:17):
I know it's weird. It's hard to say the name. Yeah. Klaviyo is a huge piece of email and marketing software that all the e-commerce brands use. It's kind of like, oh,
Leo Laporte (01:41:27):
So if you were in the biz, you would know the [01:41:30] name. Yeah,
Ben Parr (01:41:31):
Yeah. If you're in the biz, you know the name. It's been interesting. These IPOs have started happening after a year of no IPOs and so everyone's been watching these IPOs pretty closely. Are we going to have a bunch of tech companies going for I P O? Is it going to be okay? Kind of the result is the IPOs were all right for companies where they,
Leo Laporte (01:41:53):
By day two, the Instacart went right back down and lost all of their gains,
Ben Parr (01:41:59):
Right? The [01:42:00] problem, so there are two different stories here, right? Instacart had, they were valued what in the 30 ish billion range originally by private investors that went back down on their pricing to 10, which is about where it ended up at, which is not a great win for the later investors who invested late, fine for the early investors who won, but not great for if you raise a lot of money like Stripe or others, they might want to wait more. Klaviyo is unique. They were bootstrapped for most of their time and then they raised [01:42:30] some money. Their last round was about a $9 billion round before they did this. The fact that they went above that number is a little surprising and a really good sign for I think the market for if you have a cash efficient business, because Klaviyo is a profitable business by a little bit, but it's a profitable business and the market rewarded him for that. Now, how it would be long-term, I don't know, but they priced above what people thought they would price on their I P O and then they popped afterwards. So I think there's some kind of trend of if you were overpriced before [01:43:00] you get hurt, if you were really cash efficient at a profitable business, the market's rewarding you.
Leo Laporte (01:43:06):
The company's three founders in Instacart own 17% of the company before the I P O. So they all are multimillionaires, maybe billionaires after the I P O. So we're going to see more tech industry IPOs.
Ben Parr (01:43:22):
One crazy thing, since you're talking about the percentage of ownership, the C E O of Clavio, Andrew owned over 42%, [01:43:30] which is unheard of
For that level. So both the founders are multi-billionaires now. They're going to be more IPOs. They have to, this was at least a decent enough result for more to be like, okay, we should file. I think it'll be a while until you see maybe some bigger ones who have enough money like a Stripe if they file. But there're going to be others where they're like, okay, we're good enough, let's go and do it. Some don't have a choice either. Yeah, the gates are going to slowly open up. More companies are going to [01:44:00] test. We haven't seen a giant pop, but this is not a horrible result as a starting point.
Leo Laporte (01:44:08):
Arms going public right
Ben Parr (01:44:10):
Arm on top two weeks ago.
Leo Laporte (01:44:11):
Yeah, already. And was that a success? Did they do well?
Ben Parr (01:44:18):
Same kind of middling answer. I think actually they had a pop and I think they went down. It's the same kind of middling. It's not a wow. Everything is roses and candy and it's also not, [01:44:30] wow. Everything is burning to the ground. It's
Leo Laporte (01:44:32):
Kind of more normal, right? It's not hype now. It
Ben Parr (01:44:34):
Feels like real market correction has happened, which is good.
Leo Laporte (01:44:38):
Harry McCracken (01:44:38):
I am glad that ARM did go public because the reason it did that was that Nvidia when Alan said was and then was not able to acquire them. And I think that from a competitive standpoint, the world is better off with Nvidia and Arm not being the same company.
Leo Laporte (01:44:55):
Yeah, it's this week in Middling. Somebody said
Ben Parr (01:45:00):
[01:45:00] This week
Leo Laporte (01:45:02):
In middling. That's why we don't cover the markets. I don't. Yeah. Let's take a little break. I want to give a plug to some events coming up very soon in our club. Do you know about our club? If you're not a member, I would love to invite you to join. Club Twit is our way of kind of handling what is happening to podcasts across the board. Advertisers are kind of abandoning podcasting in droves. Not good for us, but we want to keep going. [01:45:30] We like what we're doing and we like the shows and I hope you do too. And that's why we're turning to our audience for the last couple of years. It's just seven bucks a month because you're giving us seven bucks. We don't need to do advertising. So everybody who joins Club Twit gets ad free versions of all of our shows, but we thought we could do more.
We created a fabulous Discord, which has become a social group for five or 6,000 of you. Not everybody who joins the club ends up at Discord, but it's a great place to be to talk to people, smart people, [01:46:00] about any subject under the sun. And we've got a lot of different forums in there in the Discord. And of course you could chat about the existing shows. Plus we're using the Discord for Club Twit events. Lou Mareska, the host of this weekend, enterprise Tech, will be joining us Thursday. Aunt Pruitt, our community manager, will interview him for an ask me anything. 9:00 AM Pacific Time sci-fi author, John Scalzi is coming up October 5th. We're going to have some fun at the end of next month for our Halloween adventure. Micah, Jason and I will do an escape room [01:46:30] 3:00 PM October 26th, we'll all be in costume and we'll be, it's an escape room in a box so you can watch us act stupid.
Stacey's book Club continues in the club. Renee Richie will be joining us. He's the creator liaison at YouTube and of longtime host of Mac Break Weekly. He'll be doing a fireside chat in November. I'll be doing one with Doc Searls and Jeff Jarvis that they call the old Farts chat. I don't know why, but anyway, that'll be interesting. And [01:47:00] I mentioned Anthony Nielsen, our AI guru. He's going to do one early next year. More events coming. You also get a TWI plus feed of shows that we don't put out in public. Like Scott Wilkinson's, home theater geeks, HandsOn McIntosh with Micah Sargent. These are all shows supported by your donation. There's no advertising. So we do 'em in the club first and sometimes they grow big like this week in space and we put 'em out in public. Paul Ott does hands-on windows at Jonathan Bennett's Untitled Linux show the GIZ Fizz with Dick d Barolo, all those in club [01:47:30] twi. That's a lot for seven bucks a month. Now that I say it, it sounds like an awful lot. Join us. We'd love to have you in the club, TWI tv slash club twit if you're not already a member. And I know you're not because if you were, remember, you wouldn't be hearing this. We cut these out too. We don't do those pitches. Hey, it was a great week at twit. This was the week that was watch.
Jim Cutler can sing, baby. That's good. We had a great week. I hope you'll be with us all next week. I will not be here. We're going on a little vacation. Well, if you could call it a vacation. We're going to Green Bay, Wisconsin. My 20 year old son [01:50:00] stepson for some reasons, Lisa's son is a Green Bay Packers fan. So for his 21st birthday, we're going to Lambo Field for Thursday night Football Packers, lions. Now, that'll be a game, a real rivalry. I can't wait. And we're going to go to the Vince Lombardi Steakhouse and we're going to go on a Segway tour and we're going to go on the stadium tour, and we're going to do a real Green Bay trip all next week. So I won't be here. Jason Howell and Micah Sargent will be taking over, and I'll be back next Sunday [01:50:30] for a twit. So I'm gone for a week.
Ben Parr (01:50:34):
Leo, can I just say I'm jealous of that trip and I'm a diehard Bears fan from a long time and I'm jealous it'll be a good game because whatever we're walking right now, I'm watching Kansas City. The scoreboard just absolutely T Swamp my Bears. Oh, I'm sorry. And it's a real painted suffering to be a Bears fan.
Leo Laporte (01:50:57):
I'm sorry, but I would hate to be the team. The Dolphins beat this week [01:51:00] by 70 to 20, so it could have been worse. I think the Kers Lion would be a good game. Broncos. Huh?
Doc Rock (01:51:09):
It was the donkeys. I mean Broncos.
Leo Laporte (01:51:11):
Oh, it was the Donkeys. Okay, good. Those poor Broncos. Oh
Ben Parr (01:51:14):
Cat. Those poor donkeys.
Leo Laporte (01:51:16):
Those poor donkeys. Yeah. It's could be interesting. Lisa and I, retirement Dream, want to go to see all the N F L stadiums slowly over a period of years. Lambo though, you got to start with Lambo. What an amazing
Doc Rock (01:51:30):
[01:51:30] Stadium. I haven't had a chance to go yet, but when we were at N A B, just driving past outside of Allegion, I was like, I need to have a reason to come. This place is beautiful. And that T-Mobile dome they built in the middle of Vegas is gorgeous. Oh my gosh. Maybe when we go to c
Leo Laporte (01:51:49):
S Go for the Super Bowl. Stay for the Super Bowl.
Doc Rock (01:51:52):
Oh yeah. C e s. Yeah. We got to find a reason to go there. Just to check it out. You know
Leo Laporte (01:51:57):
Where I want to go is the dome [01:52:00] that M S G dome. I tried to get tickets to U Two's playing this week. To open that M S G dome, but my credit card was refused. So
Ben Parr (01:52:11):
Ed, go to CS just to see
Leo Laporte (01:52:13):
The dome. Yeah.
Ben Parr (01:52:15):
I have a pitch live recording of TWIT in the M S G do in Vegas.
Doc Rock (01:52:21):
Let's do it.
Leo Laporte (01:52:22):
18,000 seats. It cost I think 2 billion to build this thing and they animated [01:52:30] it. So it's a big L e D screen. In fact, I'm very curious because we are going to the F one race in November, and I understand they're going to broadcast the race. You know how they do the race broadcast on screens all around the race course? They're going to apparently put it on the sphere, which is it cool on the inside too or just on the outside? Yeah, it looks pretty cool on the inside. And by the way, thousands of speakers, so the audio in there is supposed to be pretty spectacular. I'm
Doc Rock (01:53:00):
[01:53:00] So jealous you're going to the F one race. I know. I was looking at, they have a bunch of displays and stuff around town. They had a real car out inside the Venetian and Oh, that's going to be so
Leo Laporte (01:53:12):
Epic. It's going to be so much fun. Yeah, it's a once in a life. It's so expensive. It's a once in a lifetime.
Doc Rock (01:53:18):
Oh yes. I looked up to cheapest room at the Venetian and it was something like, oh, it's
Leo Laporte (01:53:24):
Crazy. Thousands. Yeah, yeah. But we will be staying there over at [01:53:30] the Bellagio, I guess overlooking the strip. And then we have seats on the grandstand there at the Bellagio. So that's it. That broke the bag. That's why I couldn't go see you two. My credit card said, no, no, you already did that. You're not going to go.
Doc Rock (01:53:47):
No. I'll be stalking your polar steps.
Leo Laporte (01:53:50):
We saw Joan Donovan on that promo. She taught misinformation, taught about disinformation at Harvard at the J [01:54:00] F K S school and was fired. There's been pressure across the board, political pressure against disinformation experts. Our friend Alex Stamos, who runs the internet observatory at Stanford. Joan Donovan did find a job at bu. Congratulations to BU for having the chutzpah and the guts to do this Washington Post this week. Misinformation research buckling under G O P legal attacks. It's getting harder and harder for universities to study [01:54:30] disinformation. It started with covid. They didn't like it that the government had gone to Twitter and other social networks. Facebook saying, you got to take down, please do this for us. Take down this Covid misinformation. It's a public health hazard. There was some concern about Russian disinformation during the election. The N I H has frozen $150 million program intended to advance the communication [01:55:00] of medical information citing regulatory and legal threats. Stanford University officials discussing how they can continue tracking election related misinformation through their election integrity partnership because of attacks from political operatives in the US government. It seems to me counterproductive disinformation misinformation is not political. [01:55:30] It is an attempt to trick people and why they would want to shut this down. I don't know,
Ben Parr (01:55:39):
Leo. I think you know the answer and I know the answer. There is a group of people who do want to trick people and it is in their best interest to have misinformation spread and that will only advance with all the new technologies and things coming out. And I sincerely hope [01:56:00] more money goes to this research and more people donate to it and that people are more aware because it may get worse before it gets better and it is a real threat to democracy.
Leo Laporte (01:56:14):
Yeah, I agree with you a hundred percent. It's a real threat to human intelligence period and a threat to science because scientists are scared to do research [01:56:30] because it might have political impact. It feels like McCarthyism, to be honest. It feels like it's the wrong direction for this country. I dunno what to say about it. Be aware of it. Good article in the Washington Post, Stephen Fry's, A little pissed off. Stephen Fry. We love Stephen Fry, the host of Qi Hugh may remember that now in the US the Harry Potter books are narrated by Jim Dale, but if you got him in the uk, and I kind of have to confess, I managed [01:57:00] to score them both Jim Dale's incredible version. But I also wanted to hear Stephen Fry because I love Stephen Fry and he does such a good job. So he narrated all the UK audiobooks of the Harry Potter series.
That's hundreds of hours of Stephen Fry's voice. Apparently people have used AI to learn his voice and narrate a documentary. It sounds just like Stephen Fry did it. He says he played a clip of an AI system [01:57:30] mimicking his voice for a historical documentary. I said not one word of that. Stephen Fry said it was a machine. They used my reading of the seven volumes of the Harry Potter books. And from that dataset, an AI of my voice was created and it made that new narration. He said it could therefore have me read anything from a call to storm parliament to hard porn, all without my knowledge and without my permission. It seems pretty straightforward that that should be illegal. Yeah, [01:58:00] he didn't say he was going to pursue it legally. I don't know what his remedy would be.
Ben Parr (01:58:05):
There's no avenue to pursue it legally right now. There needs to. I agree with Harry. There needs to be, it is a real problem. I already met someone who they imitated their grandkid to make a phone call about being in jail. I can literally right now make an AI voice of anyone as long as I have about 45 seconds of vocals from [01:58:30] them, and that's all that you need. You literally go to 11 labs, you upload and it's done. It's scary. Good how much this technology has advanced, but also scary what that technology can be used for in the hands of bad agents.
Harry McCracken (01:58:45):
If Steven probably wants to take money to let people do that, that's fine. So there are conceivably some acceptable uses, but doing it without that person's permission, now
Leo Laporte (01:58:55):
You own your image. I don't know if you can own your voice. I guess [01:59:00] you can. I mean that would seem to be a part of this persona,
Ben Parr (01:59:05):
But how would you even track some of them down? So the people are impossible to track down or in different countries. It's part of the complex issues related to AI society.
Harry McCracken (01:59:16):
And what do you do if they take Steven Fry's voice and then tweak it by 30%? Make it a
Leo Laporte (01:59:21):
Female Steven Fry. Yeah,
Doc Rock (01:59:22):
Right. Have you ever listened to one of these things and not been able to tell that it's not ai?
Ben Parr (01:59:29):
It's [01:59:30] getting a lot
Harry McCracken (01:59:30):
Doc Rock (01:59:32):
I listened to him. I can tell. I don't know, maybe I'm just crazy, but I can tell. But then that being said, if Frank Cado was my friend, yo, I would be doing all kind. He's
Leo Laporte (01:59:43):
The great impersonator. Yeah, he's
Doc Rock (01:59:45):
Wonderful. You don't, he's mely one of the best. He's so good. So in a way it's always existed, but you can be more prolific with it now because of the ai. But with Frank, it's hard to tell. But for the most of the AI stuff that I've heard so far, I've been able to tell. And [02:00:00] the ones that I've been listening to, which are just hilarious, is like Snoop reading nursery rhymes or they generating songs, whole cloth. And there's a certain thing about a swing in a person's voice, right? Think of it as quantization that the AI can't pull off. And so even when I heard the fake Drake, I knew it was fake Drake because it's missing the mistakes. It was too perfect. And Drake [02:00:30] is not that rhythmic, so it was easy to tell. So I don't know. To me I'd here a better one. I
Ben Parr (02:00:36):
Guess that's just for now though. For now, hundred percent. There are these companies that have raised a lot of money in this technology advance at the same rate as the large language models and the last year in particular to become really, really good. And in limited cases it's really hard to tell for sure. There's some where you can tell just the hair, but the more they're trained, the better they get. And in certain situations where you're panicked, [02:01:00] someone's calling you, there's a jail thing, it sounds mildly plausible, you're not going to have the difference. And look, you have super hearing whatever Superman put into your ears. Most people probably don't have that superpower. I always remind people, AI is at the worst it'll ever be in human history. It'll only get better and scarier and more sounding human. Actually,
What makes it scary, I think, is that people aren't going to care.
Leo Laporte (02:01:29):
Benito [02:01:30] people
Aren't going to care. There's autotune on every song on the radio now fake. Every fake single track is autotune. Everything's fake. Nobody caress. So you think anybody's going to care about this?
Leo Laporte (02:01:38):
Here's one of the YouTube channels that does this so well. They're the ones who did the Barbie Girl by Johnny Cash. Here is Hank the channel's there. I ruined it. Hank Williams singing Straight out of Compton. Straight
Hank Williams? (02:01:51):
Out of Compton. Crazy Mother named I Cube from the game called Fellas with Attitudes. When I'm called off, I got a soda [02:02:00] Squeeze the Trigger and Body are hold off you too.
Leo Laporte (02:02:05):
You can't tell me that. You would know that. That's not Hank Williams. That sounds just like him.
Doc Rock (02:02:12):
It does sound just like I didn't listen to. That's so good. Myself an expert. But that was good. I did hear that one. I forgot. I heard that one. Yeah,
Leo Laporte (02:02:20):
There's some, it's good enough that you don't care if it's ai, right? Well, you might care if you were Hank Williams or I'm talking about the audience. Yeah, the audience [02:02:30] to pay for this stuff. Yeah. Here's Snoop Dogg's reaction to, that's the Jungle Book. Ballou singing Gin and Juice. So
Doc Rock (02:02:54):
That one doesn't sound at all like Snoop, but it
Leo Laporte (02:02:56):
Was No, it was supposed to be, I think Ballou. It actually wasn't [02:03:00] one of his better ones. But Snoop's reaction was hysterical. Yeah,
Doc Rock (02:03:04):
That reaction was hilarious.
Leo Laporte (02:03:05):
Yeah, there I ruined it.
Doc Rock (02:03:06):
What these things.
Leo Laporte (02:03:09):
Highly recommended it. This is a good use of ai. This one I can get behind. John Grisham, the novelist, George r r Martin, the creator of Game of Thrones. Jodi Piol suing Open AI for systematic theft on a mass scale. Joining Sarava Silverman, [02:03:30] Jonathan Franzen, David Baldacci and others. A lot of authors very unhappy about OpenAI being able to synopsize their books or write in their style. I don't think they're going to win this case, to be honest.
Doc Rock (02:03:45):
Leo Laporte (02:03:46):
Ben Parr (02:03:48):
Even if they did win the cats out of the bag, the GDS out of the bottle. Pandora's box has been open. You have an open source, large language model out there. We have a bunch now. [02:04:00] You can't take it back. Someone else is going to add stuff to it. I get why they're doing it. And opening Eyes is taking all the arrows as we've talked about, and everyone else won't take the arrows. Yeah, it probably won't win. Even if it does, it just doesn't stop this.
Leo Laporte (02:04:21):
Doc Rock (02:04:22):
He looks like George, double R.
Leo Laporte (02:04:25):
George double R George. R r. Would you let Elon Musk's robot [02:04:30] perform brain surgery on you to insert a chip in your brain? Because he's looking for volunteers. Neuralink, which got F D a approval is now entering clinical trials. Now in his defense, they're looking for people with spinal cord injuries, quadriplegics, paraplegics and others. I still, I don't care. I still would not let a robot implant this device in my brain. On [02:05:00] the other hand, if you want to go to X or Neuralink and read all about it, Neuralink
Doc Rock (02:05:06):
At all. But if I could get a chip to fix my mom's dementia, yo, I'm paying.
Leo Laporte (02:05:15):
I do it. And I think that's perhaps the point is if you've already lost the use of your limbs, maybe you'd be willing to do something like this if it had even the faint promise that you might be able to get 'em back. If it weren't Elon, maybe [02:05:30] I just, I
Ben Parr (02:05:32):
Don't know. Yeah, Ellan just complicates the equation. But people are, yeah, if you're in that situation, if I was in that situation, probably would try most anything.
Leo Laporte (02:05:42):
Yeah, but you're right. I think at some point we're going to get, there's already implants you can do to correct Parkinson's. And in fact, I have a friend with Parkinson's. It is implanted in his brain and [02:06:00] in his phone is the software that he can put on his head and turn it off. And he turned it off for me. And the shaking, very distinctive, shaking from Parkinson's came back, he put it back and his said, turn it on. And it's imperceptible. So you really can do some amazing things.
Doc Rock (02:06:18):
And a quick shameless plug, the reason why my studio is purple is because of az.org.
Leo Laporte (02:06:23):
Oh, right on.
Doc Rock (02:06:24):
Right on. Alzheimer's is a freaking pain in the butt. And I went through one parent on [02:06:30] both sides of the family, and they're both still alive and kicking. But trust me, the amount of work and stuff that we have to put in as caregivers, the actual caregivers put in as caregivers, it's a very tough thing. And out of every three people, two of them are probably dealing with someone with either dementia or Alzheimer's or something like that.
Leo Laporte (02:06:52):
One of the reasons I went back east a couple of weeks ago is to put my mom in assisted living and the place we chose has memory care because she is early stages [02:07:00] of Alzheimer's every most of the time she's normal. Every once in a while she say something strange like, oh, you should get married and have children. Except that chip sail mom.
Doc Rock (02:07:11):
That's the one. That's the one
Leo Laporte (02:07:13):
That was a while ago. So it's an interesting experience. It's really at the beginning. So I'm with you. I'm with, and I support Alzheimer's. We do the Alzheimer's walk every year and all that on. I think this would be a good time to wrap things up. [02:07:30] What do you say? What do you say, boys?
Doc Rock (02:07:32):
I didn't mean to bring it.
Leo Laporte (02:07:33):
You brought me down. No, no, that's a positive. That's a positive. And I didn't know that's why you were purple. You've been purple forever. A hundred percent. Yeah.
Ben Parr (02:07:42):
I will leave it at, I feel like we have now learned that cyborgs exist and you can tap a brain and you can reduce Parkinson's symptoms and that's freaking incredible. And that kind of thing is going to continue. And that will be great for a lot of people who have had lots [02:08:00] of issues, lots of problems that couldn't be solved. I love science. This is why I love technology
Harry McCracken (02:08:07):
And to bring it around to ai. Again, I think there's lots of reason for optimism that AI will help with some of these intractable problems with human health, particularly relating to what happens to almost all of us if we're lucky enough to live to a ripe old age. That's right.
Leo Laporte (02:08:20):
That's right. I agree a hundred percent. And I think the Alzheimer's walk is coming up. Find out email@example.com. I'm 100% with you on this [02:08:30] one Doc, doc rock plug things. Tell us all about it
Doc Rock (02:08:36):
Right now. I'm just, as you know, working with the Ecec thing, we just dropped a new release and doing a lot on my channel and stuff is just trying to get people to start telling their story. I believe everybody has the story to tell our generation we grew up with speaking until spoken to. Nobody wants to hear from you kind of thing. And that could be further from the truth. So yeah, [02:09:00] for everybody that's out there, if you have something you want to get into or a story you want to tell, please look it. It's all out there and someone wants to hear from you. Believe it or not, look
Leo Laporte (02:09:10):
At you 10 years ago. 10 years ago. Look at that.
Doc Rock (02:09:13):
When I see that, I was like, wow, that was one of my best videos. I love that. The story behind how that video got so popular is actually hilarious. There was an artist who, everybody Is me Doc.
Leo Laporte (02:09:27):
Doc Rock (02:09:29):
Hard drive [02:09:30] on his Mac, ate it. And he was panicking. He has his crew going out trying to buy new computers, whatever. And somebody in his crew saw my video. They replaced a hard drive and they were able to do their show. It was at some big venue in New York and that musician went and told everybody, Hey, I saw this guy. If you have a problem with your mat, go check him out. Oh,
Leo Laporte (02:09:49):
Doc Rock (02:09:50):
Exploded and started.
Leo Laporte (02:09:52):
Wow. 683,000 views later. Look at you young man. That's awesome. That's awesome. [02:10:00] Doc Rock, I love you. It's great to see you. Thank you so much for being here. Thank you for being so kind to my crew out there in Denver and I look forward to visiting you in the Aloha State very soon. It's always a pleasure. Aloha, Mr. Harry McCracken. Sorry you didn't come up here. I know you got busy next time. Next
Harry McCracken (02:10:20):
Time would be
Leo Laporte (02:10:21):
Great. We'll see you, Marie. I know you'd like to come up to Petaluma and experience the thrills of Big City Living,
Harry McCracken (02:10:28):
But it was a nice stay today too.
Leo Laporte (02:10:29):
Good. [02:10:30] Yeah, it was beautiful. Fast Company's the place to catch 'em. Cover story this week. Interview with Satya Nadella. He's a global technology editor now on Masteron Social is Harry McCracken. Anything else you want to plug?
Harry McCracken (02:10:43):
I have a newsletter that comes out on Wednesdays called Plugged In, which you can subscribe firstname.lastname@example.org. And I think I will write more about Microsoft and their news from their event next week.
Leo Laporte (02:10:54):
Oh good. That's awesome. Easy. It's free. Just subscribe. Plugged in [02:11:00] Fast Company's weekly tech report with the Techn. Thank you, Harry. Great to see you. Get my luck.
Harry McCracken (02:11:08):
Leo Laporte (02:11:09):
Always a pleasure. Totally. And of course, Mr. Par, my good friend, Ben Par, the AI analyst. Thank you for indulging us on this AI conversation. It's nice to get an expert in here and get some expert point of view. I always like to challenge people like you. [02:11:30] The place to go, ben par.com, you're at ben par on every possible platform. Anything else you want to plug?
Ben Parr (02:11:37):
I mean, look, I'm going to have some big stories and big announcements in the coming months, so just go to subscribe ben par.com or pick your social network. I will post on literally everything. I got some stuff cooking, and I will leave it at that.
Leo Laporte (02:11:52):
And here is a lovely picture of you and the soon to be Mrs. Par, the engagement. [02:12:00] Beautiful pictures on there, on the blog, ben par.com and subscribe right there. See get more AI content in your inbox. While you're looking at these pictures, just give me your email address. That's all I ask. Hey, it's great to see you, all three of you. Thank you for joining us. Thanks. I wish I could see you, but I'm glad you're listening. Dear friend, thank you for listening to this Week in Tech for this week. We do this show, of course, every Sunday, two to 5:00 PM Pacific. That's five to 8:00 PM [02:12:30] Eastern, that's 2100 utc. You can join us live. If you'd like to get the freshest version of the show, watch us do it live. Live Twit TVs, the live audio and video streams. If you are watching Live Chat Live, our IRC is IRC twit tv.
I hear birds. I must be Doc Rock. Must be Aloha Hawaiian Birds singing. If you are not [02:13:00] a member of Club twit, twit tv slash club twit to get copies of this show twit TV after the fact, you can get them there on the website for all of our shows. You can also subscribe to the YouTube channel this week in Tech. Probably the best thing to do though, get a podcast player and subscribe to the audio or video feed. That way you'll get it automatically is soon. Soon as it is finished and it's finished again, as I've been saying for the last 18 years. We'll see you next time. Another [02:13:30] twit is in the cow. Bye-bye