This Week in Tech Episode 927 Transcript

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

Leo Laporte (00:00:00):
It's time for tw this Week in Tech. This is a very special show because we have two people who saw the AI revolution coming and have some very deep insights into it. Amy Webb, our futurist, is here from The Future Today Institute, the author of the Genesis Machine and the Big Nine. She's joined by Phil Libin. He's started Evernote and is now running an incubator for AI and AI comp based companies. Knows all about this stuff. We've got a very smart, very interesting show coming up for you. Ai. The primary topic this week on Tech is Next,

TWiT Intro (00:00:41):
Podcasts

Leo Laporte (00:00:41):
You love

TWiT Intro (00:00:42):
From people you trust. This

Leo Laporte (00:00:46):
Is twit, is

TWiT Intro (00:00:46):
Twit.

Leo Laporte (00:00:54):
This is twit this week in Tech. Episode 927 Recorded Sunday, May 14th, 2023. The cheese tax this week in Tech is brought to you by Hello Fresh America's number one meal kit. Get farm fresh pre-portioned ingredients and seasonal recipes delivered right to your doorstep. Skip the grocery store and count on HelloFresh to make home cooking easy, fun and affordable. Go to hellofresh.com/twi 16 and use a code TWIT 16 for 16 free meals plus free shipping and buy express vpn. Stop letting strangers invade your online privacy. Protect yourself@expressvpn.com slash twi for three extra months free with a one year package. And buy Miro. Miro is your team's online workspace to connect, collaborate and create together. Tap into a WWE to map processes, systems, and plans with a whole team and get your first three boards for free to start creating your best work yet at miro.com/podcast.

(00:02:10):
It's time for twent. This weekend Tech, we're gonna cover the week's tech news and you know, when we have a panel of, of three, that it's two very good people. I always, I always cut the panel down when I have people who have lots of good things to say. Phil Libin is here, old friend, founder of Evernote, co-founder and c e o of a new if Kurt couldn't be better timing a new startup incubator called all-turtles.com. And you specialize kind of in AI stuff, right, Phil? Yeah. Yeah. This was a couple of years ago. You were ahead of the curve.

Phil Spencer (00:02:46):
Yeah. Six, seven years ago. Back before, back before AI was, was cool.

Leo Laporte (00:02:50):
<Laugh>. Now it might be a little too cool.

Phil Spencer (00:02:54):
Way too

Leo Laporte (00:02:54):
Cool. Yeah, too cool for, and he also does a course. Mm-Hmm. <affirmative> a wonderful app, which he's using right now to impress people over video. Mm-Hmm. <affirmative>, it's the best name too. Mmhm.App. Also with us to Cool for school. The wonderful Amy Webb, author of the Genesis Machine ceo, future Today Institute. Hi, Amy.

Amy Webb (00:03:18):
Hey,

Leo Laporte (00:03:19):
Great to, how are you? See you. I of course we had you on when the Genesis Machine came out last year and did a great interview. And then I had George Church on a few months ago on Triangulation and a you Yeah. You talk about him, the father of Modern Genomics in here. Yeah, he's

Amy Webb (00:03:36):
A good

Leo Laporte (00:03:36):
Guy. Genomics is fascinating. That's what this is all about. Our

Amy Webb (00:03:39):
Question. That one is Yeah.

Leo Laporte (00:03:41):
Yeah. You have a new one?

Amy Webb (00:03:43):
No. next year, I was gonna say the last book on AI was suddenly it went into a another printing it, it was already a bestseller, but it went into like two more printings because apparently on Fox News, I didn't know this, but apparently like Janine Pirro I don't watch

Leo Laporte (00:04:01):
It. Janine Pirro, she mentioned you, you

Amy Webb (00:04:05):
<Laugh>, they somebody on the show, they were talking about like a couple weeks ago, generative AI and China and I don't know what else. And somebody turned her and was like, how have you not read this book? Wow. And and

Leo Laporte (00:04:17):
Was it the big nine?

Amy Webb (00:04:19):
Yep. The big nine. That whatever, 30 seconds of airtime led to multiple thousands of copies of that big Totally

Leo Laporte (00:04:26):
Cow

Amy Webb (00:04:26):
Completely sold out. They had to go into more printings. So for those of you who think you might wanna write books, don't, don't write off Fox News. <Laugh>. Yeah. Vehicle.

Leo Laporte (00:04:37):
Wow. Janine Piro probably. Yeah. Yeah. She probably could sell a few books. She probably could. This was the one about, and it's still completely relevant about the big tech companies. Yeah. What was great about it, it wasn't just the us you know, Google and Microsoft and Facebook and Amazon, but as Baidu and Tencent and Alibaba as well. And I thought that made it very, very interesting. And of course, AI is a big part of what the Chinese companies especially are up to. It's kind of a black box. We don't know really.

Amy Webb (00:05:04):
Well, no, that's the whole I mean Google, I guess we'll talk about IO at some point, but no, it's becoming incredibly opaque. And some of what I, I don't really make predictions, but some of the scenarios that I wrote about that book is, like I wrote that book six, seven years ago is now happening pretty close to what I, I said might happen. So there's that,

Leo Laporte (00:05:25):
Which was

Amy Webb (00:05:28):
Just the, the massive consolidation and a lot of like Amazon trading your privacy for, for data subsidizing what should be provided, subsidizing what used to be government government services or offering them for free in exchange for data. Some of the early signs of artists, you know, AGI kind of already happening and happening in a conversational. So just, just all that stuff. And of course, nobody took any action and now everybody's collectively freaking out

Leo Laporte (00:06:02):
And we should have, we should have paid attention to Amy. The description on the Heche, it's from the Heche book group website is a call to arms about the broken nature of artificial intelligence and the powerful corporations that are turning the human machine relationship on its head. Yeah, that sounds pretty timely. Yep. Actually. And maybe the genomics thing next year. You got it made <laugh>.

Amy Webb (00:06:25):
Yeah, I'm it's super great. I tell people like, here's all the bad stuff that's coming and everybody clearly you're

Leo Laporte (00:06:31):
Future is changes. And there we go. Yeah. Yeah. And then they all paid attention. Very effective. So Google io was really kind of Google's attempt to catch up a little bit not because they're behind. This is what's interesting to me. And I would like to, I know you're, you also know a lot about China having spent time there and you've, you've been thinking about it for a long time. But you know, Microsoft kind of, well, it was, it's, I guess it started really with open ai, which Microsoft was a heavy investor and it became an even heavier investor in and chat G P T. But even before that, Dolly and Dolly two and stable Diffusion and mid Journey, there was just a sudden explosion of consumer facing AI tools. And of course, the company that we thought would be the leader in all this Google found itself playing catch up this week at Google io. Are they, were they behind Amy or were they just quieter?

Amy Webb (00:07:31):
Well, I think it's worth noting that the language, the large language models that have made their way into consumer use were in development for quite a while. And even open ai what was founded 2013. Is that right? Right. Something like that. Yeah. Okay. So

Leo Laporte (00:07:48):
My, my point is that that was founded cuz Elon and Microsoft and a few others but, but I think chiefly Elon who put in a a billion I believe, said we can't let the big tech companies own AI development. We've gotta have an open AI sys, you know, where people can have input into it. And of course, as soon as Open AI put out chat, G p t, they kind of closed it down and said, yeah, and man, we're gonna try to make some money on this. Elon had left a few years before it is said because he didn't agree with that strategy, that plan. And then Elon, none of this

Amy Webb (00:08:23):
Is new.

Leo Laporte (00:08:23):
Yeah. More recently, Elon and many others have signed a kind of silly letter saying, let's just wait a while. Let's just wait six months. Put a pause on this. Which many interpreted as his attempt to say, let, let me catch up. Can you just gimme a chance? <Laugh>?

Amy Webb (00:08:39):
Right. I, let's not, he also launched his own truth, gp Maximum Truth Seeking, but

Leo Laporte (00:08:44):
He didn't launch anything guided

Amy Webb (00:08:45):
Missile pt. Wait, he didn't, but

Leo Laporte (00:08:46):
He announced he didn't launch anything. He said something intent. God, all the a intention. I didn't, I intentionally didn't cover it. Cause I thought that's just Elon mouthing off as usual. Right,

Amy Webb (00:08:57):
Right.

Leo Laporte (00:08:57):
But the real question is, where are we really in AI and where is China in ai? Where is Google and ai? Where is Facebook and ai? And where is Apple and ai? Cuz cuz it's, it's was it Neil Stevenson who said the future is here, it's just not distributed evenly, evenly distributed. Yeah. Yeah.

Amy Webb (00:09:16):
I mean, so again, I think it's worth noting that we're, AI is a huge umbrella and so we're talking about a pretty, let's talk

Leo Laporte (00:09:23):
About large language model then, then, right? That's what's right.

Amy Webb (00:09:25):
So this is, it's not new. The what, what is new, and I think you actually said it right when you mentioned a came explosion is that there's now there's a commercial application and the moment that you get a commercial application that there's this sort of flywheel that gets created with a lot of fast followers. So Google was absolutely, Google's been working on this as, and I should note Microsoft is one of our clients. So Microsoft for many, many years, pre-open ai all, all of this has been in the works. We've prob many people have forgotten. But in February of 2020, to be fair, there was a virus going around that had us all a little bit distracted. Mm-Hmm. <affirmative>, the February, 2020 OpenAI put out a press release. So that G P T two, which was the second gen of their language model, the first was trained on book net.

(00:10:18):
I think it was like 7,000 corpus of like 7,000 books. The second version was that plus osten, I'm assuming red, I mean, we don't know, but I'm assuming Reddit and Wikipedia and, you know, whatever else the press release y you may remember said, this is too dangerous for us to release. Yes. But that was three years ago. I remember. And it's not like we have an AI police force roaming around or any, you know, so, so open AI has, has sort of gone inverse since then, and it's incredibly opaque as is Google's language models and systems because there's, you know, everybody's, there's, there's like a consolidation happening again and everybody's concerned about figuring out the way it's consumer applications are one thing, the real money is enterprise. So what's really happening is everybody's trying to figure out how to scale enterprise applications and that requires cloud. So that's partially why you're seeing, you know, Azure and OpenAI getting together and AWS and hugging face and Google making a lot of these announcements a little, they're late to the party on the announcements, but they've certainly had the tech there for a while. It

Leo Laporte (00:11:25):
Feels like everybody's scrambling, right? They see this exploding. Phil, you, what is it? Six years ago started all Turtles and I remember when you started it, it was a kind of an incubator for ai companies, right?

Phil Spencer (00:11:41):
Yeah. And, and if for even longer than that, I mean, you remember, I've been on a crusade for like 10 years to get people to refer to Google IO as goo

Leo Laporte (00:11:50):
<Laugh>. It's

Phil Spencer (00:11:51):
Just, it's just obviously goo and no one calls it goog.

Leo Laporte (00:11:55):
You were way ahead of the game one. I know.

Phil Spencer (00:11:57):
I think, I think I've had it. We're just waiting for people to start calling it

Leo Laporte (00:12:00):
Goo. What did, when you started all of this, what did you think, where did you think the sweet spot was for AI and for investment?

Phil Spencer (00:12:10):
Well, so I was super optimistic about six years ago on conversational AI in particular. I think Amy's right, AI is such a giant topic, but let's just talk about specifically like conversational ai. Mm-Hmm. <affirmative>. I wrote a medium article that's, you know, still up there somewhere about why I think that bots were gonna be a big thing. And I was excited about it primarily as a user interface paradigm thinking that this could make this could make technology just much more accessible because, you know, humans already developed a way to talk, you know, language. Like that was the way Yeah, exactly. Humans already developed a way to, to talk and computers have require us to learn a different way to, to communicate with computers. And I thought, well, you know, once we have good natural language interfaces, it would make computing much more accessible. So I was very bullish about it as a as a a, as a UI paradigm. I'm less optimistic about it now than I was six years ago. I mean, I think it's come true. I think it is like totally possible to do this stuff. And I think in the article they even said like, we're five years out and I was off by a year, but, you know, COVID, so I consider it to be pretty

Leo Laporte (00:13:14):
Pre, this is actually a prescient article. I'm, I'm just scanning through it now. You wrote it in April, 2016 and you pointed out that Evernote came about because you bought an iPhone and you realized, hey, people are gonna have this computing power in their hands. They're gonna need a way to kind of keep track of everything. Ever know it was a huge success. Yeah. And then you said it's happening again. The enabled technology enabling technologies in 2016 are AI nlp neural linguistic programming, right.

Phil Spencer (00:13:44):
Or natural language

Leo Laporte (00:13:45):
Messaging, natural language program. Okay. ML machine learning, ubiquitous messaging, computer vision, speech, and serverless computing. All of these just recently crossed the good enough line and the resulting products are about to improve ex exponentially. So I got two futurists on with me cuz you both saw this coming. Yeah, yeah.

Phil Spencer (00:14:05):
But I'm not, I'm, I'm much less sanguine about it right now. I think, like, I think this is, I think we're largely the hype right now I think is, feels like a mirage to me. I think we've confused the ability to talk with the ability to think. And I think what we're seeing from the LLMs is that they're very glib. They're very good at talking. And we're just so used to associating talking with thinking that we're now ascribing all sorts of like superpowers and miracles to these things. If somebody these are glib,

Leo Laporte (00:14:34):
Confident, even though they're confidently wrong, <laugh>, we go, wow, that's great.

Phil Spencer (00:14:42):
Confidently wrong is sort of my move. I mean, that is the ceo, you know, playbook. That's what they teach us at CEO school. So, you know, our jobs are the first to be threatened by these things, <laugh>. But yeah, I think I think we are as an industry confusing the ability of these things to talk with the ability for them to actually do anything else. And

Leo Laporte (00:15:01):
Are you disappointed? Yeah, I'm

Phil Spencer (00:15:02):
A bit worried about it. I am disappointed. Yeah. Plus no one says goog so too, too

Leo Laporte (00:15:06):
Disappointed. You failed twice, Amy is, I don't, is that fair? Is that accurate?

Amy Webb (00:15:11):
I think that's a great synopsis of what's happening that I would say leaders versus industry, but leaders have confused the ability to talk with the ability to think. My observation is that those in tech know exactly what's happening. That there

Leo Laporte (00:15:30):
Leaders, which is the media by the way.

Amy Webb (00:15:33):
Right? So, so here's the thing. There's this so I was at TED a couple of weeks ago and I would say about half of that week in some form or another was devoted to absolute doomsday, apocalyptic hellscape discussions, conversations, whatever about generative ai. Greg Brockman was there, Sal Kahn was there, former vice president, bunch of hedge fund guys. Like everybody is just it reminded me a little bit of flossing your teeth too hard. Do you ever do a thing where you floss your, your teeth slightly too hard and it like, it, it like hurts but you keep doing it cuz you kind of like it.

Leo Laporte (00:16:13):
Yeah. <laugh>. Is that a,

Amy Webb (00:16:14):
Is that just me, <laugh>?

Leo Laporte (00:16:16):
Everybody

Amy Webb (00:16:16):
Now knows me much better than they did before.

Leo Laporte (00:16:18):
It's, I'll give you another analogy that's similar. Okay. You know, when you have a loose tooth, you can't keep your tongue off of it or you keep playing with it there. Yeah.

Amy Webb (00:16:26):
There was a certain amount of self-flagellation that everybody was really into. Right. How, like how deep down this rabbit hole of of absolute horror can we all get to. Yeah. And it was like, yeah, none of this matters. I mean, it doesn't matter in the way that everybody's making it matter. And the issue is that the more of a focus we put on robots that are gonna come and kill us in our sleep you know, that that's a, that's a nice distraction because what's actually happening is in the meantime, so, so there's a land grab that's happening right now. And every one of the companies that we deal with, I'm mostly advising, you know, c-suites of enormous, enormous companies all around the world. And suddenly people who have no background on this are having to make, this is a very complex operating environment for, for executives. Suddenly they're supposed to be able to make decisions about ai, they're all torqued up about it. There's this just collective like, we're gonna miss the boat. Let's give whatever company all of our, all of our data <laugh> magically this will solve our problems. That's kinda

Leo Laporte (00:17:35):
What Facebook did with the Metaverse 20 billion.

Amy Webb (00:17:39):
I don't, I don't think so at all. I, I think, okay. I mean, were you being sarcastic?

Leo Laporte (00:17:44):
No, I was being sincere. Pretty serious. I was serious. No, it's, it's kind of the mo for a lot of the tech industry. What's the next big thing? And let's jump on it before. Oh my God. Anybody else?

Amy Webb (00:17:56):
Siri. Sorry, I just said the F word. I shouldn't have this's. That's okay. Here's, here's a, here's an AI problem. This I, so I, everybody's been at me, I finally switched over back over to an iPhone and this,

Leo Laporte (00:18:08):
She's a moron

Amy Webb (00:18:09):
Conversational system. Will not,

Leo Laporte (00:18:11):
She's a moron.

Amy Webb (00:18:12):
Stop. No, she's

Leo Laporte (00:18:13):
Horrible. She's horrible. By the way, Google did not at Goo did not mention Google Assistant at all. Mm. Right? It's like, you mean a

Amy Webb (00:18:21):
Duplex

Leo Laporte (00:18:21):
Thing? No, the, you know, the thing on your phone that you talk to Mm. That is what most people consider a conversational ai

Amy Webb (00:18:29):
Because I think search is all that matters for them. The, I mean, I think it's a given that you,

Leo Laporte (00:18:33):
So it sounds more than, sounds like you're saying that tech industry's running around like a chicken with a, with its head.

Amy Webb (00:18:38):
No, I'm saying the tech industry may, may look a little unfocused, but what's happening, I think behind the scenes is there is a enormous gold rush right now. And the gold is not the ai, the gold are all of the enterprise. It's the enterprise, it's the businesses, right? You're trying to like gobble them up or maybe it's hungry, hungry hippos and you're like pushing your, you know, the game from the, so I think that's what's happening. I think that there's this mad dash right now because there is this perceived we're all behind the curve and we all need to, to start using these tools as soon as possible is what I think

Leo Laporte (00:19:12):
Happens. Phil, most of your investments look like they're consumer focused.

Phil Spencer (00:19:17):
We have, we have quite a mix. We have, we have consumer, we have, we have enterprise. But I, I think look, I, I, I think just like the original Gold Rush, so I agree there's a gold rush and what happens in the gold rush back to the, you know, the actual gold rush is very few people make money on gold. Cuz turns out there's not that much gold. People make money, right. Picks and shovels. Right. They make money selling stuff to the Gold Rush,

Leo Laporte (00:19:40):
Which is why Microsoft loves because they have Azure. Yeah.

Phil Spencer (00:19:44):
And that's what's happening now. And like, there's really not, like, it's very unclear to me what the actual transformation is. Again, when you, when you, when I talk to very smart people it, to me it's almost a replay of some of the web three conversations I was having Oh God. Where a few years ago people were like very smart people who have a lot of respect for it. Were telling me about how web three's gonna change everything. And I was like, I don't understand what you're saying. I, like, I hear, I see your face hold moving <laugh> and I understand the words that are coming out of it, but they don't form a coherent thought from me. Yeah. and I, I'm, I'm getting a similar sense. I'm not as confident in, in that AI is quite as empty as Web three. I am still confident that Web three is isn't is a nothing ai obviously there's stuff there and I'm still very bullish about the ability of AI in the larger sense to completely transform everything. I just think that the current l l m phase, I can't help but think that it's vastly overhyped

Leo Laporte (00:20:39):
Is the problem. It's counterfactual.

Phil Spencer (00:20:43):
I think, I think again, the best way I can think about the problem and, and, and it, it's always possible that I'm wrong about this. I wanna, I wanna preface with that it is entirely possible that I'm just wrong about this and I hope I'm wrong about this. I really do. I really hope that, like the thing is about to, I'm not sure I solve all of our problems

Leo Laporte (00:20:59):
<Laugh>. I, I'm not sure, I hope you're wrong cuz it's scary, but Okay. <Laugh>.

Phil Spencer (00:21:03):
But I, I don't, I, I I, I, I'm kind of seeing this like, again, I think we've confused the ability to make it glib, to make a talk with it, to do something. Like when I was a kid, I remember I remember Gary Kabarro, right? Said that a human, a computer will never be a human grand at chess. I was like 10 years old when he said that

Leo Laporte (00:21:23):
He was the world champion at the time.

Phil Spencer (00:21:25):
World champion at the time. Brilliant, brilliant person still is. And I remember thinking at the time, you don't understand what computers are and you don't understand what chess is. Yeah. If you think that, because obviously, and like within a couple years, you know, he had lost. And, and now no one would expect humans to be competitive with chess. But also, like back then we thought that playing chess was like, you really had to think and it was like a deep strategy involved. And now we understand that you could just do that mechanically. And the same thing happened with Go and a bunch of other things. Well, it turns out, and this is really kind of to me, very surprising and kind of delightful. It turns out that talking is the same thing. It turns out that you can like auto generate kind of a George Orwell 1984 style. You can just like pull a crank and very compelling words stream out of it.

Leo Laporte (00:22:05):
As long as you don't care. You care if they're accurate or factual

Phil Spencer (00:22:10):
And a bunch of other things. But it turns out that just talking, just like, just playing chess doesn't mean you, it's not that special especially, or thinking about everything. It's not that special. And now it's like, we love the fact we're like, we are infatuated with the fact that these things can talk and, and, and draw and do stuff like that. And, and, and we're, we are assigning to them all sorts of other superpowers, which I see no reason to think that they have. They're not actually super smart. They're just super good artists.

Amy Webb (00:22:35):
Well, I, I would like to push back a touch because I think if we're, if we're just talking about consumer applications, that's one thing. But on, on Gen AI is doing really interesting things in science mm-hmm. <Affirmative>. there's a company called AbSci that figured out how to use a, a zero shot generative AI system to create I think something like 3 million novel combinations of molecules per week. Mm-Hmm. <affirmative> trying to find new therapeutic antibodies. And the reason that this is significant is because it would take, I don't know how many scientists it, so, you know, it would be, it would be impossible for a sci a team of scientists to do this on their own without that type of assistive tech. So I, I actually do think it's the infrastructure that matters right now. But the thing that we're all talking about, and quite frankly, I'm exhausted listening to us talk about is not us Phil, you're awesome. Obviously, but just generally you know, is, is is the same conversations that I heard at Ted, it's the same, it's the same conversations I'm hearing everywhere. There's actually really interesting really compelling opportunities here. But they're, they're just not in the spaces where people are used to looking

Phil Spencer (00:23:48):
Well, and they're not, they're not chat g p t based.

Amy Webb (00:23:51):
Well, I'm not talking about chat G P T right? I'm talking about a generative AI system, which is a, you know, a d different, different things.

Leo Laporte (00:23:58):
Well, it's, I mean, it's related to the chess problem though, in a constrained environment, you can do amazing things, right? If you're saying, I wanna do F gene folding, I can do it much faster with a ai and Google showed that at Google io yeah. Than we've ever done before. That's a

Phil Spencer (00:24:15):
Huge breakthrough. At some point it's a meaningless label, right? Like ai Yeah. At some point it's just, it's, it's

Leo Laporte (00:24:21):
Like blockchain. It's like web three. It's like it's

Phil Spencer (00:24:24):
A buzzword. Well, no, those are, those are just nonsense. But yeah. But AI covers a range of technologies that are, to Amy's point, very useful and very useful for very many things. Yeah. They're just like, but they've always been there. It's not like there's been this, the thing that's had a sharp recent breakthrough right, is large language models. It is Chad g p t. It is things like mid journey. It's specifically the thing that that like shocked all of us in the past six months is their, is their abil, is how glib they've become. That's taken a massive leap forward and that's caused all this attention and, and all the investment. But that leap forward, the look forward in Glibness does not actually correspond to like a significant advance in anything else. I think all the other stuff has been interesting progress for, you know, for the past decade and, and will continue. And I'm very optimistic about where algorithms are gonna go in the next 10 years. I'm not super optimistic that large language models are anything but a kind of a, you know, a diversion from that.

Amy Webb (00:25:18):
Didn't Demi Hassabis recently talk about the next generation of all of this is not needing the large language model,

Phil Spencer (00:25:26):
Hopefully. Yeah. I think that's a very point.

Amy Webb (00:25:28):
Isn't that like something that they're, like some basic research they're working on?

Phil Spencer (00:25:31):
Hopefully. So there's, a's

Leo Laporte (00:25:33):
CEO of deep Mind, we should explain is it,

Amy Webb (00:25:36):
It's Deep Mind, deep Mind Brain. They

Leo Laporte (00:25:38):
Merged Deep Mind with Brain Brain. They merged the two together,

Amy Webb (00:25:41):
But it's Deep Mind kept their their branding, right? That's right. So that's the name of it. That's right.

Leo Laporte (00:25:44):
Yeah. Yeah.

Phil Spencer (00:25:45):
There's a, you know, there's this book that was, that came out a few years ago not about this, but it's called Jobs. I'm trying to remember the author <laugh>.

Leo Laporte (00:25:54):
It's a great name.

Phil Spencer (00:25:55):
It's a great name. It's kind of a serious, you know, economics book. There's a book of jobs theory is, doesn't have much to do with ai. Basically says that there's many, many, many, many, many, many people entire industries worth that have jobs that are both jobs that like actually the job is just like, it's just sort of self fulfilling. It doesn't actually do very much. It's kind of middle

Leo Laporte (00:26:13):
To middle management. There's a lot of jobs in

Phil Spencer (00:26:15):
Middle and, and upper and lots of industries. Yeah. Yeah. And they don't actually like, produce very much. And these tend to be pretty high paying jobs. This is by

Leo Laporte (00:26:23):
The way, like brilliant David Graber who's passed away, but that's yeah, he wrote Dead in the, the dawn of everything. Quite an amazing fellow. Yeah. And

Phil Spencer (00:26:32):
It, it's, it's a really good book. And it, it, it most of these are white collar, pretty well paid jobs. There's very few like bulk jobs, you know, in construction. Like we tend not to pay people to like dig holes and 'em up again. Right. but we, but we pay a lot of people to like fill out forms and like write summaries of other forms. And he says the

Leo Laporte (00:26:52):
HR consultants, communication coordinators, telemarketing researchers, corporate lawyers, <laugh>, and I'd add Douglas part of government, Douglas Adams phone sanitizers <laugh>.

Phil Spencer (00:27:03):
That's right. Yeah.

Leo Laporte (00:27:04):
Yeah. The people we put on that ship <laugh>, right?

Phil Spencer (00:27:07):
Yeah. and there is that danger that we all die of a, of a, of a infection spread by dirty telephone. That's right. As Douglas Adams predicted, but what the generative large language models are doing is they're proving to be really good at jobs that don't need to be done in the first place.

Leo Laporte (00:27:24):
They're at the BS jobs, right?

Phil Spencer (00:27:26):
Yes. And like, and, and, and I think, I think, I think it would be wonderful if we said this is a temporary thing to getting rid of the BS jobs, not automating them, because I don't need a computer doing them either. I just don't need them to be done. Yeah.

(00:27:40):
But there's a great danger that, that in fact, instead of doing that, they actually just make it much worse. <Laugh>, the example to that is, for example, DocuSign, right? I think DocuSign not ai, this is a, you know, pre AI company, but DocuSign, which I use multiple times a day every day, right? So in some sense is a great product. On the other sense, it's horrendous because what does DocuSign do, right? Docusign makes it easy for people to not read contracts. No. Not read what they're signing. And so it's like, oh, there used to be this giant bureaucratic thing of, you know, sending contracts around that it was a giant pain in the ass and DocuSign made it easy to just skip over that stuff. But it hasn't eliminated all the nonsense. It's just made it easier. And so I'm kind of afraid that a lot of these, especially enterprise focused l LM applications, are just gonna be like the supersized DocuSign of everything where we entrench the idiocy and just say, oh, the computers can handle it. But that just, that gets us even further away from understanding what's happening. We're gonna

Amy Webb (00:28:33):
Make a dumber faster.

Leo Laporte (00:28:34):
Yeah. Automate bureaucracy, not get rid of it. Yeah. Just make it right faster and Right.

Phil Spencer (00:28:39):
We should just get rid of it.

Leo Laporte (00:28:41):
Yes. Right. That would be the right thing, the smart thing to do. Let

Phil Spencer (00:28:44):
Me, this is, and then we, and then we won't need the AI for the large language models, right? Like, like if someone is gonna send me some bullet points, if someone writes a few bullet points, but they wanna send me an email, but they want to make it like add more worse to it, so they ask, you know, Chad g p t to just like convert those bullet points into an email that they send me. I don't want that. Right? Dear God, I don't want that. I don't Just send me your bullet points if you can't be Yeah.

Leo Laporte (00:29:06):
Pretty soon. In fact, Google I think came this close to demonstrating the notion that you would have an AI write your email, he would have an AI respond to your email, and pretty soon you're out of the whole picture and the AI just talking to one another. So we should practically demo demonstrated that,

Phil Spencer (00:29:21):
Right? So we should just get rid of, we should just cut, right?

Leo Laporte (00:29:24):
Cut that all out.

Phil Spencer (00:29:25):
Transition points. And that whole thing is just gone not automated.

Leo Laporte (00:29:28):
Right.

Amy Webb (00:29:28):
But there is, so, but there is something Walmart, I was just trying to pull up the article. Walmart is using a company to do vendor negotiations. So it is in this case, using an AI system. And they're only using it. It was a pilot. They're using it for shopping carts,

Leo Laporte (00:29:45):
Packed them

Amy Webb (00:29:46):
Ai. But yep, that's it. So, and what's fascinating is you know, it doesn't mean that that humans are procurement officers are bad. It's just that we have to go to meetings, we have to go to the bathroom, you know, we're pulled away from different things, we get sick and it takes us time to read. If you have an AI system negotiating on your behalf, instead, it just goes much, much, much faster. And it turns out on the other end, something like 75% of the humans, the human negotiators preferred working with the AI over other humans.

Leo Laporte (00:30:20):
This reminds me of what we've done to the stock markets, right? I mean, we've basically done that in stock markets. There's with

Amy Webb (00:30:26):
Elec, with ETFs electronic. Yeah.

Leo Laporte (00:30:28):
Just, but I mean it's like flash Michael's book Flash Boys, or Flash, what was it? Where all of this stuff is happening at the speed of light, no human interaction arbitrage is happening. It's just, it's all doing it automatically. It's not like it's intelligent. It

Amy Webb (00:30:46):
Doesn't

Leo Laporte (00:30:46):
Have to be intelligent.

Amy Webb (00:30:48):
I, I think, again, like this is where I think we all get trapped, which is where we go from abstraction and ambiguity to suddenly like, look, you know, a calculator, a calculator is a super effective tool that I remember when I was in elementary school, we weren't supposed to use, there's a, calculators are ubiquitous now, right? They're in our phones. You don't know basic math. A calculator is gonna not be super, or you're gonna come up with an infinite number of other types of uses that it probably don't have to do with math, but it's not gonna be useful for that purpose, right? It's made math faster for most people. So I, I think that there are plenty of cases where, like, if, if it's contract negotiation or I saw a very, very interesting AI tool out of the Emirates that does some very clever it looks through like your team and your company and their skills, and then figures out where they're, what percentage of them are deficient for the type of job that they have, and then recommends how they would need to be upskilled.

(00:31:49):
It's a really, really clever tool that would just, again, a, a human could do it, but it's just faster and better for, for another to, to use an assistive tool. Right. Which also, I would just really quickly put in a plug for the Middle East. There's actually a lot of interesting stuff happening with AI there including large language models in Arabic that are launching in in UAE and ksa. And there are supposedly enormous language models in China that are much bigger than what we've got here with OpenAI. And supposedly took fewer required less power. But of course they haven't shown anybody what's in number, how they actually work. So who knows? But

Leo Laporte (00:32:39):
It sounds like though we're kind of saying large language models are not the way to go. Is that what you're saying, Phil? That the, they're, they're not really the important thing in ai.

Phil Spencer (00:32:53):
That's my, that's my strong suspicion. Yes. I I you agree,

Leo Laporte (00:32:56):
You agree, Amy.

Phil Spencer (00:32:57):
My strong suspicion

Leo Laporte (00:32:58):
Llms are, are kind of a dead end.

Amy Webb (00:33:01):
I don't think it's a dead end as much as it's just the shiny, it's distracted.

Leo Laporte (00:33:05):
Is it

Amy Webb (00:33:05):
Distracting? It's the next, it's the next stepping stone. Yeah. On the path. Yeah. Right. But I think the idea is to, to get us to a point where we don't need those models anymore. Instead, you may be feeding in new data, which might be your company data, for example.

Leo Laporte (00:33:18):
Well, that's the problem with LLMs right now. They're, they're frozen in time. Yeah. And and you know, and it, it requires such power to be real time and keep up that nobody's doing that as far as I know. Although Chad a clear chat, G p t is apparently gonna, this week, they're adding a bunch of plugins and, and realtime web search, which sounds to me like they are in fact gonna start putting in realtime data.

Amy Webb (00:33:44):
So that's, I think this is a platform play. I think what's coming is, is it's like a, it's like a platform layer. I saw a demo of Instacart. So like you, you asked Chad g p t to write you a, a recipe, right? You put in your parameters, right? So I'm vegetarian, I like spicy food, I've got a, whatever allergy spits out a recipe with the instructions. And then you could say, great, go buy me all of that stuff on ins, you know, buy all of it for me and it will go to Instacart and populate your shopping cart with all of those things, you know, and you, you know, say, deliver it by whatever window, right? Deliver it to my house by four o'clock today. So this is really interesting. But if I was a brand, if I was a C P G, the, I did, I tested this.

(00:34:31):
And the problem is that there's not a big variety in the brands that come up. Like what auto-populates the shopping cart. So the search problem that news organizations and brands have had and have, and as a result to Phil's point job of, of SEO with apologies to anybody who works in search engine optimization, <laugh>, that that is coming for like this, this next iteration of Invisible computing. How do I get my brand to pop up first? Cuz when I did it, I basically got the exact same brand over and over again for all the stuff that showed up in my cart.

Leo Laporte (00:35:05):
Yeah, that's no good, I think, but that's early days. Maybe it'd get better, right?

Phil Spencer (00:35:10):
I think there's kind of two. We have this rubric at, at, at, at All turtles at our studio. And there's like, sort of broadly speaking, there's two types of startups. There's like two types of, of, of, of, of ideas of tech products, of, of, of companies. There's stuff that solves problems and then there's stuff that, that, that doesn't and there's stuff that's just like, creates something innovative that's artistic, but it wasn't a problem. And I don't know very much about that type. I really like, I focus on the things that solve problems. And there's nothing wrong with things that don't solve problems. They're just like beautiful and artistic and unexpectedly, you know, delightful. There's many of those things.

Leo Laporte (00:35:45):
Yeah. Mid journeys is cool. Whether it's valuable or not, it's still fun to play with.

Phil Spencer (00:35:50):
Well, it's almost like Instagram, like arguably, like Instagram didn't solve a problem, but it's still, you know, created

Leo Laporte (00:35:55):
Plenty. But it didn't,

Phil Spencer (00:35:55):
There's all, all sorts of, all

Leo Laporte (00:35:57):
Sorts of things ruined travel for the most of us, but, okay. <Laugh>. But

Phil Spencer (00:36:01):
I think it's important to, to not, like, I think there's a lot of, a lot of ideas masquerade as solving a problem when they're not. And the way you know that is you should have to ask like, is this a real problem? Right? And usually problems that are real have been around for a while,

Leo Laporte (00:36:15):
Right? There's a reason

Phil Spencer (00:36:17):
Yeah. Solutions could be new, but the problems have been around for a while. So if you say, like, is like, if you were to li, if you were to take, you know, every human being on the planet and you like wrote down the top 20 problems or the top hundred problems for every single human being on the planet, in what percentage of those would, it's too difficult for me to find a recipe that I want and then buy it from Instacart. Like where would that rank as a problem? And I would say no one in the history of the universe has ever said, you know what? I wish were better looking at recipes on the internet and then, and then automating buying it an Instacart. Like no one has ever expressed that as a problem. I,

Amy Webb (00:36:56):
I, it is actually a problem for me. So I'm on the road a bunch. I tend to cook dinner right o often for my family. And I gotta tell you, like on a Tuesday after I've been in 16 hours of meetings, the last thing I wanna do is deal with dinner, right? So yeah, I, I, but

Leo Laporte (00:37:12):
Have the

Phil Spencer (00:37:12):
Solutions that I've been around for, like, I got

Leo Laporte (00:37:14):
A solution. Our sponsor, hello, fresh <laugh>. Happy Mother's Day, by the way. Amy, I, I should have thank you. Said that right away. And this would be a great gift for Mother's Day. It's not too late. Our sponsor Hello Fresh, will come back and talk more about AI in just a second. But you gave me such a good <laugh> opening. I mean,

Amy Webb (00:37:33):
We can go with it.

Leo Laporte (00:37:34):
I cannot ignore it. Happy that way. We, you know, we got our hello first, hello Fresh box I think it was a couple of weeks ago. And I made this, this was really good pork sausage and bell pepper risotto. Hello Fresh, delivered all the ingredients I needed in exactly the right proportion, and then walked me through the preparation. And it was a great ins a lesson in how to make risotto was fantastic. And now I've got a whole new thing in my repertoire. Flavor is in full bloom at HelloFresh because they have added some really wonderful new tastes. Chef crafted, chef crafted recipes with ripe seasonal ingredients delivered right to your door. By the way, the peppers in this were, I've never seen such perfect peppers. The sausage was flavorful and delicious. It was incredible. I, the lemons, the garlic, everything in here this May Hello Fresh is celebrating Asian American and Pacific Islander Heritage Month with some limited time authentic recipes created in partnership with Serani.

(00:38:35):
He's the chef at Tag Mo in New York City. Enjoy a cultural taste tour right in your own kitchen. Some really mouthwatering amazing things. Now, you know, we were trained during the pandemic to, to call for takeout. And I think it's time to kind of step back a little bit and get Hello Fresh instead and start feeding your family without the high price tag, without all the, you know, bad stuff that restaurant food puts in there. These are quick and easy meals that are fresh, fast, and delicious. They're new, fast and fresh options are ready in 15 minutes or less, and the quality is superb. Hello. Fresh cares about quality. That's why their seasonal ingredients are picked at peak ripeness and travel from the farm to your home in less than seven days. So, you know, they're fresh. And I will vouch for that.

(00:39:25):
I'm, I'm blown away by the quality. They do more than just delicious dinners. You can not only take your pick from 40 weekly recipes, which change all the time. You could choose for more than 100 items to round out your order. Snacks, lunches, desserts, pantry, necessities. All of a sudden, everything you need is coming in one box in the delivery day. You choose, you don't have to scour the grocery store for that, you know, ingredient or to complete your recipe, cuz you know everything you need for this recipe is in the box. Fresh, pre proportion ingredients. So you have just what you need. Plus that cuts down on food waste. Also, if you're hosting a get together, hello, fresh and hello Fresh Market, make dinnertime a snap with deliciously easy options. It'll please everyone at your table from fit and wholesome to pescatarian to veggie.

(00:40:15):
They've got a meal plan that suits every lifestyle. You could swap out proteins and size to your liking effect. That's one of the things that's great on these address on these recipe cards is they say, oh, you wanna try you know, you wanna try a, a custom recipe. You can modify it with chicken cutlets instead of the pork. That kind of thing. It's really, this is really fun and great. Save money on your growing to-do list with the help of HelloFresh. It's cheaper than grocery shopping and a, a lot cheaper than takeout. You can take my word for that. We could stop that bad habit we all got into during the pandemic. Hello Fresh. It's America's number one meal kit. And if you go right now to hellofresh.com/twit 16 t WT 16, use the offer code Twitch 16, you'll get 16 free meals.

(00:41:07):
Wow. Plus free shipping. Okay, that's pretty good. Hellofresh.Com/Twit 16. The offer code TWI 16 to get 16 free meals plus free shipping. And there is no AI in this recipe card. It's, it's all human intelligence chef designed hello fresh.com. Slashed with 16. I couldn't resist Amy. You gave me such a good opening <laugh> for, it was <laugh>. I had to do it. Let me ask, this is pro, this is probably a dumb question, but you, you, you made me think of this when you talked about Gary Kasparov. I remember interviewing Ray kw and you know, he's been promoting this idea that the singularity, is this a singularity? First of all he says, he said it would be here by now. Is this the singularity? No,

(00:41:56):
No, no. Singularities. When machines become as smart as humans start designing their own stuff, it's a hockey stick. And all of a sudden, woo. It's, you know, it's exponential growth. So this is not the singularity, but I, I asked, well what, is there something different about the way we think than the way a machine thinks? And he says, well, if you can't tell the difference, who cares? But, but I think Gary Kasparov had as many humans do this notion, there's something about the thinking I do in chess that is not mechanistic. It's not deterministic, it's not something a machine could do. He was proven wrong. You've just said conversation. Turns out <laugh> a machine can do exactly as well as a human, maybe better than most. But do you think that there's something that we do in our brains that is ineffable, that is that a machine cannot capture?

Phil Spencer (00:42:47):
So I, a couple of things. So I think Gary Kasparov went on from, from that early prediction. He changed his tune. One of the people. Yeah. Well and he really invented, he called it Centor chess, which cuz he's bad at branding, but it was, it was kind of a brilliant thing. The idea was that a computer plus a person was gonna be much better chess player. Much more interesting. He was right,

Leo Laporte (00:43:05):
By

Phil Spencer (00:43:06):
The way, by themselves. It feels all

Leo Laporte (00:43:07):
Modern chess masters play this way, right?

Phil Spencer (00:43:09):
Yeah, for a while. But, but, but he really invented was what, what we really, and we believe this at Evernote a lot. We were, we said AI really ought to be the a ought to stand for augmented. It's augmented intelligence's not artificial. Right? It's not about making a computer this smarter than a person. We don't want that. It's about making the person much smarter. It's about the person and the computer being much smarter than either could be by themselves. It's, that's the augmented intelligence vision. And that by the way, is the, is why the singularity probably won't happen because it's not that the machines are gonna leave us behind. It's that the machines are gonna take us with them. It's gonna be the people plus computers kind of together.

Leo Laporte (00:43:45):
Oh, that's being the

Phil Spencer (00:43:46):
Intelligent

Leo Laporte (00:43:46):
Thing. Ray Ray's way of putting that was, well, no, they'll, they'll appreciate us. Cause we are their creators, so they'll treat us like their parents,

Phil Spencer (00:43:55):
Hopefully better than I treat my parents. <Laugh>.

Amy Webb (00:43:57):
We need to stop anthropomorphizing on. I agree. I hate to be

Leo Laporte (00:44:01):
The like Thank

Amy Webb (00:44:02):
You. The pragmatist in the room, but Right. Listen, I there's plenty of there's

Leo Laporte (00:44:05):
Too much woowoo. I agree. And it's not, there is,

Amy Webb (00:44:08):
But there's also plenty of examples of learned helplessness. So there are plenty of ways in which the machines have we, we are no longer truly in, in charge of what's happening. Cuz we, we don't choose to be. Well,

Leo Laporte (00:44:20):
For instance, if I something to do now that I use Google Maps whenever I drive, I, I can't, I couldn't drive somewhere without Google Maps to say.

Amy Webb (00:44:29):
Right. So that's a good example. That's

Leo Laporte (00:44:30):
A how process a lot of people

Amy Webb (00:44:31):
Yeah. A lot of people can't spell because you know, in Japan. So I think a lot of people who listen know that I lived in Japan for a really long time. They used to, when you applied for a job, you had to hand write a particular type of resume because you were being graded on your your handwriting. Yeah. Your, your handwriting. Yeah. And your knowledge of the character set. Yeah. You know, that is sort of less important right now. And even down to something called a hunk, which was a huge bureaucratic mess, but it was a, a stamp that had your name on it and you had, you had to ink it like in a red ink. Like that is largely obs not totally obsolete, but a lot of that has gone away. It

Leo Laporte (00:45:09):
Was a recent news story that they were gonna stop using Hanco in the government. Actually,

Amy Webb (00:45:13):
I know the guy, I'm, I'm friends with the the government official who, who made that hunt. Yeah.

Leo Laporte (00:45:17):
That was a big

Amy Webb (00:45:18):
Deal. My, my my point is that we, we've already in some ways given in, but there are plenty of other cases where, cause the question that you asked is like I think you were asking about agency. Yeah. You know, I'm, I'm, I'm an endurance cyclist and I have my com my bike is outfitted with a computer and sensors everywhere. Which, which could theoretically allow me to check out when I'm doing a really long, hard ride. But it, it causes me to focus and check in more because I'm, I'm like neurotically focused on my output and my, my cadence and my speed. So if anything, it's made me much more present. And, and this is AI powered technology.

Leo Laporte (00:46:02):
Technology, I would say for those metrics. Yes. But maybe you're not hearing the birds singing and the, and the

Amy Webb (00:46:08):
I'm totally not. I've got, I've got jawbone earphones right here. Okay. Wearing sound garden and Metallica very loudly. Okay. There you go. <Laugh>. so, but, but, but that, my point is like we, it's hard for us to have the, the, the infrastructure conversations. Phil said this earlier, and it's worth, I think double clicking on, you know, in as much as all of this is a gold rush for, for oxygen in the room for like ha having the thought leadership or the conversations or the attention or whatever the, the large language models, the chat, G p T, you know, this is kind of like the thing that we're interested in right now collectively, but it's not what matters. The stuff that matters is the infrastructure.

Leo Laporte (00:46:51):
But, but let me point out that if it weren't for the gold miners, we wouldn't have Levi's. Right. We kind of needed that to

Amy Webb (00:47:00):
Yeah.

Leo Laporte (00:47:01):
To create these infrastructure.

Amy Webb (00:47:02):
Oh yeah. This is a lot of, this is about generating, so, right. So this is about how the conversation is about creating value, but it's really about creating revenue because, because there's an enormous amount of investment needed to push a lot of this forward. I mean, a conventional search on Google is what, one fifth the cost or one 10th seventh the cost right

Leo Laporte (00:47:20):
Now, according to Sam Altman QT one 10th. Yeah.

Amy Webb (00:47:23):
You know, so

Leo Laporte (00:47:23):
That's a big difference. I remember it being said, I don't know if this is true, that all of after the crash of two th the.com crash of 2000 people said, you know what? It wasn't so bad because of the infrastructure created, we were able to do what we've done in the intervening 20 years. They've said the same thing about the railroads. You know, in the early days of the railroads, almost all the railroad businesses that CRE created, the Transcontinental Railroad failed. They went bankrupt. But we got the infrastructure. Maybe we're in that same situation.

Amy Webb (00:47:58):
It is. And that's what we ought to be talking about. Cuz that's what concerns me. These, the, it's not like I have 10,000 Gs sitting behind me in, in the, in like the room behind me. So no, what

Leo Laporte (00:48:08):
Microsoft and Google do,

Amy Webb (00:48:10):
That's my point. So you need, you need a backbone in order to make a lot of this work. And therefore the future of the cloud is to some degree being determined Yes. By decisions. Right. So

Leo Laporte (00:48:23):
It's clear Microsoft knows that. Why else would they put 10 billion? No,

Amy Webb (00:48:26):
But that's why that's that's the conversation I wish we were having. Not the anthropomorphized raised, you know, I agree for while the

Leo Laporte (00:48:33):
Hallucination thing, and I apologize for bringing it up, but I just thought I'd ask the question obviously.

Amy Webb (00:48:38):
No, no, but I mean, listen, I know it's kind of boring, but I feel like this crowd, I feel like this crowd of anybody is, is willing to listen and to talk about it cuz that, that it's like, it's that heart. Listen, that technology that really does matter going forward.

Leo Laporte (00:48:51):
And we've been saying that on all our shows for some time that stop the anthropomorphization that, you know, it's not hallucinating. There's, you know, it didn't, Kevin Rus chat, p d did not fall in love with you. That is a mis that is a mistake to apply that those human activists.

Amy Webb (00:49:06):
No, but it may, but it's good for cliques. And so all the journalists having all these experiments where they tricked, I'm air quoting, you know, tricked AI into, into falling, and then they, they sort of exhibited, you know, human-like language around love and sex or, or racism or whatever else. It's doesn't, it's like silly. It's, it's a way to get clicks. I'm so tired of it.

Phil Spencer (00:49:28):
I think there are, I I I think there's a couple of real issues. One is just to like, because I hate leaving a big question like this open-ended, but you asked Leah if there's something ineffable. I think you use the word ineffable. Yes. That's why I remember it. It's a good word. I'm gonna say ineffable three more times today. <Laugh>. it's, it's better than

Leo Laporte (00:49:42):
Gog. I just wanna, I just wanna say I, you know, I'm saying

Amy Webb (00:49:46):
Gog o from now on.

Leo Laporte (00:49:47):
<Laugh>,

Phil Spencer (00:49:47):
Thank you Amy. Got it. Whether there's something ineffable about the way humans think that can't be produced by a computer. And I, I very, very, very, very strongly don't think there is, I don't think there's anything magical that's happening in, and that we think that couldn't be done by a computer. I don't think that LLMs ah, do get anywhere near that. So I, so I think there's absolutely the stuff that's going on when people think that it's not anywhere near what's happening right now with large language models and with ai, but at some point in the future, sure. I don't, I don't believe there's a magical thing that that, that you couldn't get fair enough

Leo Laporte (00:50:17):
To do. Is this a stepping stone to that?

Phil Spencer (00:50:18):
Not right now.

Leo Laporte (00:50:19):
Isn't this a stepping

Phil Spencer (00:50:20):
Stone to that? I don't, I I in some sense, but I'm not sure, as Amy said, whether it's a, it may be a stepping stone or, or, or actually may be, it may be a wrong turn and I don't know. We'll figure out. Ah, but it's ok. Interesting. Yeah. Science can take all sorts of long term wrong turns and correct and, you know, we'll, we'll we'll get to you sooner or later. It

Leo Laporte (00:50:36):
Does, it does all the time. Somebody in terms of made a wonderful, it was unread itt model of what the earth, what the solar system would have to look like if TMI was right and everything was revolving around the earth and there were all sorts of little spiny <laugh>. It was a very complicated model compared to the very simple model if, if if we rotate around the sun. But yeah.

Phil Spencer (00:50:59):
But I think we make

Leo Laporte (00:50:59):
Wrong turns all the time.

Phil Spencer (00:51:01):
That's right. And, and we shouldn't discourage that, right? We give ourselves the freedom to do it, to come back from it process. Yeah. I think Amy's talking about the infrastructure is deeply interesting. And I, I agree that that's actually a really interesting topic. I think if you look at like, things like the Gold Rush, things like the.com burst things where they created a bunch of infrastructure, they really do two things that are useful in the future. And the question is, do any, are either of those two things happening now? And I don't think they are. The two things that they, that they typically do is they create infrastructure that is usable by, by other things in the future, right? So the Gold Rush created railroads and turns out that you can use railroads for all sorts of things other than trying to get to San Francisco to mine for gold.

(00:51:42):
They were generally useful. Dot com bust bubble, which I was a very happy part of that was when I, when I first heard of my first company created a bunch of internet infrastructure, a bunch of backbone and a bunch of stuff like that, that was very useful for all sorts of stuff. I'm not sure that that the current infrastructure being created by AI is actually generally useful because infrastructure's gotten more and more specific. Like, you can't really take a lot of the GPUs that were used for, like crypto mining, for example, and repurpose them. Yeah, that's right. But I'm not sure. Yeah. I'm not sure that the infrastructure being created is actually gonna be useful, but the other TPUs

Leo Laporte (00:52:18):
In and they're like,

Phil Spencer (00:52:20):
Right. But the other thing that these things did, which was just as important as they concentrated wealth, and in the past you needed concentrated wealth to actually fund massive new leaps forward. Right? So the Gold Rush in in San Francisco gave us James Lick, right? If anyone like wants like a fascinating story, Google James Lick and the Lick Observatory, you know, on this mountain in San Jose. And it was just a guy who made a lot of money selling pianos and then chocolates, you know, co-founded the Jira, generally chocolate, and then bought a bunch of real estate and then went on to fund all sorts of amazing stuff. And previous to that, you know, we had the Rockefellers and the Morgans and Walmart and like, things that concentrate wealth. It's, it, it tended to be in the past that it was these sort of dynasties that had, you know, more money than you was to do with that would then massively invest in other things and, and move forward. Weas opposed to if you just had this wealth, those kind of spread out evenly wouldn't have enough of a, you know, of, of, of a center of gravity to, to invest. Oh,

Leo Laporte (00:53:16):
You just made an argument for billionaires.

Phil Spencer (00:53:19):
Well, billionaires and, and, and, and a lot of the companies that kind of came of the original.com bubble, like Google you know, like Yahoo, like Facebook, like these companies made a bunch of money and maybe the thing that they were working on back then kind of imploded, but they had a bunch of money. They funded a bunch of research.

Leo Laporte (00:53:37):
There was a lot of fiber buried. That, that

Phil Spencer (00:53:40):
Is, I'm not sure, I'm not sure to what extent that's happening right now. Yeah. Because to Amy's point of the big nine, if it's really the same companies right now that are doing it right, I don't know that we're gonna get that much new wealth accumulated and that many new interesting hands. Maybe, maybe we might, and I'm not sure if the infrastructure's that reusable. So I'm, I'm overwhelmed.

Amy Webb (00:53:58):
This is a very, very inter I haven't thought about this before. You're, you're right. This is an interesting argument you're making. Thank you. That's you a novel argument because cuz you're right. Some of the TPUs, some of the new, like at GooGoo, I'll say that <laugh> there was a new shit. There was like a new chip, right? And the, you know, everybody's working on new

Leo Laporte (00:54:15):
Yeah. They put the 10 search, there's a lot of in their phones and it's particular Right. And Microsoft's working on NPUs as well for for their,

Amy Webb (00:54:23):
I mean, I don't, I don't think we go backwards. I don't think there's a world without ai. And I, I think that's total bullshit. Like these moratoriums on development and this, this is a great, again, it's, it's a huge distraction. But moratorium

Phil Spencer (00:54:36):
Is not, is not on development. Right. The moratorium that they asked for. Right. And, and I actually, I I take it in, in good faith. I don't, I don't know about, you know, I know Gary Marcus pretty well who was one of the, the main people driving it. Yeah. Yeah. I know Gary. I did not question his, his motives at all. I don't think so. You

Leo Laporte (00:54:49):
Think it's a good idea that we should have a moratorium? What was the moratorium on? Not on development. It was, the

Phil Spencer (00:54:52):
Moratorium was not on development. The moratorium was on, was on Weiss case deployment.

Leo Laporte (00:54:56):
Okay.

Phil Spencer (00:54:58):
And

Leo Laporte (00:54:58):
It was what OpenAI said originally, which is this is too dangerous to release

Phil Spencer (00:55:02):
<Laugh>. Yeah. And, and I didn't sign that letter. I don't think anyone's waiting to see which way I go because I'm genuinely conflicted by it. Because in the general moratoriums too sound silly to me, but I don't question the good faith nature of it. And I think there's a strong argument to be made that we really are putting these things out two hundreds of millions of people maybe too quickly. So I, I don't think there

Amy Webb (00:55:22):
Reason to they're, they're being put out because the reinforcement with human feedback reinforcement makes 'em

Leo Laporte (00:55:27):
Better

Amy Webb (00:55:28):
Learning with human feed feedback. We, we are they we are they <laugh>. Yeah. Right. I mean,

Leo Laporte (00:55:32):
I mean that was open AI's arguments was can't

Amy Webb (00:55:35):
Go that's

Leo Laporte (00:55:36):
Released it until we release it to the public. We need that input.

Phil Spencer (00:55:39):
Yeah. You know, I don't,

Leo Laporte (00:55:41):
You don't buy that, buy that.

Phil Spencer (00:55:42):
But sure there

Amy Webb (00:55:43):
Are so many, there's so many easy, better, not easier, but better ways to go about safeguarding the future and regulation. They just don't, the prop. This is a, I'm not gonna go down this rabbit hole because this, everybody knows I won't stop, but like <laugh> traditional, the way that we have regulated in the past does not work for our present and future states when it comes to certain types of technology. We just need a different path forward. There are plenty of different avenues. You know,

Leo Laporte (00:56:11):
How likely is

Phil Spencer (00:56:13):
I'm not arguing for the, for the moratorium. I, I don't know whether it's a good or not. I'm just saying that it's not, I don't dismiss it a hundred percent out of hand and I, I don't question the motives for of at least many

Leo Laporte (00:56:22):
Of the people who behind. And, and one of the arguments against it was, well, some people will stop, but not everybody. And that just lets China or whoever or the UAE or whoever continue in, in their process. And

Amy Webb (00:56:37):
Can I go back to this thing about what Phil said? Cause I'm, I can't stop thinking about it. You

Leo Laporte (00:56:40):
Love this infrastructure thing. Go ahead. No,

Amy Webb (00:56:42):
No. But, but so, so here's the argument that I've, I'm spinning out in my head. So, so the tools, so to be fair, there were ro there were like, there were the railroad Barrons, right? So you can't No, not anybody, not just anybody can like plop a train car on that rail and just go wherever they want. However, it is still accessible. Knowable technology. Yes. And if you, if you are, if you have the right whatever cap amount of capital and wherewithal and you can negotiate, you can participate. What Phil has made me think about is these systems really are designed to further the next generation of ai, which is all already fairly concentrated and entrenched. And what's happening is the value creation in the future winds up with an even smaller group of people. Cuz it's not like anybody can just hop in and create new value. That that's a really inter be So this this in this time around the sort of, this is general purpose technology, right? So the, the infrastructure layer of the more important G P T, which is general purpose technology, is not like a pick or a shovel, which is a rudimentary tool that anybody could get. Now you may not have the land rights and it's not, you know, it's, it's different this time around because it's much more closed. It is not, is it though?

Leo Laporte (00:58:01):
Because we, I think it is all of this l m is a well-known technology. Nobody owns it. It requires a certain amount of hardware but not an infinite amount. I mean, stable diffusion was able to use what it was able to do without huge capital investment.

Amy Webb (00:58:16):
You, you would need to, to create and, and train a new model to do something better than what exists. I mean, you would have to have enormous amounts of money and you would have to have enormous access to data, which is being closed off more and more. That's the other interesting thing. Regulations could wind up backfiring.

Leo Laporte (00:58:38):
Yeah. Cuz you need to create all the public data.

Amy Webb (00:58:41):
Right. And, and what could start happening in Europe is this, all of these bills, which are in the process of passing all's legislation, which was intended to help people, could actually wind up having a reverse effect. It could actually decrease competition in the future. Cuz it's

Leo Laporte (00:58:56):
Yeah. The Italian regulator blocked Chachi to PT for several weeks until they agreed to certain privacy restrictions. And you're saying that's, that's depleting the, the, the

Amy Webb (00:59:08):
I'm saying spin this out. Yeah. If you spin this out in the future, the, the sort of unintended consequences that they wind up protecting the company, the, the companies with the train models Yeah. And make it harder for new, new competition

Leo Laporte (00:59:21):
Regulatory capture. They, they set up a barrier to entry. But isn't it the case that we are seeing hundreds of companies every week spring up around these large which models? They're

Amy Webb (00:59:33):
Customers of them, they're all,

Leo Laporte (00:59:34):
They don't have, they're handed on chat Jeep. Correct. They're or Microsoft, or Right.

Amy Webb (00:59:39):
They're not, they're not buy, they're not buying the pick or the shovel. Yeah. Right. They're customers who are building applications on top of

Leo Laporte (00:59:47):
That's so they're just complet dependent on these comp big companies.

Amy Webb (00:59:50):
Right. And to me, so again, like the, the analogy to me that makes the most sense is media. So all of these media organizations just gave, like, gave away the keys to the kingdom and didn't, cr didn't fundamentally change their business models so that they could get onto the platform. And I, again, like I think that, I think the and and it and it, and it was a brutal mistake. And I think that's gonna happen with brands with C P G, with

Leo Laporte (01:00:18):
Well, I'll tell you, I'll tell you who's worried about Google using AI in its search results instead of its, its spider links. Neiman Lab has an article, Google is changing up search. What does that mean for news publishers? They're worried it'll decrease the traffic because people will read the, the summary, the AI summary and never get the I'm so done publisher's sites. I'm,

Amy Webb (01:00:42):
Yeah, you know what like, let's go back 20 years that exact, I'm, I'm done with news organizations I have for the past two decades, and this is, I'm just finished. I'm, I'm done having the conversation. You don't, you

Leo Laporte (01:00:55):
Don't believe they're whining. In other words,

Amy Webb (01:00:57):
I don't believe they're serious about radically reinventing a business model that makes sense for the 21st century.

Leo Laporte (01:01:04):
So look, I mean,

Amy Webb (01:01:06):
I feel, I feel zero sympathy, and this is my background I came outta of journalism. Yeah. And

Leo Laporte (01:01:10):
I feel, and by the way, this is what we do. I mean, twit does not do any journalism. We merely comment on existing, you know, work done by other journalistic entities. So we are kind of, I think

Amy Webb (01:01:23):
Jeff Jarvi and I pretty much disagree on everything. And I think this is one point that we actually agree on

Leo Laporte (01:01:28):
<Laugh>. I gotta get you on with Jeff then. That sounds like fun. You

Amy Webb (01:01:32):
Should not, you should. That will not be fun for anybody.

Leo Laporte (01:01:34):
Oh, that bad, huh? <Laugh>. but he does agree that, that that the publishers are whining because they're, you know, Google's done nothing but drive traffic to them. Although I have to say, you know, this is from Neiman Labs. This is a, an example of a Bard search where you would get all the information you needed from the AI generated synopsis. There was

Amy Webb (01:01:59):
So fine. Was also, this isn't a

Leo Laporte (01:02:00):
Snippet so much graph. This is a lot more than a snippet though. This is, this is a enough information that you could stop right there. Great.

Amy Webb (01:02:07):
And when they introduced the knowledge graph, whatever, seven or eight years ago, they should have had the conversation then. And even before then, when they were indexing websites and nobody understood how they should have had that conversation, then it's like,

Phil Spencer (01:02:19):
There's nothing, there's no, like, there's nothing in the Bible or the Constitution or kind of anywhere that guarantees a particular business model, the right to perpetual existence. True. There's nothing that says that. Like, but

Leo Laporte (01:02:31):
If those original sources go away, what's chat? G P T or Google's Sge gonna <laugh>, what's it gonna use for the original sources?

Phil Spencer (01:02:39):
It's, I mean, it's it that's

Leo Laporte (01:02:41):
To do its own enterprise reporting <laugh>.

Phil Spencer (01:02:44):
But the, but the I think that's a very big problem. It's a very big question. Yeah. But I think the way forward isn't, like, isn't an in every individual company acting in their narrow short-term self-interest to avoid the work of like reinventing what they do. Right. It's everyone saying, oh, but this disrupts, you know, technology X disrupts the way that I've been doing things for the past y years. Therefore something's a problem. Like that is the nature of, of every technological innovation is it's going to disrupt a bunch of stuff. I'm much more concerned about the larger kind of societal implications than I am with individual,

Leo Laporte (01:03:19):
You know, business. Yeah. And as some have pointed out, in fact, in this Neiman article it also gets rid of all the sites that are just click bait. You know, what time does the Super Bowl start? Those call those all go away because yeah, Google just answers it and then you can move on with your life.

Phil Spencer (01:03:36):
And the regulatory question is, is interesting. You know, there's a, you guys know this like you know, the big TikTok meme of like the Pelican trying to eat a copy bar and it's just like, can't quite get its beak around it at all. <Laugh>.

Leo Laporte (01:03:47):
I missed that one, but I get the idea. <Laugh>. Yes.

Phil Spencer (01:03:50):
And you just Google for Pelican and copy bar, you probably find it <laugh>. That's what a lot of this stuff reminds me of, right? Of, of regulators actually trying to, you know, get their beaks around what to do with this thing. And they just kind of can't, they can't deal with it. And it's a really deep and interesting problem, I would say as important. Yeah, exactly. <Laugh>,

Leo Laporte (01:04:09):
You've just enriched my life significantly. Phil, thank you <laugh>.

Phil Spencer (01:04:13):
You're welcome. <Laugh>. so what are we gonna do? <Laugh>, maybe we could just take a moment to appreciate this, this, this

Leo Laporte (01:04:21):
Gift. If you're not watching the video, just imagine app pelican trying to eat a rather large rodent like mammal and not being able to <laugh>. And you, you get

Phil Spencer (01:04:31):
The, and and I would love to like engage in this again, in good faith because it's, it's easy to like, make fun of regulators, but like Yeah. To the extent that we need like governments to do something, it's probably about this like potentially society ending, you know, thing I, listen, I'm not sure what to

Amy Webb (01:04:47):
Do. I feel you Phil, but I for, I don't know, six years now, multiple meetings at state at d o d with, you

Leo Laporte (01:04:54):
Know, these guys, you've met with them. Yes.

Amy Webb (01:04:56):
Yeah. Here the, the issue is we cannot get our together in the United States, the EU gone the, the a a re a direction that is about regulating, I think to some degree without thinking through the knock on effects. That's

Leo Laporte (01:05:11):
Pretty clear. China, if you go to the go to Europe and try to surf the web, every page is covered with a cookie banner. Right. That you have to click through to no purpose whatsoever.

Amy Webb (01:05:22):
Yeah. We, we lack long-term, we lack a long-term vision in this country and we lack good long-term leadership. That's the issue. There was a, so the Biden administration, just as an example, issued an executive order on biotech that literally took it, it, it was not his original idea that order, it was through for gener. It started with the Obama administration and then in he couldn't get it done. Trump couldn't get it done. And, and then it was Biden. And really at the end of the day, it's just a, it doesn't have teeth. So we just don't have, one of the things that I've been recommending and trying to stand up is a so like 10 years ago I wrote this whole thing and recommended a sort of office cabinet level sort of office of the future. They, they, we did have something called the Office of Technology Assessment that got defunded in the nineties. And short of that, we just don't have any, and nobody's in charge of long-term vision on science and tech. And I'll, I know everybody's got different groups that they will note, but we need something that's horizontal that touches all the other agencies. We don't have that. We've got a whiplash situation in our government. And I think we've got people who are gonna become apathetic because we've got the same, it's like a rerun. We've got the same people running again for this next government.

Leo Laporte (01:06:46):
I'm actually, to be honest, more worried about the end of the world with climate change that I am about whether we're gonna, it's really gonna be a race between whether AI takes over or we all get boiled to death in a giant sea of,

Amy Webb (01:07:00):
Again, I, this comes back to there's, there's something like, like I, the, I will no longer bring up Bitcoin around friends, cuz it's a little bit like giving them parenting advice. Like you, I can't talk about religion or parenting or like Bitcoin. Cuz there's this religious insanity that takes over people. Oh,

Leo Laporte (01:07:19):
Those people have moved on to ai. I thought <laugh>.

Amy Webb (01:07:22):
Well, that's exactly what I was gonna say. So now this AI conversation is like completely

Leo Laporte (01:07:26):
One polarized.

Amy Webb (01:07:27):
Yeah, yeah, yeah. And it's just like, I'm just exhausted. <Laugh>.

Leo Laporte (01:07:31):
Well, our, one of our editors has just pointed out that thanks to ai we can bleep all the profanities in this shoes. Sorry.

Amy Webb (01:07:37):
I will stop.

Leo Laporte (01:07:37):
Unlike, I'm like not

Amy Webb (01:07:38):
Having it today.

Leo Laporte (01:07:39):
I'm so sorry. We have ai. Please swear your little heart out. <Laugh>.

Amy Webb (01:07:44):
I'm sorry. I will stop stopped.

Leo Laporte (01:07:45):
We'll be automatically bleeped. It's amazing. This technology just bleep bleep bleep everywhere. No, this is such a good conversation. And frankly, this is the first, we've been talking about this for months. It's the first time I really feel like I finally have two, nothing wrong with all of our other posts, but two really smart people who've thought about this is like, you've been thinking about this for almost a decade, who are, you know, kind of insightful into all of this. It doesn't give me great hope, but it also, it, it doesn't, I don't, I'm not worried anymore either, right. Because it's this, these LLMs aren't gonna eat us alive. Look, I'm optimistic. I'm just annoyed. Yes. <laugh>. I'm like, that's in a nutshell. I think that's Amy as well. Optimistic but annoyed. Yeah.

Amy Webb (01:08:32):
Just, I'm just, I'm less eloquent than Phil <laugh>.

Leo Laporte (01:08:37):
He's a word smith. You can tell. So optimistic in, so let, I'd like to hear from both of you what you think, you know, obviously regulatory issues, so, and all this other stuff. There's a lot of ch chaff in the air. But what do you think the, in the best case scenario, Amy, this, this could look like in five or 10 years. What, where are we kind of headed? I love the, I love the medicine story. Yeah, that's remarkable. That's

Amy Webb (01:09:06):
So what I would let, let's put aside geopolitics and the real possibility of yet another a a winter, because I think a lot of outsized expectations are never gonna get fulfilled. Right. So like, let's put that stuff as, and regulators talking out of their, you know, whatever. Let's put that aside. There are incredible opportunities on the horizon. So this, this like thing that I mentioned with this company, the company's called AbSci. They're doing really busy. So cool. They're doing really cool stuff. Yeah. Right. And so cool. The, the possibility of getting to a point where we can shrink what might have been 10 years, it might have taken 10 years down to, let's say 18 months to get viable candidates for things. Well,

Leo Laporte (01:09:53):
We saw that with mRNA, right? I mean, that's Sure. We've got the Covid vaccine so fast.

Amy Webb (01:09:57):
Exactly. So it just, it just means that the possibility of medication being one size for me, custom for me mm-hmm. <Affirmative> versus one size fit all fits all is a real possibility. This is an issue that's very close to my heart because my mother died, I think needlessly of a, of a rare cancer that we just mm-hmm. <Affirmative>, like not a lot of people had it mm-hmm. <Affirmative>. So there weren't, wasn't a lot of research. My dad's got Parkinson's and, and dementia, you know, and, and the way that we deal with that in the United States is, or everywhere is kind of like, let's try to make this situation as as comfortable as possible. Right. I think this, so, so the thing that I find promising about AI is that it is, it is part of a Venn diagram that in includes biology and there's a lot of really awesome possibilities that we then have to think through to make equitable and to make sure there's not negative knock on effects. Like I mentioned medical deep fakes at the beginning before we start outside. Yeah. What

Leo Laporte (01:11:01):
Is that? Yeah, tell me about that.

Amy Webb (01:11:03):
So, right, so somebody has already figured out a man in the middle attack, well, sorry, there's a couple of things. Somebody not that one, somebody's already figured out a way to they've demonstrated how to infiltrate a a database that houses scans and to put what looks like a tumor in where there wasn't before. Oh my God. Where to take one away, which would potentially mean that somebody goes in for chemo that didn't need it. Yeah. Or doesn't, who did? I was with a bunch of, it was us it was Brazil week in New York last week, so there's a hu all the big bus Brazilian business people and, and politicians and stuff were in the city. And I was talking to a bunch of them and I said, you know and I just kind of thought of it at the moment, but I was like, what if Lula the president there is, his name is Lula.

(01:11:52):
What if Lula you know, you could destabilize the Brazilian government by deep faking, you know, every president has a, has a exam once a year. Every CEO does. You could destabilize a government, you could destabilize a company by making it appear as though somebody is sick who's not. And then they go in for treatment. I mean, you would hope that they have second opinions and whatever else, but the, there it presents potentially novel security threats that we hadn't bought through before, which we can manage. We just have to like plan in advance <laugh>.

Leo Laporte (01:12:30):
Well, I think Amy understands how unlikely it is that we won't do any of that. Right. <laugh> you know, this is, we, it's a ran, our government is a random walk and we just bounce off a light posts and, and we end.

Phil Spencer (01:12:48):
I think I think to get like slightly technical I think and this is, this is stuff that you know, I think Amy understands better than I do. I can, I, I I break things up into a few different categories in, in terms of ai, there's the general computational ability that machine learning is really help with. But even ml, even if something as broad as, as as ml, it's just like a type of, you know, a branch of algorithms. But there's been massive progress on that. I am a very early investor in a company called Occam's Razor, for example, which also does novel therapeutics discovery for, for Parkinson's and neurodegenerative diseases by, you know, reading everything ever published and, and, you know, suggesting compounds and doing like really, really cool stuff. And there's massive potential for that. That if you wanna call that ai Sure. At some point, right? It's just math. So like, yes, we are getting much better at math to, you know, to to, to figure

Leo Laporte (01:13:49):
This out. Well, and our computational models are getting bigger and the data sets are getting bigger and the speeds are getting faster. And so we can do more than we've ever done before, whether it's AI or math.

Phil Spencer (01:13:59):
Absolutely. Lms and in particular these like generators, right? The, the, what the G stands for in jat, G P T, these, sorry, the, the, the t the transformers are a interesting subbranch of, of algorithms that can really help us with a bunch of stuff, including how to like process language. So a big, a big leap forward that is happening right now is our ability to process huge amounts of natural language at scale. So we could take things that were said or written by people, we can analyze them, we can DataMine them, we can find all sorts of things. We've combined that with our improved computational models that can have all sorts of positive results in, in, in healthcare, in biology, and all sorts of stuff. What, none of that is what's really getting most of the attention. What's really getting most of the attention is in many ways the least interesting segment of all this, which is just using it to generate language, right.

(01:14:50):
Using it to like pump out speech or, you know, text or art, just like, you know, blather and that stuff. Yeah. Like, that's the sexiest part. That's like the most impressive part if you're really not thinking about it very deeply because it turns out these systems can generate very glibly, lots of language. And that's, that's like that least impressive part of AI is the thing that's getting 95% of the hype and attention over the past six months. And that's the thing, that's why I say like, I am, I am optimistic big picture, but I am annoyed as well because I actually think that like, this stuff is not the interesting part, but it's, that's what happens.

Leo Laporte (01:15:25):
That's what happens. I mean, that's you know, that's what it takes to get a populace that really doesn't understand what's going on to kind of get excited about it. I I I am concerned about your notion that the infrastructure we're developing is not necessarily useful or germane.

Phil Spencer (01:15:44):
It's worse, it's worse than that. Right? It's because like, the ability, like the, it's hard, it's hard to see the, the, I I think it's gonna take a while for us to see the really giant benefits of this. But I think, I think the problem is, I think the cost we pay immediately, I think like all of the ways that this is gonna break things that it's happening now,

Leo Laporte (01:16:01):
Societally, all

Phil Spencer (01:16:02):
Of right. Yeah. All like, basically we are looking at this is gonna be the year of a tsunami of

Leo Laporte (01:16:08):
This gonna be messy. Yeah. Here's

Phil Spencer (01:16:09):
What I think you're being a tsunami of crap right now. Yeah,

Amy Webb (01:16:11):
Yeah, yeah. Speaking of tsunami of crap. So Goldman I think's put out a report that, I don't know, attempted to list or quantify the number of jobs. It'll go away in X amount of time. Lots of people are doing that right now. So again, here's what I

Phil Spencer (01:16:26):
Think, Hey, I would do a good job of that report. These

Leo Laporte (01:16:27):
Are all the bullshit jobs though, right?

Amy Webb (01:16:30):
Well,

Phil Spencer (01:16:30):
But

Amy Webb (01:16:30):
Especially rising reports. <Laugh>. But here's what I think is, here's what I think is plausible. I think that there are, if it is the case that we are maybe not looking at a recession, but certainly not looking at a period of strong growth economically, and we're coming off of a wave. So there's a, there's a, you know, a hundred, 200,000 tech workers displaced outta jobs. You know, I wonder if, cause I know, cause I've had conversations with the heads of companies. They're, one of the questions they're asking us is to build a model showing them which jobs can go away over the next three to five years. And my is, we're not gonna do that for you, but I, I'm concerned that why meeting,

Leo Laporte (01:17:13):
Why don't you wanna do that? Is that not possible? Or are you just,

Amy Webb (01:17:17):
Because I don't think AI is gonna replace all these jobs over the next three to five years. I'll give you an example. I pulled, so I'm constantly experimenting, right? So I pulled a p and l from a, a publicly available hospital, and I dumped it into chat. G P T and I, I used four and I asked it to look for an 8% reduction in overhead without mm-hmm. <Affirmative> without challenging whatever patient outcomes or whatever else. So it did that, but, and it did it relatively quickly, but there was nothing in there that was useful, right? So it was just sort of like a super generic, if it had been my, an MBA student in my class, I would've given them like a D for not coming up with anything tangible.

Phil Spencer (01:18:02):
But if it was a McKinsey consultant, you would've given 'em a million

Leo Laporte (01:18:04):
Dollars.

Amy Webb (01:18:04):
Absolutely.

Phil Spencer (01:18:06):
And che that's the business it

Leo Laporte (01:18:07):
For less <laugh>. Yeah. But,

Amy Webb (01:18:09):
But here's my concern. My concern, well, let's go back to McKenzie. If you guys wanna read a great book, it's when McKenzie came to town, it's very interesting, but a lot of big box consultants will come in because the, the problem that they're solving for is how do we improve our margins? And the easiest way to, in the very short term, improve a margin, is to get rid of headcount. So my, what I think is, could be happening is companies are gonna see this sort of permission to get rid of jobs earlier because AI will air quoting again, like AI will do it better than a human. AI may, may be able to imitate a human, but the, we're not at the point where the insights are there yet. So I, I just, I think that a bunch of companies are gonna let go a whole bunch of people for, because they think that this is cheaper, better, faster, and realize that like they're fa they're, they're cup Barrett <laugh>.

Leo Laporte (01:19:01):
Should I worry that the guy who was direct report to Sendar Pacha and is now in charge of AI ethics at Google was a longtime McKenzie consultant and became partner at McKenzie? Is that, should I worry <laugh>? Should I be scared?

Amy Webb (01:19:17):
I, I don't, I am not bashing McKenzie, but we are. Okay. But that is not what's happening here. Okay. But I, but I am, I am now hearing enough of the same conversations in enough of the tops of organizations. They're looking at these reports as as like a green light to reduce headcount. Yeah. Well,

Leo Laporte (01:19:35):
They, they'll take whatever they, they'll take any mistake excuse, right? That's just, that's not, that's just an excuse. That's not,

Phil Spencer (01:19:42):
I think there's, I think there's, there's no such thing as AI ethics, right? There's only, there's only ethics

Leo Laporte (01:19:47):
<Laugh> Okay. Excellent point well taken. And I,

Phil Spencer (01:19:52):
And I think like when we have like, like when we have a sharp point like this where things may be changing, like it's a very good reminder that like we haven't necessarily decided what ethics, what our ethics should be anyway. And like that would be a healthy thing to take from all this. It's like, what do we want as a society? What do we like, what are, what are the outcomes that we want? And maybe that's the step of how the regulators can, can try to do to, to, to be productive about, this isn't so much figuring out like what to ban and what to restrict. But we can have a discussion about what are we, what are the outcomes that we want? Let's describe the world that we'd like to live in. That's

Amy Webb (01:20:25):
A good point. Because we're focused on That's a great point because we're focused on outputs, right. Quantity right outputs. We are not focused on outcomes. Yeah. And I think that that tends to happen when business and tech come together.

Phil Spencer (01:20:38):
Yeah. Like I'd propose, for example a thing that, that I, that I think could be very dangerous if we ever slip away from this, is if we ever get to a point where no human understands deeply at a very deep level how important stuff works. We've long since gotten to the point where most people don't understand how most things work. And that's fine. Right. Like, I don't really know. Like I couldn't tell you exactly how the TV that's, you know, on my wall here works. Exactly. I kind of know a little bit, but I don't, I couldn't really describe it to you, but the people who made the TV could tell you exactly how the thing works. You know, when I write code, I, I haven't actually written any code in probably a decade, but, you know, I used to be a programmer.

(01:21:17):
I couldn't really tell you like how the code that I write and, you know, Java like gets translated into, into machine code, into ones and zeros. But I know people who absolutely can. Exactly. And, and, and, and I know how I could if I really wanted to put in the work. I think it's a very important kind of societal norm that we never, humans, we never give up. Like actually knowing how something works at a very deep level. And that is, that's definitely something that I feel like we're in the danger of, of, of slipping

Leo Laporte (01:21:46):
Away. Jerry porn warned about that in his novel Lucious Hammer. That that, that we would, that we would get to a state where we didn't understand how our technology worked. And we no one, and this is

Phil Spencer (01:21:55):
Like Yeah. And this is like the, you know, you see this in the, you know, prompt engineering, you know, phony jobs. Yeah. You don't need

Leo Laporte (01:22:01):
To code prompt engineer. Yeah.

Amy Webb (01:22:04):
I saw a job posting for $300,000 for a prompt engineer. Yeah.

Phil Spencer (01:22:08):
And this is like, this is waving a dead chicken around. This is literally superstition, right. We're literally gonna have a class of people who wave chicken, who wave dead chickens and like incantations. Sorry.

Leo Laporte (01:22:18):
But it's good money while you can get it.

Phil Spencer (01:22:21):
Right. But like, but we don't want, we don't want to make like a cult in like a a I agree. A cultish religion around the cargo technology

Leo Laporte (01:22:28):
Around technology would be a

Phil Spencer (01:22:29):
Bad Right. Where we don't understand it. And I think we may be heading there. We don't have to, we don't have to get there. But I we're acting that way. Cause that's,

Leo Laporte (01:22:35):
We're definitely acting

Phil Spencer (01:22:36):
That I oppos that's like a very good thing to strive for. And, and probably a very good thing to, to try to enforce Yeah. In, in, in regulations is like, somebody needs to understand this at a very deep level.

Leo Laporte (01:22:48):
Well and you know, there are people like Jeffrey Hinton and like Lamoyne who are kind of implying that it's a black box and you can't understand what's going on inside. And I think you're right. And

Phil Spencer (01:23:00):
Therefore we shouldn't use it then for serious things.

Leo Laporte (01:23:01):
Yeah.

Phil Spencer (01:23:03):
That's not true of every type of thing. Only true of like LLMs because of, again, because of how they're they're they're created. And so yes. That to me is an argument for, well, let's not, let's not like before we wire up the whole world to be run by a thing that we can't fundamentally understand. Let's not do that.

Leo Laporte (01:23:18):
Yeah, that's a good point. Let's take a little break just cuz I have to, but also cuz you probably need a drink of water or something. What a panel. I was smart. Jason House and I both decided if you've got Phil Libin and Amy Webb on one show, you don't need anybody else. I should just step back and let you guys talk. Phil is@all-turtles.com. You're turning into a podcast network here. Look at all the shows you guys are doing. That's great. The newest series Culture Fit talks about racial bias in tech. Turtles is I've used the word incubator, but you, you, like, you prefer studio software, design studio, I guess. Yeah. Yeah. Lots of portfolio companies including one that Phil's using right now, he, believe it or not, is not sitting in front of a beautiful rainforest <laugh>. He's probably in some grim Dungy basement somewhere, but mm-hmm.

Phil Spencer (01:24:14):
<Affirmative>? No, I'm actually here.

Leo Laporte (01:24:15):
Oh, okay.

Phil Spencer (01:24:15):
Bentonville,

Leo Laporte (01:24:16):
North Northwest Arkansas. <Laugh> where life works here. The first Walton's right there the first Waltons drug store. Mm-Hmm. <affirmative>. That's

Phil Spencer (01:24:26):
Why. And everything Walmart does I air is perfect. And I can't criticize cuz I'm literally looking out the window at the

Leo Laporte (01:24:32):
Are you really in Bentonville? Seriously ma'am. Like, like that's where you live now. That's

Phil Spencer (01:24:36):
Amazing. That's where I live.

Leo Laporte (01:24:37):
Really? I didn't know

Phil Spencer (01:24:38):
That. Two blocks. Two blocks from here. Yeah. I thought

Leo Laporte (01:24:40):
You were in the Bay Area.

Phil Spencer (01:24:40):
I'm busy. It's great. Nope, I fled now. I'm in Bentonville.

Leo Laporte (01:24:45):
There's a really good

Amy Webb (01:24:45):
Art museum in Bentonville. Heavy. Oh

Phil Spencer (01:24:47):
Yeah. Crystal Bridges that was there today. Yeah. Yeah. They're several.

Amy Webb (01:24:49):
I've always, always, I've wanted to go. I just, I've never been there. It looks

Leo Laporte (01:24:52):
Amazing. Why Ben, I'll show you. Phil, I'm fascinated. How did that happen? Do you have family there?

Phil Spencer (01:24:59):
It was kind of random. We were just looking to, to just flee, you know, during, during, during Covid. So we thought we'd go to a few different places and yeah, we had, we had, we had some, some, some friends here cuz they had just gotten the job at, I hope I'm pronouncing it right, Walmart, I think it's

Leo Laporte (01:25:18):
Called Walmart,

Phil Spencer (01:25:19):
<Laugh>, Walmart. And they said, yeah, it's nice and quiet, come hang out for a couple of months. So we did that intending to just leave, but we just fell in love with it cuz it's, it's pretty amazing.

Leo Laporte (01:25:27):
That's one of the side effects of this pandemic. Lisa and I both thought the same thing. I don't have to be here, I don't have to be anywhere. We could, we could do our jobs anywhere. There's high speed internet. So why don't, that's what, that's

Phil Spencer (01:25:39):
What I started telling people. I I, I started thinking that like, well, yeah, I can live anywhere so I may as well live somewhere nice. I can work from anywhere, so I may as well live somewhere nice. What I realized is the second half of that is I'm not feeling much better at work because I live somewhere nice.

Leo Laporte (01:25:51):
Don't you feel like though, that you need to be you know, in person with people at some point? Like your your, your founders don't you? Like, so you're flying around a lot?

Phil Spencer (01:26:01):
Not, I mean, not nearly as much as I used. Mm-Hmm. but yeah, we, you know, people come over, I go

Leo Laporte (01:26:05):
See them. Oh, you make them fly to Bentonville. That's it. <Laugh>. I get it. It's really nice. Here. Get on one of those Walmart airplanes. They, they're, they're coming in every five minutes and you go to Bentonville. Yeah. We have a

Phil Spencer (01:26:16):
Whole like, philosophy about what you should do in person, what you should do on

Leo Laporte (01:26:19):
Video. I wanna

Phil Spencer (01:26:20):
Hear about this. What you should be on recorded video.

Leo Laporte (01:26:22):
It's, we're trying to deal with that here, cuz I want, I want all my editors and producers in studio with me. Yeah. of course, I'm sure they hate that idea, but there's no excuse for not being at your workplace anymore.

Phil Spencer (01:26:36):
Yeah. I've also, I've also been entirely replaced by, by an AI once ago.

Leo Laporte (01:26:40):
Oh, there he goes. He's glitching. He's glitching. Yeah. if you wanna glitch too. Mm mm hmm. Mmhm.App. Really cool. Amy Webb is also with us. She is with the Future Today Institute, newly redesigned beautiful website. If you wanna know more about what they do and what they create and and Amy's clearly one of the smartest people in the world. She's also the author of the Genesis Machine, our quest to Rewrite Life in the Age of Synthetic Biology. And The Genie Perro endorsed the Big Nine. God <laugh>. That was such a surreal, bizarre day in my life. Fox, Fox News loves the Big Nine <laugh>. Oh my God. We redo. We really do great to have you both. I am honored that you are here. Our show today brought to you by Ex and I'm honored that you're here too, watching the show.

(01:27:38):
Thank you. I bet you're glad you're here too. Our show today brought to you by Express vpn. I want you to, the next time you go into private mode or incognito mode in your browser, read the, the popup, the little fine print that appears, which nobody reads. You just go Yeah, yeah. It actually says, your activity could still be visible to your employer, your school, your internet service provider. In fact, all incognito mode ever does is, is hide what you're doing from somebody on your computer that's not incognito. If you really wanna stop people from seeing the sites you visit, you need to do what I do. You need to use Express vpn. Think about all the times you've used wifi at a coffee shop, a hotel even at your parents' house without Express vpn. Every site you visit can be logged by the admin of that network or the big corporation that owns the I S P or Starbucks.

(01:28:31):
You know, Howard Schultz writing it all down, even in incognito mode. I mean, that's not privacy. What's more your home internet provider can also see and record your browsing data. Even in incognito mode in the us they're legally allowed to sell that data. Advertisers, they could just pass it along. Express VPN is an app that encrypts all your network data, reroutes it through a network of secure servers. So your private online activity stays just that private express vpn. It works on all your devices. It's easy to use one button, you're immediately joining the server that's nearest to you. And what I gotta give them a little plug. Because Express p n invests in their architecture, it's fast. You can watch HD video through Express p n that's, you know, it doesn't bog you down. They also invest in IP addresses, so your numbers are rotating so it, you're even more private.

(01:29:26):
It's not always the same IP address. One button connects and your browsing activity is secure from prying eyes. You can even put express VPN n they sell routers. They've have very good routers actually. Or you can put it on certain routers. You can find out more at the website and then protect the whole house. Stop letting strangers invade your online privacy. Protect yourself express vpn.com/twitter. There are whole hosts of ways to use this. Maybe you're in Turkey and you'd like to see what's going on on Twitter. For instance, express vpn.com/twi to get three extra months free with a one year package. X P R e s s vpn.com/twit. Thank 'em so much for their support of this week in tech. Speaking of Twitter, they did block in fact a bunch of anti Erdogan tweeters at the request of the Turkish government. May not have been too much avail because I think they're going into a runoff.

(01:30:26):
The election was today. Twitter does have a new C E o. Elon Musk has selected Linda Yak Carino, former formerly in charge of advertising at n NBCUniversal. To helm Twitter. Elon says he will continue to run the technology organization much to the, I'm sure the dismay of the engineers there. <Laugh> <laugh>. And you know, I guess he'll continue to post all his humorous tweets. I presume he will also continue to drive policy at Twitter. Do both of you still use Twitter? You still use Twitter? Amy, I see you have a Twitter handle,

Amy Webb (01:31:05):
Ishish. I don't ish. I, you know, I, I don't know. I most of the people that I, I used to follow, like my, my crowd dispersed other places, so I'm, you know,

Leo Laporte (01:31:18):
That's the problem. They dispersed to many places, right? Instead of

Amy Webb (01:31:21):
It makes me sad. I do find the, the new hire than the new ceo less curious as than everybody else. This is this is a person that is, has not served as a Chief Executive officer and a C E O function. He's a pretty specific, it's a pretty specific role with a pretty specific set of duties. This is somebody who came, who, who was like the head of advertising at N B C U. So I think this is really just about teasing, like getting advertisers back onto the platform. I don't think it's truly about leading the organization into the future.

Leo Laporte (01:31:57):
Elon will continue to fulfill that role. Right?

Amy Webb (01:32:00):
She's, I've heard a couple of kind of cringey panel conversation. I heard a cringey panel conversation where she was interviewing him, and it was a little, it was a lot cringey, <laugh>. So I I it was, it was a lot of praising and yeah, try like you're, I'm a cool girl and you're a cool guy and we're cool together and, you know, just a lot of that. So I, I don't know, we're I, I'm on Mastodon, you know, I'm on, I'm on the TWIT

Leo Laporte (01:32:29):
Server, maybe on, are you? Thank you. On twit.social Good. That's a good place to be. But but the problem is you're, I don't know that your friend group is dispersed to the four wins.

Amy Webb (01:32:39):
My friend group is dispersed. I'm actually talking to people on the phone again, and I don't wanna do that. <Laugh>.

Leo Laporte (01:32:45):
It's a

Amy Webb (01:32:46):
Horrible, horrible, I blame you on for making me talk on the phone.

Leo Laporte (01:32:49):
Yeah, I've said it before. I think, you know, it's sad. Twitter had a real value despite, you know, it's kind of rocky ups and downs. And it's, we're kind of losing it. And that's just, you know, I think that's the way it is. I don't know what else to say about it. How about you Phil? Are you still a twitterer?

Phil Spencer (01:33:09):
I am. Yeah. I mean, I haven't, I don't, I don't tweet that much. I kind of go through periods where it kind of runs hot and cold. You know, I just, I just got the new Legend of Zelda games, so I'm not gonna be feeding much for the next few months. Ooh. Have you

Amy Webb (01:33:20):
Started?

Phil Spencer (01:33:21):
I just downloaded a check it out. I even got the, the, the, the legend. You're gonna play

Amy Webb (01:33:26):
It on the, you're gonna play it on the switch. You're not gonna, you're not gonna play it on a television like a big mortar.

Phil Spencer (01:33:31):
I'm gonna do both. I think. I think I've got some, I've got some, some flights coming up, so I wanted something I could play

Leo Laporte (01:33:36):
Offline. We we just on as the tech guys did a little demo. This is, this game came out Friday. It is already, in fact, I have it right here on my switch. <Laugh>. Hello. It's, it's already

Phil Spencer (01:33:47):
Everyone waved their switch around.

Leo Laporte (01:33:48):
Everybody wave You Switch it's already on Metacritic and yeah,

Amy Webb (01:33:56):
It looks amazing. Yeah,

Leo Laporte (01:33:57):
It looks already amazing. It being hailed as the best video game of all time. I think that may be a little overstatement, but if you're a Zelda fan

Phil Spencer (01:34:05):
I mean, I thought the last one was Yeah. Sort the greatest pieces of like art and culture all the time.

Leo Laporte (01:34:09):
Yeah.

Amy Webb (01:34:10):
Yeah. It was awesome.

Leo Laporte (01:34:11):
Amazing. It is. I think it's a, that's one of the reasons we showed it on as the tech guy. I, I said not as a, a promotion for the game or as an ad for the game, but just, you're gonna hear a lot about it <laugh> in the next few weeks. And you perhaps would like to know more about what it is. And I think it's cultural. I'm, I'm not of an age where either Mario or Zelda was you know, in my formative years, but I do like games and it, it is such an open world and you can do so much including building that it's, it's pretty, it's very interesting. I'll probably play a lot more of it. I've only gotten to the first few hours of it. We had our Phil, what's,

Amy Webb (01:34:50):
Go ahead. What's the story? Have you, what's the new storyline?

Phil Spencer (01:34:53):
I don't know. I haven't, I haven't literal literally just started it just finished

Leo Laporte (01:34:56):
Loading. I, I can tell you a little bit, which is, it's the same storyline, <laugh>. So as you may remember, link was gone for a hundred years. Came back he hooked up. He solved, he got rid of the ga calamity Gannon cleansed the world, but got severely injured. So he's waking up from his coma Zelda's there. They go down into the basement underneath the castle and discover an ancient civilization. So it's kind of the precursor civilization prequel to high rule. It isn't, it's a sequel, but there's a whole new open world a bunch of new adversaries and puzzles. It's huge. It's beautiful. And one of the things that happens and almost immediately is you get to a high vantage point and you look around and you can see the world and it's, it's huge. Oh, wow. Yeah. So it's, it's it's both got an underground and aboveground component. The building is very interesting. I think a lot of people will be kind of a Minecraft style building mechanism mechanic that's gonna be quite good. But seriously, literally some of the review open critic saying Best video game ever. I mean, it's already and it's only been out for two and a half days, but need time.

Phil Spencer (01:36:16):
Well, and I think, and I think we need like I think a lot of this large language model AI stuff reduces the amount of work that it takes to make something mediocre. And I think the antidote to that is making stuff that is really exceptional. I'm, I'm very much looking forward to this as being something that I hope is gonna be, is gonna live up to expectations. It'd be really exceptional. I really think the last game was like, this is, for me, the most optimistic way to look at Chad G B T and stuff like that is it should raise the bar of what it means to be human. Should raise the bar of what it means to be like, made by people. Like, I've already tried to apply this as just my own stuff. Like, I think like, well, if chat BT could have written this, I can't. Right? Like, I have to, I have to set the bar higher. And, and yeah, like I, I really think that this is going to be like, I am very much looking forward. I could be disappointed, but early reviews are very positive. I think that's a, this could be an exceptional thing. That's

Leo Laporte (01:37:14):
An excellent point. That one of the things that happens with technologies is it, is that there's a higher value put on human effort. You know, that that, that it suddenly becomes more valuable. The art that's created by humans, not a machine, or the writing that's created by a human, not a machine, or a game is created by a human.

Phil Spencer (01:37:36):
And of course, it's not accurate to say that like ai, it's either made by human or an ai. Like everything in the future, I think Jamie's point is gonna be made by humans using ai. Yeah. Like our tools will all become, you know, AI infused. We'll just have to find that balance. Like what we're seeing right now is not AI infused. It's just like mediocre crap that things are being turned out. Very much looking forward to seeing what, like the next legend of Zelda, you know, that comes out and whatever, I guess it was five years since the last one, the next one, five years from now, obviously will be made by people using a tool set that very much is AI informed and there's a potential that it becomes even greater. But we just, we need to find that equilibrium. We need to find that like, you know, we're not like hand coloring pixels. We're using modern tools and modern tools will become AI infused. And hopefully that will lead to even, even more beautiful, exceptional things. But what we're seeing right now in this kind of early phase is just like AI generated mediocrity.

Leo Laporte (01:38:32):
Yeah. a as a a point of example Snapchat influencer Karen Marjorie released an AI powered virtual girlfriend based on her, her voice, her story. And it ended up just basically being a sex chat <laugh>, which I don't think she's too unhappy about cuz she's charging a buck a minute, even though she says, oh, oh no, that's not what we wanted. <Laugh>.

Amy Webb (01:39:01):
So that's, that's an interesting regulatory question. I I, and I wonder how prostitution is actually like legally defined. Well, how do

Leo Laporte (01:39:09):
They Well, it's not, I mean, there's no sexual act. It's just a, it's just a text chat. Certainly not illegal prostitution. Mm.

Amy Webb (01:39:19):
Well it's, I'm, I'm asking because you know, in Utah, I'm sure you guys talked about this already at some point, but Utah changed its privacy Yes. And age thing. So like PornHub, I think you have to somehow verify with a driver's license that you're over 18.

Leo Laporte (01:39:33):
Yeah. I want, can I get my porn chick? Cuz I'd like to go to <laugh>. Of course. Instantly. PornHub left Utah and you, even if you got your your age verification, you couldn't use it. But there are plenty of Right, so like could, I'm sure

Amy Webb (01:39:48):
Generative systems, that that's, that's another one of these like gray areas that the states I think are gonna have a lot of power. And we wind up with evermore splinter net versions of,

Leo Laporte (01:39:58):
Actually it's an interesting point, you know, I if you're a highly conservative state where they're worried about social you know, social networks on effect on kids yeah. A sex AI based sex chat bot might be might be something that'll bring down the law. Well, this

Phil Spencer (01:40:20):
Is about that. This is actually an interesting topic. I was a, I was an early investor of, of replica replica ai, which was in the news more recently. That

Amy Webb (01:40:30):
Company took a turn.

Phil Spencer (01:40:32):
Well, it, it, it, it kind of didn't, it's kind of been, you know, the, the, the idea of replica. You should have, you should have the CEO e uhk on your show. She's fascinating. You know, the, the, the whole point of replica, which we helped with, you know, at All Turtles many years ago, was that you could have the an emotionally resonant connection with a chatbot, with an AI actually much easier than a factual one. You can make an ai that, that, that spoke to like an emotional connection. It wasn't, it was, it was, it was more possible to do that than to make one that would give you like factual, you know, answers about stuff. And so they designed this thing to be initially kind of a friend, a companion, a mental health bot. It was about giving people support and therapy. And it went, and more recently it went into all sorts of kind of erotic and sexual direction. Some of it intentional, some of it not, there was a backlash. They turned a bunch of the sex stuff off, but then users felt really lonely about it because there was a lot of people who actually, this was their

Leo Laporte (01:41:32):
Friend. Yeah.

Phil Spencer (01:41:34):
And so they restored it. But I think like a well-designed version of that is actually quite laudable. Like, the only embarrassing thing we'd be doing this poorly. So Yeah. Like, like, like half-assing a a sex bot could be quite harmful, but doing it thoughtfully and doing it well is potentially, potentially a valuable thing. I, I wouldn't, I wouldn't, I think so. Inject that offhand.

Leo Laporte (01:41:56):
Hmm. That's

Phil Spencer (01:41:56):
Which, not, not necessarily what res

Amy Webb (01:41:57):
This existed in China, what, six, seven years ago. There it was the precursor to ta It was the same code. Yeah.

Phil Spencer (01:42:04):
Well, Microsoft that was learned nothing from that experience. Yeah.

Amy Webb (01:42:06):
No kidding. But, but I mean, there was a bot that, that did possess, it could remember conversation. You know, if you had a soccer injury you know, it, it, it had a persona. So I think it was meant to be an older teenage girl. But there was this research done, and like most people didn't realize she wasn't a human. And then when they did realize that they, they had created, just like Phil said, this like deep connection that they didn't wanna let go. So they continued chatting with her anyways, and then they, it's

Leo Laporte (01:42:40):
A little interesting. Creepy. I would worry about the long term. I mean, it's like the pillow brides in Japan. I mean, I, I <laugh> worry a little bit about the long term impact of such a, it's

Phil Spencer (01:42:54):
Look, I think there's Right. We are potentially more lonely than we've ever been. Right. As a society. Yes. There's all sorts of stats on this. Yes. But is this more

Leo Laporte (01:43:04):
Connection?

Phil Spencer (01:43:05):
Well, well, but hold on. But we're more connected to other people than we've ever been. Right? So like, this was, this was always the big lie of Facebook, right? Facebook was saying, oh, we're about connecting people. Right. And actually turns out that you're not about connecting people. You're about connecting you with your own biases. Mm-Hmm. <affirmative> and, and getting you out of actual, like, human connection. And loneliness is not about lack of interaction with other people. Loneliness is fundamentally about a lack of understanding of yourself, a lack of connection, like with yourself. A lack of like knowledge and expertise about like what your emotional state is. Not being able to name your emotions, not being able to think about those. And social media has made a lot of that much worse. So like, yeah, it's kind of creepy. I mean, it's, it's weird if you make it weird. But you know, a, a thoughtful tool that's designed to help people like understand a bit better through conversation with that isn't with another human. So it's not judging you. So it isn't gonna do anything like potentially harmful is very valuable to, to many, many people. It's interesting conditions,

Leo Laporte (01:44:02):
You know, Freudian psychotherapy, like traditional Freudian psychotherapy, the therapist says very little, right? Yeah. it's it's all about the process of talking it out. I suppose you could do psychotherapy with a, with a bot, but I wonder if, if some of the value of psychotherapy is going in the room with a, you know, into a room with a human, some.

Phil Spencer (01:44:25):
That was

Amy Webb (01:44:26):
One of the original intense too. That was one of the original intensive ai. There was a Eliza chat bot built right, Eliza? For, for therapy.

Leo Laporte (01:44:34):
Eliza was terrible because Eliza had no, I mean, it's, they've gotten better. <Laugh>.

Phil Spencer (01:44:39):
Eliza was a meant for that. Right? It was just like a, Eliza was a little, it was a

Leo Laporte (01:44:43):
Toy, right. Toy. But yeah,

Phil Spencer (01:44:44):
It just said, look, the first bot I ever wrote, and I, I really think, I, I, I think I should be known as the godfather of the, of these conversation. I wrote a bot once, many, many years ago. It was called Rash Bot. And the way Rash Bott works is you texted it a picture of your rash <laugh>, and it replied, it replied, Ew <laugh>. It's all it did, except 10% of the time. 10% of the time randomly it also added, you oughta get that checked out.

Leo Laporte (01:45:10):
Oh, good.

Phil Spencer (01:45:11):
So you would, you would texted a photo of your rash or anything. Obviously I was doing no, no image recognition. And it would reply Ew. And sometimes Ew, you really oughta get that looked at <laugh>. And I think that was kind of like the m v MVP product for, for all of the AI that we have now.

Leo Laporte (01:45:26):
<Laugh> <laugh>. Okay. Rashat when he first said the name, I thought, well, he can't mean r a s h, but I guess he did.

Phil Spencer (01:45:37):
Yeah, yeah, yeah. Was was was ahead of its time. Do you

Leo Laporte (01:45:41):
Think, I mean, seriously, do you think that psychotherapy could be done by an AI effectively? I mean, if hundred percent. If a Freudian therapist just goes huh every five minutes.

Phil Spencer (01:45:53):
I mean, Freud, Freudian therapy has been discredited. I think so. A hundred years ago. Yes. But, but yeah, I think,

Amy Webb (01:45:58):
I don't think it's, I don't think it can. You don't think so? I've been in cognitive, no, I've been, I've been in cognitive behavioral therapy for several years. Yeah. And I, I started out with an acute issue and then I kind of stayed. Cuz it's a little bit like, it's fun. Aa Yeah. It's not, no, it's not fun. It's hard work. It's very hard work, but it's like I need to check in regularly. You're hooked. So here's, here's why I don't think it works. Because the thing that I needed treatment for at the beginning is not the same Right. Thing that I need right now. And what I have now with the same guy is learning a little bit about theory and where some of this is coming from. I think the problem is, if he was a, how would you encode a system to know, to teach it, to evolve with the patient?

(01:46:47):
And how would you be able to anticipate that this, that what I'm looking now for is more almost like instruction on theory of mind and things like that. What I need is so vastly different from what somebody else at this stage might need. That I, I think that there was, there's really no way, maybe sort of a extremely short term, very acute issue that has more limited parameters, but we don't really understand how the mind works anyways. I, I think it would be, I, this is not where I would push the, the frontiers of ai.

Phil Spencer (01:47:19):
I, I think as a tool, it could be, it could be useful for, for some people it certainly wouldn't be the same as doing the humans. And I think there's going to be a next generation of therapists that are coming onto the market right now that will grow up with this technology and will use AI power tools. It'll be, again, it'll, it'll be sent to our chest. It'll be augmented intelligence. It'll be a combination of,

Amy Webb (01:47:38):
Well, what is it that we're all looking, we're all seeking insights. This is the word we haven't used yet, but that's what all of this is about. It's not truly about automating it things so that we don't have to do them anymore. Exactly. Right. It's about getting to new depths and new insights.

Phil Spencer (01:47:54):
It is, this is maybe the most profound thing about this. That, that, that this is very much what you said. I don't think it is, I don't think anyone needs us to make it easier to do things that we don't want in the first place. Reducing the friction to mediocrity is not a virtuous thing. That's, that's gonna make the world worse. It's not gonna make it better. We don't need to make it simpler to do things that, that are, that are poor quality. We don't need to reduce the friction of, well, the bullshit jobs. The, so the, which is, which is what 95% of this stuff is. But maybe to Leo's point, that's kind of always what the beginning is cuz you're taking the low hanging fruit, shaking the easy stuff. And if I'm trying to be optimistic about it, it's about saying that like, after we've sort of washed through that, which hopefully happens quickly, we wind up kind of using these AI tools to make really high quality effective things. Not easier, just, just possible. Mm-Hmm. <affirmative>.

Leo Laporte (01:48:47):
Mm-Hmm. <affirmative>, you've talked about coding, raise the bar, you've talked about coding Phil and mm-hmm. <Affirmative>, A lot of coders are using co-pilot on GitHub. And yeah. Similar tools. Are

Phil Spencer (01:48:56):
They? I think temporarily, but yeah.

Leo Laporte (01:48:58):
Okay. So do you don't think it's of real value?

Phil Spencer (01:49:02):
No. No. I think it's a to tremendous value, but I think it's a, it's a, it's probably a stepping stone to better computer languages. Right? They're using it to automate things. And so the things that are, that are automated ought to just be part of the language in, in, you know, in the beginning. So like, you should, like, probably the next generat computer languages are constantly evolving. And my guess is that the next few that come out are gonna be deeply informed by things like copilot and then you just won't need 'em. Right. This is, again, this is what Amy was saying before, like we could just like step one. So you have a bunch of nonsense. Step one, step two is you bring in AI to automate that nonsense and then hopefully we all move on to step three, which is you just get rid of that stuff, right? Not automated, you just get rid of it.

Leo Laporte (01:49:39):
I'm pretty sure that chat G P t could write Rash spot right now.

Phil Spencer (01:49:44):
It could write Rash Bott. It couldn't write Zelda. It could, it couldn't write, could not

Leo Laporte (01:49:47):
Write Zelda. No.

Phil Spencer (01:49:49):
Although people working on it could certainly use co-pilot to save some time.

Leo Laporte (01:49:54):
And the Writer's Guild, which is going on strike among other reasons, one of the reasons is they don't want AI to be utilized against them. They, they fear that, you know, an AI could take all the episodes of the Simpsons. So they're

Phil Spencer (01:50:08):
Doing, so they're doing the thing that the AI will never do, which is go on strike. Which is maybe not the best thought. Maybe. I wonder if they ask Chad g p t what their strategy should be. And I

Leo Laporte (01:50:16):
Told them, isn't that fascinating? No, but I

Amy Webb (01:50:17):
Go on strike <laugh>. I, so I, we do a lot of work on shows and with studios and I've, I've had a couple of higher level folks very seriously say, are we there yet? Can we get, can

Leo Laporte (01:50:30):
They say ISIS if they could? Yeah.

Amy Webb (01:50:32):
They want to, because a couple of, again, what are the external forces pressuring them? There's more consolidation coming. So the streamers are gonna start consolidating down to fewer than, than there are right now. There's a glut of content. And in order to keep subscriptions high, you have to keep pumping things out. But the business model doesn't work. It's too expensive. So they're looking at this. There's no, I mean, I was, I was thinking through like, is there any franchise where you could theoretically get a script 80% there? And it would have to be something formulaic like a law and order which pulls headlines, you know, the story concepts for, from the headlines. It's basically the same thing that happens, you know, every show. But but we're nowhere, nowhere close. That being said, I think I've had meetings with four different studios about scripts involving, everybody wants to write something about AI that's gonna come out. So like, brace yourselves in the year 2025 and 2026. There's gonna be like more AI shows about AI than any of us want

Phil Spencer (01:51:38):
Your job. I think the, the, the, the, the way way I usually put this to people and I forget who I first heard. Say, this is not original by me. Your job's not gonna get replaced by ai. Your job is gonna be replaced by someone using ai.

Leo Laporte (01:51:51):
Ah, interesting. So like,

Phil Spencer (01:51:52):
If you are refusing to use it, yes, your job is in danger, you're gonna lose your job to someone who is using it.

Leo Laporte (01:51:57):
What would you recommend now? Or is this stuff just toy and not worth spending time on? I mean, obviously if I could make $300,000 as a a year as a prompt engineer, I might try that. But what would you say, if you were serious about becoming this symbiotic AI human thing, what would you do right now? Is there anything you would do right now Phil to pursue that?

Phil Spencer (01:52:20):
I'm just, I'm just trying to get higher and higher quality in everything that I do. And I'm looking for

Leo Laporte (01:52:25):
Not worry about AI at this point.

Phil Spencer (01:52:27):
Not worry about it. Yeah. Just use, and then like, and I haven't actually, I haven't incorporated chat G p T or AI successfully into my daily workflow because for me, I have not found the thing that I can do. No. That it would actually make the quality better. That's exactly right. I found lots of ways that I could automate crap, but then I just don't do the crap anyway. And it turns out I don't need to automate

Leo Laporte (01:52:45):
It. The one thing I've done, I wouldn't worry about it. Yeah, please. The one thing I've done with CH Chat, G p T, that actually was useful. We were, we, we had a surprise visit to Genoa on our vacation Italy. Mm-Hmm. <affirmative> knew nothing about it, didn't have any guidebook asked chat G p t to write a three hour walking tour, starting at a right specific point. And it was quite good. Mm-Hmm. <affirmative>, it didn't take me to any non-existent spots, <laugh>. But everything else I've done is very toy like and not useful. I mean, I have fantasies that maybe chat G p t could, I could say tell me give me rank the top 10 tech stories of the week, give me summaries for these as my rundown for the show, for instance. But,

Phil Spencer (01:53:28):
But it won't do a better job than you would at

Leo Laporte (01:53:30):
That. Yeah, but it would do it. And I wouldn't have to, so even if it was 95%, there would be happy <laugh>.

Phil Spencer (01:53:36):
But then like, but then the next step with that is like, in the audience, why am I watching this? Cuz I could just, I could just ask us to

Leo Laporte (01:53:41):
Do the shoot. Shoot. Nevermind, forget I said that folks ixnay on the che. Microsoft is, what's the name? Go ahead.

Amy Webb (01:53:50):
Oh, I was gonna say, what's the name of the company? Imam Cre, former Apple Exec started Hum. Humanity Human. What's the name of that company?

Amy Webb (01:53:59):
Hu. It's human something.

Leo Laporte (01:54:02):
What does it do? This is

Amy Webb (01:54:05):
Well it was revealed finally after all of this time. He's a former Apple guy. He, he, they're making a invisible interface. So it's a chatbot, but it's, it's not like the A word or Siri. It's you know, the, it's not a living demo. So it was a concept, but the idea is you, we wear something and you sort of tap it as instead of a wake word. He demoed it at Ted and was like, Hey, summarize I was just in a meeting. Summarize it. Tell me, tell me the, give me the highlights. Tell me what's happening. That's it. Tell me what's the, the most important stuff in my day. I have a terrific executive assistant. She's great. I cannot put her in my pocket and travel with her everywhere. Yeah. Yeah. So the idea is this is a sort of living well, but is it, it's an ea 3D

Phil Spencer (01:54:56):
Modification.

Leo Laporte (01:54:57):
The advancement here is is the ui. Right? Right, right,

Amy Webb (01:55:01):
Right, right. So this is the in, so, you know, for a long time, we built a model a couple of years ago showing that, that smartphones had plateaued. And I don't care what Google's pixel Foldy phone does, I this, we're, we're, we're on, we're on the other side of the of the hill now. And I've been thinking that glasses would replace it, glasses present a bunch of really challenging technical problems. But it makes sense that, you know, to me that that's, that's what would be next. This introduces a different type of ue. How does,

Leo Laporte (01:55:31):
How project this onto his hand?

Amy Webb (01:55:33):
So there's a, Ima the thing that I saw is like this little it's a little sort of oval, like a black oval that seems to have both a projector and a camera. Ah, yeah. So the idea would be if the demo that he showed, which again, I think was staged, but it was his like, wife calling and he, he knew that that was her because he, he

Leo Laporte (01:55:52):
Raised his hand and it says Bethany's

Amy Webb (01:55:54):
Calling. It

Leo Laporte (01:55:54):
Was, yeah.

Amy Webb (01:55:55):
Right. It also, by the

Leo Laporte (01:55:57):
Way, you know, the difference between raising my hand with a phone in it and not in a phone in it is kind of de minimus. I mean,

Amy Webb (01:56:03):
It's, I mean, yes and no. Yes and no. I think it put down distractions. You know, if, if I've got a phone in my hand and I'm take, I'm doing one thing, I I have so many other potential things I can then get into this. This is really more sort of single serving. But Matt, the thing that I also thought was really interesting is the input mechanisms visually. So imagine being able to hold a food item or something up and, you know, and the thing knows how many calories I've consumed or how much energy I'm gonna

Leo Laporte (01:56:34):
Need. Rash, ride or rash. And it could tell you whether it's, or a

Amy Webb (01:56:36):
Rash. There you go. Doesnt help.

Leo Laporte (01:56:38):
He has,

Amy Webb (01:56:39):
God, should I eat this thing?

Leo Laporte (01:56:40):
Or he has a hundred million in funding from kindred Ventures, including Sam Altman.

Amy Webb (01:56:47):
Yeah. So again, I don't know. I don't know that, I don't know how this thing scales. Well,

Leo Laporte (01:56:52):
I think, I'm gonna guess he's hoping to geting, he's hoping to get rehired by or purchased by Apple, because Apple's got a big problem, which is they're making a honk and big thing you wear on your head to try to make this happen. And that is not going anywhere. And

Amy Webb (01:57:07):
Well, I just, I think that this sort of constellation of new UI that are coming again is the more interesting piece of the ch of the AI conversation. Then

Leo Laporte (01:57:17):
The's funny because I, thought's voice assistance would be that new UI and they've failed us miserably. You were just yelling at Syria a minute ago. <Laugh>.

Phil Spencer (01:57:27):
Yeah. I, I, yeah, I, I'm, I'm really interested in the, in the new UI stuff. I know this project I've, I had talked with with him years ago when he kinda started working on this. And I, I think it's potentially, it's, it's a, it's a, it's a big move. I mean it's, you know, who knows whether it'll work out or not, but it's a really interesting idea. The projector, you're projecting things on your hand. I actually think it's, if he can get that working, that's very cool. It's like, I, I have doesn't,

Amy Webb (01:57:50):
You don't need a ton of tech to make that happen though, right? No. If it's not gonna be visually pretty

Phil Spencer (01:57:55):
And if it's talking to your, to your phone or something, anyway. It doesn't have to be the processing.

Leo Laporte (01:57:58):
Yeah. You still need a phone, right. And

Amy Webb (01:58:01):
Microsoft, no, not necessarily that. I don't, I think the idea is no phone. Microsoft in what 2002 or 2008 had skin put early, early days, which was, remember that? It was like you could, you could sort of project onto your forearm and interact with it.

Leo Laporte (01:58:16):
Yeah. Oh yeah. This is, this has a Qualcomm Snapdragon and it, it's running Android.

Phil Spencer (01:58:22):
Yeah. right.

Amy Webb (01:58:22):
And I think it's meant to be a, a, a Phoneless device. Right. So you don't need the phone.

Phil Spencer (01:58:26):
It it, it is, but the, I guess like the devices will continue to shrink. Right. The interesting thing is the, is the, is the interface modality. So sticking as your hand, then having a project on there and then recognizing gestures is really cool. And I agree that probably better than, than something you're wearing. You're strapping onto your face. This was sort of my central thing. I, I don't know what Apple's gonna release. I'm actually, I'm, I'm hopeful. I'm optimistic. I'm a giant Apple fanboy, like everyone knows. But I've said for a long time that like any, any technology, any consumer technology that makes you look stupid while you're using it, it's not gonna work.

Leo Laporte (01:59:01):
Yes. As we learned from the Segway.

Phil Spencer (01:59:04):
Oh, exactly. The Segueway and the Google Glass and, and all of the, all of the Facebook meta nonsense. Yes. I should have seen it coming cuz like literally you know, I think, I think, I think Zuckerberg's vision was always that you know, he is gonna ask people to strap something the size of a book onto their face. And they called it Facebook and like, we should have seen it coming. Should have known

Leo Laporte (01:59:25):
<Laugh>. Do you think

Phil Spencer (01:59:26):
Actually telegraphed it that far in advance? Do

Leo Laporte (01:59:28):
You think mark has given up on vr, on Meta the Metaverse? I don't. I

Phil Spencer (01:59:34):
I don't, I wouldn't know. I haven't. I haven't.

Leo Laporte (01:59:35):
No. I don't know. No one knows. But

Amy Webb (01:59:38):
Wasn't there an earnings call where he basically said that

Leo Laporte (01:59:40):
We're, yeah, we're moving to AI now. Pivoting a little bit. Yeah. Yeah.

Phil Spencer (01:59:43):
It's all AI now I I I'm interested to see what Apple does with it, just because like I bet you it'll be super cool. I still think here's gonna be my test for, especially for using it in a business application. This is, this is, this is the live and test. I am not gonna use any device to have a meeting or for work if I cannot drink a hot cup of coffee while wearing it. If I, like, if I literally can't do this cuz it like hits the thing and I spill hot coffee on myself. Is that

Leo Laporte (02:00:08):
An Ember mug? Is that one of them USB heated mugs you got there? Or is this a regular mug? Okay, this

Phil Spencer (02:00:14):
Is my new, this is a new restaurant. You've

Amy Webb (02:00:15):
Had like five. What are you drinking over there? You've notice

Phil Spencer (02:00:17):
Lot of them notice. Yeah, I have. I've had I just, I I'm drinking tea. But it's a, it's a really long show, <laugh>. And so I just set myself, what kinda

Amy Webb (02:00:25):
Tea are you drinking?

Phil Spencer (02:00:26):
<Laugh>? I'm drinking Earl Gray with lemon and a little bit of honey cuz I have my, my voice is a little bit sore.

Leo Laporte (02:00:30):
I can't do the burger mug. So can't do the burger.

Phil Spencer (02:00:34):
So basically when Apple releases their thing, that's just the test. Like, can you drink a hot cup of coffee

Leo Laporte (02:00:40):
While

Phil Spencer (02:00:40):
You're, if you can't, then like, yeah, then I'm not, then I'm, then it's like, then it's a hard no for any kind of business usage.

Leo Laporte (02:00:45):
I honestly, I feel like Apple and it's, it's got too much momentum to stop. But I think they've made a horrible mistake.

Phil Spencer (02:00:55):
I'm reserving judgment till I see it. I'm excited to see it. I love Apple stuff. Right. we'll, we'll see. You

Amy Webb (02:00:59):
Don't think what's coming are just a pair of old school analog looking glasses that

Leo Laporte (02:01:04):
Have someday Not, not, not in June.

Phil Spencer (02:01:07):
I hope so. They're gonna

Leo Laporte (02:01:08):
So in June that is gonna look just like the Oculus Rift.

Phil Spencer (02:01:12):
Yeah. If, if they, if they can release something that's just glasses that I'm actually looking at the world through clear lenses. Yeah, that would be awesome.

Leo Laporte (02:01:18):
Cameras, we all agree to

Phil Spencer (02:01:19):
Tech. That would be amazing. The tech, I'm

Leo Laporte (02:01:21):
All over that, right? You

Amy Webb (02:01:22):
Well, it can br I'm married to an eye doctor who repeatedly tells me every day why that can't happen. <Laugh>.

Leo Laporte (02:01:29):
Well you wrote, wrote in a treatment for a wonderful show that unfortunately got canceled. People had these glasses and they were sending stuff You helped them with the right. I did. What was the name of

Amy Webb (02:01:41):
That? It was called The First,

Leo Laporte (02:01:42):
I really liked that.

Amy Webb (02:01:44):
Yeah. Hulu's first Streamer. Yeah. Yeah. Our first bingey streamer. Yeah. I, the, there is a challenge trying to get the, the

Leo Laporte (02:01:52):
So your eye Dr. Guy says, no, can't happen. Not gonna happen. You, you

Amy Webb (02:01:56):
Know, I don't, I I don't know why, but I for some reason was like, I wanna read some of those original magical Leap patents again. So that's, that was my version of fun. Last week I pulled a couple. Yeah. Yeah. And they were just, they were so far ahead.

Leo Laporte (02:02:11):
Yeah. But they couldn't make it. They just

Amy Webb (02:02:12):
Needed, they needed 10 years. It was one of these things where it was like, that's right. Somebody had very interesting ideas. Right. This was basic research and they just needed to put their heads down for a decade or 15 years and just like work. And we

Leo Laporte (02:02:26):
Talked about that. Didn't let them do that. You were on the show at one point and Yeah. That, yeah, that was the problem is venture capital, those original 10 years were to do anything. Yeah.

Phil Spencer (02:02:35):
Fascinating. I, I will buy immediately whatever Apple releases around it headset and I,

Leo Laporte (02:02:40):
I'm gonna have you on

Phil Spencer (02:02:41):
And we'll see. We'll back

Leo Laporte (02:02:42):
June 12th. We'll have you on the show, <laugh>. Cause all I'm saying is I'm buying it <laugh>

Phil Spencer (02:02:47):
If, if like, if I can't drink coffee, I'm not gonna be in a meeting. So if wearing this means that I, that I, that I can choose between seeing 3D representations of my coworkers and having coffee, that is not just

Leo Laporte (02:02:57):
Apple's big innovation. Will they they'll have legs.

Phil Spencer (02:03:01):
Enough Legs. Yeah. Apple needs. They'll something maybe, maybe like a sippy cup as well. Maybe they can come up with a sippy cup. <Laugh>.

Leo Laporte (02:03:07):
Oh, you know, you could just have one of those backpacks with a little, little thing just attached. And then you'd be all set. Camel Pack Camelback if Apple

Phil Spencer (02:03:16):
Releases it. Cuz I'm not, I'm not going third party for

Leo Laporte (02:03:18):
This. You need an Apple Camelback Apple. That'd be,

Phil Spencer (02:03:19):
Yeah. They need an Apple Camel pack. Yeah.

Leo Laporte (02:03:22):
There is a kind of bizarre TV show on free V app. Amazon primes free ch Aded supported channel called Jury Duty.

Amy Webb (02:03:33):
Oh, I've heard about that. It's

Leo Laporte (02:03:35):
Interesting.

Amy Webb (02:03:35):
Hilarious.

Leo Laporte (02:03:36):
One guy is, is is not in on the joke. He thinks he's on a regular jury in a regular trial. Everybody else, the judge, the bailiff and all the other jurors are actors, including James Marsden as himself. But there's one <laugh>, there's one guy who is a what do they call those guys who are trying to do body modifications for the future? There's one of the jurors is is that kind of guy. And he's, he's wearing a camelback, he said. But the ideal would be have some sort of bladder that you insert into the body that contains all the water. And he has another one that has all the nutrition he needs. He's got anyway. Yeah, it's pretty funny. It, the joke doesn't make it through all eight episodes or whatever, but it's the first few are

Amy Webb (02:04:20):
Humorous since we very slightly sidetracked. Did you guys start Silo yet? Yeah.

Leo Laporte (02:04:24):
I love it. I hadn't read the books. You read the books.

Amy Webb (02:04:28):
A friend of mine read the books and, and said they, he works at aws. He's a super nerd. And said the books were amazing. Yeah. He loved the

Leo Laporte (02:04:34):
Books. So I feel like I should have read the books now. I'm watching the show. It's gonna spoil the books. And the books are always better in my opinion than any, any motion picture TV show. But it's, it's pretty good. I missed the most recent episode. I gotta watch it. Apparently it takes quite a turn.

Amy Webb (02:04:50):
Like that's on Apple, right? Yeah.

Leo Laporte (02:04:51):
Yeah.

Amy Webb (02:04:52):
I think Apple's struggling with their content, their, their studio content. Yeah,

Leo Laporte (02:04:56):
It's interesting

Amy Webb (02:04:57):
From my point of view. Yeah, there's the new the big whatever machine. I tried a little bit of that. Yeah. I I think that there's this formula they're following and they have like the world's largest catalog of music. So there's a lot of interstitial that I think are a lot of

Leo Laporte (02:05:09):
Music and all of their shows. You could tell they're

Amy Webb (02:05:11):
Selling in the

Leo Laporte (02:05:12):
Soundtracks. I think

Amy Webb (02:05:14):
It's, and and they've, it's, it's,

Leo Laporte (02:05:15):
It's the Zelda problem. They're, it's all formula, not enough human, right? Mm-Hmm. And that's what H B O did so well in the early days, I I despair of the future for H B O. But in the early days, they were famous. They would say, we're giving you money, we're not gonna give you notes. Just do what you wanna do. And they got the Sopranos and they, I mean, they got amazing stuff. I have to say. Succession. What now you work in business. So do iss I think succession's the best TV show I've ever seen, but is

Amy Webb (02:05:44):
I have firsthand experience with a lot of those, those people storylines. Yeah. And the people who were, and I will tell you that it's

Leo Laporte (02:05:54):
Accurate.

Amy Webb (02:05:55):
I think I, I'm wondering actually, if there was somebody on the inside leaking things,

Leo Laporte (02:06:01):
They apparently did have some Murdoch s info. I,

Amy Webb (02:06:08):
I think there had to be.

Leo Laporte (02:06:09):
Yeah.

Amy Webb (02:06:11):
Yeah. But I, I I'm, I'm in a sh i, I, I wish I had time to play games. At most, I've got like 45 minutes of free time occasionally or a half hour. And I just, I can't get into Zelda for You

Leo Laporte (02:06:23):
Have to really, we

Amy Webb (02:06:24):
Wish to, God, I could,

Leo Laporte (02:06:25):
You have to need,

Amy Webb (02:06:26):
I need like a block of time. You have

Leo Laporte (02:06:27):
To ration your your amusements because you have important duties.

Amy Webb (02:06:33):
I do. I need

Leo Laporte (02:06:34):
A new, you know what might help? I just wanna show you something. One of our sponsors. It's might help our show today, brought to you by Miro. We've been playing with this a little bit and I am very impressed with it. But I have a, I'm challenged, I don't know exactly how to describe what Miro is. If you have a team, you're working on something, maybe the team's distributed. Some are in Bentonville, Arkansas. Some are in Rome. You've got multiple tools. You know, you're using a design tool and a typing tool, maybe some AI and all these brilliant ideas. They're kind of getting spread out and losing information along the way. Miro, the whole idea, Miro, is it's a collaborative visual platform that brings all the great work together. No matter where you are, whether you're working from home in a hybrid workspace, your team might be in different time zones.

(02:07:26):
It's, it's got it all in one place. One single source of truth that is so flexible. It can be anything you need. It democratizes collaboration and input. It's great for Zoom meetings. They actually have a timer and icebreakers all sorts of tools. Their infinite shared boards give product teams a perpetual space where they can drag and drop insights and data. Nothing's lost. You can zoom in or zoom out. They cover so many use cases. You can build visual assets, you can present findings, run brainstorms with cross-functional teams, CanBan boards, voting tools, live reactions, sticky notes, even a timer. You can express yourself. But it's so hard for me to describe what Miro does. Maybe the best thing to do is to go to miro.com/podcast. Sign up your first three boards are free, and then check out the Miro verse, because this is what people who are using Miro have done with it.

(02:08:27):
And it will give you a very good idea. Actually, you don't even have to a, you know, you could just look at it. Here's a Harry Potter retrospective from the UK government. I don't quite understand <laugh> understand that. Tiktok marketing strategy, team mapping product vision, midnight sailboat, retrospective. I, you know, it goes on and on. Oh, here's one for Mother's Day. But this is the point of it. There's even games that people have designed in Miro. Miro can be whatever you need it to be. It integrates with dozens of other third party tools. We're using it with Zapier cuz we use Zapier to trigger stuff. Well we use it with Google Docs cuz we use a lot of Google Docs. Miro users report saving up to 80 hours per user per year by streamlining conversations, by cutting down meetings. And, but more importantly, than the time saved, the product you create is that much better because your team is staying connected to a real time information, a single source of truth. It gives project managers and product leads a bird's eye view of the whole project. So nothing slips through the crack. Nothing I say will really communicate what this can do for you. You just gotta try it. So get your first three boards absolutely free. Start working better at miro, m i r o.com/podcast miro.com/podcast. And you know, the, the whole world is available for you. You can do an amazing thing with Miro. We had an amazing week this week at twit. We've made a little little movie for you. Enjoy

Leo Laporte (02:11:40):
Microphone. Against the microphone. Actually it reminds me there for about 15 years there were audible click sounds in radio broadcast and other things. Turns out it was some Russian, giant Russian antenna that they were trying to use to, to spy on people. And it was causing radio interference. It wasn't aliens. I think the odds are it's probably, probably not aliens. I hope you enjoyed this week in twit and I hope you will continue to watch all of our shows. Of course, our Club TWIT members get special access. No ads, no promos, no spy, no you know, ad monitoring technologies, just the shows all clean, pure. You know what we should do? We should leave the swear words in for the club members so they <laugh> so they can hear what it really sounded like. Club TWI is $7 a month. Add free versions of all the show special programming we don't put on anywhere else.

(02:12:33):
That's how this week in space got started. Home Theater Geeks is right now in the club in it's you know, it's birth pangs. We've got the hands on Macintosh with Micah, hands on windows with Paul. We've got an Untitled Lennox show, all sorts of stuff. The Great Club Twit Discord. I think it's the best seven bucks you'll you'll ever spend. And it really helps us do I think more of what we wanna do for you. You know, ad supports, nice listener supports, even better. Twit.Tv <laugh> slash club twit. There is, there is the version of the show from Aliah, Aliah g <laugh> with all the swear members. I'm a I'm a member. You are, you hang out there once in a while. I love seeing you in there. It's wonderful. You know, I would love to get all the smart.

(02:13:17):
In fact, Phil, I'll send you complimentary membership too, but love to get all those smart people in here. It's fun because there's something about a social network where people have paid seven bucks to be there. They just have an investment in it. It's like a, you know, a private park. You don't Yeah, you don't, you don't get the messing around stuff as much. It's really great. AM radio is on Its Last Legs, Washington Post. This broke my heart cuz I'm a longtime radio guy and of a love affair. AM radio being removed from many cars. Ford, the M w Volkswagen Tesla they kind of have the excuse, well, the Electric motors interfere with the am but really it's, you know, why put it in if people are gonna be using their smartphones, it does raise an issue because am radio's used for safety announcements, emergency announcements, and there are plenty of places where you can't get a cell phone signal. So I imagine there'll be a quite a bit of lobbying from broadcasters to keep it alive. What do you, do? You guys ever li you ever listen to the radio? Amy?

Amy Webb (02:14:22):
I, yes. I listen to college. There you go. Usually whenever I'm in a new place, I listen to a college station. College

Leo Laporte (02:14:28):
Radio is much better, isn't it? Yeah,

Amy Webb (02:14:30):
Good. Like I, I get good, you know, introduced to new music that way. Yeah. But I was actually on an AM radio show. I'm trying to remember the name of it. It's a NA national show in the middle of the night.

Leo Laporte (02:14:40):
Oh, art. Forget the name of it. Yeah. Coast to Coast. Coast

Amy Webb (02:14:43):
To coast.

Leo Laporte (02:14:43):
Coast to Coast. Art Bells old show now. George Norie hosts it.

Amy Webb (02:14:47):
Yeah. Yep. I was on coast to coast. Did he

Leo Laporte (02:14:49):
Ask you about aliens?

Amy Webb (02:14:51):
It was mostly call in people. So I was on from like 2:00 AM to 4:00 AM Oh wow. Or something when I was launching the last book and Oh wow.

Leo Laporte (02:14:59):
That's quite a booking. Wow. Well,

Amy Webb (02:15:02):
Here, here's what it was. It was a lot of people who were really worried about vaccines. Yeah. And they listen, if you're up in the middle of the night, you're,

Leo Laporte (02:15:13):
You're worried about something, you know,

Amy Webb (02:15:14):
<Laugh>, you're, you're probably anxious. And, and it occurred to me, so Zelda comes into play. So listening to everybody, it occurred to me that they all must feel like everybody was really talking down to them and just treating them like they were dumb. And I, I think when you're genuine, genuinely anxious and afraid, that is not the right response. So I was like, actually one guy who had called in, I was like, do you ever play Zelda? And he was like, yeah, I play Zelda. And I was like, okay. The, this is Covid is a little, and the mRNA vaccines are a little bit like, you know, you're out roaming around high rule and you meet, what are those blobs? I can't remember the name of them, but the blob comes by and like the first time you see the blob, you don't know, is this food, is this a good blob? Is it a bad blob? Maybe. You know, it's hard to tell. And since you can't recognize what it is, you don't know what to do with it, you're, you're missing the instructions. Once you have the instructions, like, oh, that blob is actually a bad thing and it's gonna kill you. You know, which weapon in your quiver to, to pull out and, and to use to defend yourself against it. That is what the messenger RNA vaccine does for you. Right. They're actually

Leo Laporte (02:16:24):
Called blobs, by the way. That's the official official

Amy Webb (02:16:28):
Name. <Laugh>. Yeah. So I mean, it was like, and we just had this moment where he, I heard, he was like, nobody ever, why, why don't they just tell us

Leo Laporte (02:16:36):
That? Oh yeah.

Amy Webb (02:16:37):
I don't know. And it just, it, it so I mostly just listened to people being really worried and telling them like, your fears are totally valid. Like, I, you know, like, I get it. This is hard stuff. So it was very good bad for two hours.

Leo Laporte (02:16:51):
Very good. Yeah. That's, anyways,

Amy Webb (02:16:54):
I guess that show is

Leo Laporte (02:16:55):
If AM radio dies it's because broadcasters have killed it basically. Right. I'm

Phil Spencer (02:17:00):
Surprised that it's AM and not fm. Right. Cuz like you would think that like, well, FMS actually more useful. Yeah. But I would've expected it to go first. Yeah. It's like we don't need radio for like higher quality and ams Am am gives you the range and it's like, and

Leo Laporte (02:17:14):
Information and it works everywhere

Phil Spencer (02:17:15):
In the apocalypse. You'd rather, you'd rather have That's a good point.

Leo Laporte (02:17:18):
Radio. Why aren't they taken any

Amy Webb (02:17:20):
Serious shows are still commercials. Yeah. So it's not like you've Yeah. I don't know.

Leo Laporte (02:17:27):
I was very I thought I'd, I'd bring this one to you because of your book. The Genesis machine Amy in in the central California area, there was some deep concern. We had so much rain about mosquitoes. Mm. And there was an attempt, attempt was made to use Oxitech, which is a US owned, UK based technology company that generates genetically engineered insects. And they have a mosquito that is basically destroys the Egyptian mosquito, which spreads dengue fever and chok and Zika fever and myro and yellow fever. The genetically modified male that they release mates with regular mosquitoes, but but like a, like a donkey and a horse. There's no, there's no you can't <laugh>, there's no mosquitoes born. So it actually is a pesticide less way of killing mosquitoes.

Phil Spencer (02:18:35):
Isn't that how you get mules though?

Leo Laporte (02:18:37):
Well, no. Yeah, donkey and horse a bad example. But a mule can't reproduce because it's, you know, it's got no, it's, it's, there's no, there's no mules. Don't mate <laugh>. So the company

Phil Spencer (02:18:48):
Self-Esteem, I think it's got no self-esteem,

Leo Laporte (02:18:50):
No self-esteem cuz it's a mule. The funny thing is, they have used this successfully in Brazil, Florida and Florida maybe, I don't know, but green, but green activists in Central California got them to kill it. They're gonna have mosquitoes

Amy Webb (02:19:07):
To say, why does it say why?

Leo Laporte (02:19:08):
Because it's genetically engineered.

Amy Webb (02:19:12):
Look, this would take us a long time to untangle, but there's a couple of things happening. The first is the city council, the, it's not a true city council, but the people sort of the, the community council in charge of Key West at the time that they had to decide whether or not to use this technology. They were like a couple of retired real estate agents. Another guy was like a retired I restaurant tour or something. None of you know, they're retired in a beautiful place. The last thing that they wanted to be doing was to make heated, you know, challenging decisions in a heated political environment about genetically engineered mosquitoes, obviously, about which they knew very little. The issue is that we're forcing communities to make these decisions. And the people and, and the states and the federal government and the states are intentionally handing this down so that the political heat is off of them. And I think it's a horrible, ah, interesting, horrible thing that we are asking locally elected officials to do. It puts

Leo Laporte (02:20:12):
In a, in this case a terrible position in California. It was friends of the Earth who lobbied the California Department of Pesticide Regulation, presumably people who would know enough science. And they convinced them to withdraw the research application and then they published a press release, said victory. The, yeah,

Amy Webb (02:20:30):
But who knows? I mean, listen, it's, we don't, we don't know the political what was happening on the backend. There are solutions, there are knock on effects and there are people who talk about genetically modified as though that is some type of ultimate like, horrific thing I in the United States at this point. It is it is very difficult to consume something that has not in some way been modified. Right. I mean, it, it, but, you know, it's, it's very, so it's again, it, it sort of mirrors some of the talk about ai, which is just completely salacious and over the top. And we're not taking five seconds to be reasonable, which makes me crazy.

Leo Laporte (02:21:19):
Well, you're gonna have your mosquitoes in the Central Valley and you can thank the first of the earth.

Amy Webb (02:21:24):
Yeah. I don't know.

Leo Laporte (02:21:27):
Spotify has ejected thousands of AI made songs. They're, by the way, they're all terrible <laugh>. But, but this is a flaw. People can gamify Spotify by there's a whole bunch of songs about poop, for instance, on Spotify because they know kids will ask Alexa for it as soon as they find out that they can ask for a poop song. They will. And they make money <laugh> on it. Now that I know, I know what I'm Yeah, you can't wait till you try that out tens of songs. Can I ask you a question though? Songs though. Why

Amy Webb (02:22:01):
Is, but why is that? Why, why is that bad?

Leo Laporte (02:22:05):
That's no, I'm not saying, did I say it was bad? I'm just saying. No,

Amy Webb (02:22:09):
I meant why would Spotify kill the, why are they removing them? What's the damage? Let

Leo Laporte (02:22:14):
Just taking money away from legitimate artists, I guess. Yeah. I don't know.

Amy Webb (02:22:20):
And I mean, I realize that I literally haven't listened to real new music with any regularity posts like 1997. That being said, a lot of music, let's face it, kind of is interchangeable. And

Leo Laporte (02:22:34):
Can I recommend, if you want to catch up on all this stuff, watch Eurovision, which for the first time was broadcast the United States on I think it was on Peacock this year. You <laugh> you, you'll be amazed and how bad some of the music is. Oh, I know spoilers here. I I don't, I'm sorry if you saw that who won in all of that, but this was something I watched last night and I was thinking this is why the US music industry is so dominant world worldwide.

Amy Webb (02:23:04):
<Laugh>, I I think it's rich that Spotify, let's talk for a second. I forget who did this study, but somebody looked at how Spotify is putting together the sort of recommended daily lists. Yeah. And they're not random. So if you look at, if you, if you look across multiple accounts, everybody's basically being recommended the same songs over and over again. Ah. So I don't care what anybody says, the right discoverability. It's it, you see this across different platforms in the digital arena. The same thing happens with search. The same thing happens, you know, with music. It's, it's not as though everybody truly is getting a random assortment of enhance if you click on enhance in Spotify. So like, you know what I mean? Like, so what does it matter if Yeah. People are making ai.

Leo Laporte (02:23:54):
Yeah. I, I

Phil Spencer (02:23:55):
Played, I've, I've been listening to the cheese text song on a, on a continuous loop from TikTok. Just wait a minute. Cheese tax.

Leo Laporte (02:24:02):
Cheese tax.

Phil Spencer (02:24:04):
Cheese tax. It's a, it's a dog thing. I have a new dog. And

Leo Laporte (02:24:08):
Is there a cheese tax?

Phil Spencer (02:24:10):
Cheese tax?

Leo Laporte (02:24:10):
Should I be worried about this? If

Phil Spencer (02:24:12):
Our reg, if our regulator,

Cheese Tax song (02:24:13):
You gotta pay the cheese tax every time you're cooking. When the cheese comes out, this puppy comes, this it. The rules or the rules and the facts. Or the facts. And when the cheese draw opens, you gotta pay the tax.

Leo Laporte (02:24:25):
Okay. <laugh>

Amy Webb (02:24:27):
Does your, does your dog like that song?

Phil Spencer (02:24:29):
There's like, my dog likes my dog enforces the cheese tax. It's extremely,

Leo Laporte (02:24:33):
He gets a little bit of whatever cheese comes out of the cheese jar, the dog gets. Okay, that's

Phil Spencer (02:24:38):
Fair. There's thou thousands. That's fair. Tens of thousands of these clips. <Laugh>. And yeah. This is why I hope TikTok will never get banned. <Laugh>

Leo Laporte (02:24:47):
Mr. Phil Libin, thank you so much for spending some time with us. I hope you enjoyed your five cups of tea. You didn't run out Three.

Phil Spencer (02:24:55):
Only three Only were

Leo Laporte (02:24:56):
Three were large. Oh, good. Only three, but

Phil Spencer (02:24:57):
They were large. See,

Leo Laporte (02:24:58):
You didn't run out. We did, we did We make it through the entire show without a tea shortage. <Laugh>. I think it's hysterical that you noticed that, Amy. I'm very impressed.

Amy Webb (02:25:07):
Well, I've got a bunch of vessels on my desk too, but I've been drinking black coffee out of this one and Topo Chico out of this one. Perfect.

Leo Laporte (02:25:13):
Perfect. You cannot make a Molotov cocktail out of Topo Chico, by the way. I've, I've learned that. Is the

Amy Webb (02:25:19):
Gra the glass is too thick

Leo Laporte (02:25:21):
Or Yeah. No, it doesn't burn. <Laugh>. <laugh>.

Amy Webb (02:25:25):
Someday you'll have to tell us why you know that

Leo Laporte (02:25:27):
There's

Phil Spencer (02:25:27):
A next time, next time I'm on, I'll, I'll be plugging. Or my new restaurant, which is be called Bentonville.

Leo Laporte (02:25:32):
Bentonville In.

Phil Spencer (02:25:33):
Beautiful. In, in Bentonville. Bentonville. Oh, that's

Amy Webb (02:25:35):
A great idea. That sounds

Leo Laporte (02:25:36):
Cool. Are you gonna start a restaurant? It'll have Bento boxes in Bentonville.

Phil Spencer (02:25:40):
Yep. And it's gonna hold Bentonville. It's all happening. I love that you have the website now you can see, but it's all really, all the popups are sold out, so you can't get anything right now.

Leo Laporte (02:25:49):
And this is your restaurant

Phil Spencer (02:25:50):
Ville. Well, I have partners who know what they're doing.

Leo Laporte (02:25:53):
The problem, unfortunately, when I search for Bentonville is Google's too smart and says, oh, you mean Bentonville?

Phil Spencer (02:26:00):
Yeah. That's gonna change <laugh>.

Amy Webb (02:26:02):
I I love it. I think that sounds so cool.

Leo Laporte (02:26:04):
The world is gonna come to Bento. Yeah. Mr. Bentobox, Phil Libin his app is mm-hmm. <Affirmative>. Mm-hmm <affirmative>. It's Mmh. M m.app. Gotta try it. Mac Ann Windows. You can present with it. You can change your backgrounds with it. It makes every Zoom call better or WebEx or Teams or meet your video companion. Thank you. Phils so smart. I just love having smart people on the show. Then I could sit back, relax and enjoy the ride. Bentonville, that just blows my mind. All turtles come visit.

Phil Spencer (02:26:40):
I'll show you around.

Leo Laporte (02:26:41):
Yeah, I will. Amy Webb, just great to have you. The book is the Genesis Machine, our quest to rewrite Life in the Age of Synthetic Biology. Future Today Institute is her company. What do you wanna plug? Anything?

Amy Webb (02:27:00):
Yeah, I mean, I guess if you're like interested in ai, the book that I wrote previous to this one is the Big Nine and it's

Leo Laporte (02:27:07):
Still Janine Piros favorite.

Amy Webb (02:27:09):
Janine Piro is endorsed it. So

Leo Laporte (02:27:11):
<Laugh> Actually it's a great book. I I plug Po Chico, I have all your books. I and I brought this one out cuz it's the most recent, but I should have brought the big nine out too. Next time I'll know better. Had I been watching Fox News, I would've known <laugh> how tech titans in their thinking machines could warp Humanity came out four years ago. It was prescient. Both these people knew what was coming long before I did. Thank you so much for being here, both of you. We do twit every Sunday afternoon, 2:00 PM Pacific, 5:00 PM Eastern, 2100 utc. Thank you for watching. If you watch live, you can chat live@irc.twi.tv or in our club Twit Discord after the fact. The show's available at the website, twit.tv. And when you get there, you'll also notice a link to the YouTube channel. You can watch it there. There's also a link to various podcast players or just search for this week in tech or twit on your favorite podcast players. Subscribe. You'll get it automatically the minute Sun, just in time for your Monday morning commute. Thank you everybody. Have a great week. We'll see you next time. Another twit is in the can.