Loading summary
Transcript1,123 lines
- [00:00]
A
It's time for Intelligent Machines. Paris and Jeff are here. Our guest this week, MG Siegler, longtime venture capitalist with Google. He now writes about AI for his own newsletter, Spyglass. He's a great interview and he's coming up next. Then we will talk about the new Google image model. We play quite a bit with it. It's called Nano Banana. And how do you load the dishwasher? I guarantee you somebody thinks you're doing it wrong. All that more coming up next on Intelligent Machines, podcasts you love from people you TR trust. This is twit. This is Intelligent Machines with Paris Martineau and Jeff Jarvis. Episode 834, recorded Wednesday, August 27, 2025.
- [00:47]
B
Goo Gaw.
- [00:49]
A
It's time for Intelligent Machines, the show where we talk about AI and robotics and all the smart little doodads and gee jaws surrounding you these days and all of us these days. Speaking of surrounding, I'm surrounded right now by Paris Martineau, investigative reporter, Consumer Reports. Hi, Paris.
- [01:08]
C
Hi, Leo. Don't.
- [01:10]
A
Long time no see.
- [01:11]
C
I'm in your wall.
- [01:14]
A
It's only been a few days. You were great on Twit on Sunday. Thank you for joining us.
- [01:18]
C
Always lovely to return to the mothership.
- [01:21]
A
Yeah. I noted at the time that that's where we first found you was as a regular on Twitter. And I said when Stacy left this show, I said, gosh, I'd love. I don't think Paris will do it, but let me ask. Let me ask. And you did. And I'm so glad, of course.
- [01:35]
C
Now, who's that man who was playing a speech to text thing into the microphone?
- [01:39]
B
Because I was correcting Leo. He was doing a gift gif thing to another word. Is it.
- [01:49]
A
Who are you asking?
- [01:50]
C
I don't even know what we're referencing, to be honest.
- [01:53]
A
I said all the doohickeys and jee jaws surrounding you. Well, what do you say? Gee whiz or gee whiz?
- [02:00]
B
I said I used to say gee gauze, but I certainly just don't say G Jaws. That's completely.
- [02:05]
A
So what if it's G Goz. Did you ever think of that? Say it. What are you playing? He's saying galga. He's saying it wrong. What the hell's that? That's not what I say.
- [02:21]
B
Gee gaw. But it's spelled. Spelled Dugov.
- [02:24]
C
And who are you?
- [02:26]
B
Oh, hi, I'm the pedant.
- [02:29]
A
The pedant. Jeff. He's a professor. It's okay, he's allowed. Jeff Jarvis. That goes with the job title emeritus professor. Is Even worse than Professor Emeritus Craig Newmark.
- [02:39]
B
Newmark.
- [02:40]
A
It's gotten to the point now, I don't even have to say it. Emeritus professor of journalistic innovation at the Craig Newmark Graduate School of Journalism at the City University of norc.
- [02:50]
C
Norc. Today's a bad day for language.
- [02:52]
A
He's now at the State University of norc. Beautiful Stony Brook. All right, get serious because we have a serious guest on. Do you mind? Do you mind? MG Sigler is an old friend, been on Twit many times in the past. He then left for a great career as a venture capitalist. Were you at True Ventures? Google Ventures. Where were you?
- [03:15]
D
MG Google Ventures, yeah.
- [03:17]
C
And is that Google or Googa.
- [03:21]
B
Jugle Or Google.
- [03:22]
A
Now he's adjourned to London, which is pretty cool.
- [03:27]
D
Yes, hence the darkness here.
- [03:29]
A
But even before you were a vc, you were, I think a very successful, well known tech journalist. And I think you have a little hankering to get back into the game.
- [03:42]
D
Yeah, I did the VC job for a little over a decade and that was good fun.
- [03:49]
B
How was that transition?
- [03:51]
D
It was good. I mean, you know, now it became sort of a cliched thing right there, became this sort of TC to VC thing. I was at TechCrunch for a long time.
- [04:00]
A
Yeah.
- [04:01]
D
And it became sort of a more well trodden path afterwards but. But honestly it set me up well. I didn't go to school to sort of become a journalist. I stumbled into it and so when the opportunity came to sort of move onto a tangential career, I jumped at that. And yeah, it's sort of finding interesting companies to write about was the same high level mandate. Obviously there's some differences in the, in the day to day, but yeah, it worked well I hope. Knock on wood for me. And now I'm just still love writing so trying to.
- [04:37]
A
Well you're great and you still a little bit have your hand in business. It says according to Wikipedia you're on the board for Slack.
- [04:44]
D
Well, I was before it went public. Yeah.
- [04:46]
A
Wikipedia. And you invested in Medium and Secret, which is cool.
- [04:50]
B
God bless. Yeah.
- [04:52]
A
No longer 500 dish words. I did enjoy that though. I love the conceit that you could get anything done in 500 words or less.
- [04:59]
D
Thank you. I was always, I've always been trying to, you know, come up with little mind hacks to get to get me spurred writing again.
- [05:05]
A
But this current newsletter, Spyglass Insight from Afar, not so far, just across the Pond, is actually well worth reading and I've been trying to get you on for a while. Because you write a lot about AI. By the way, I did have to say I agree with you about Google's distractingly smarmy Pixel event. We talked about this last week.
- [05:30]
D
Yeah, that was a weird watch again from afar, wasn't there? Of course. But yeah, it was.
- [05:37]
B
I don't know.
- [05:38]
D
I don't know how I feel about taking a late night talk show host to sort of do these things.
- [05:43]
A
Well. And look at Jimmy Fallon, your picture. I love what you picked. Adrienne Lofton was beautifully put together. Really great presenter. She's of course CMO at Google. But Jimmy looks like. I thought he was jet lagged until I found out the event was in Brooklyn. It was just, I guess he had a late night on the subway. I don't know what happened.
- [06:02]
D
Yeah, it just doesn't look all there. He was rough.
- [06:05]
A
He.
- [06:05]
D
Yeah, the whole thing just had a weird, weird vibe to it. Very strange. I don't think we'll see them do that again anytime soon. No, I think they got the message.
- [06:14]
A
Yeah. Pretty clear. What do you think they were trying to do? It was my impression that Google Thought still has ambitions of being a consumer device. The Pixel. Yeah.
- [06:25]
D
Yeah. I mean, you know, obviously everyone from the. For the past couple decades has been trying to emulate the Apple model of doing these things. The Steve Jobs keynotes and you know, I think Amazon and some others have tried to switch things up a little bit and had varying degrees of success. Some. Some less so. But I don't, I didn't hate the overall notion of what they were trying to do. You know, bring on someone who is establish Jimmy Fallon doing a sort of sit down and see if they can take some of the weight off the shoulders of the people who, while they're fine presenting, that's not what their actual job is.
- [06:59]
A
Right.
- [07:00]
D
And you think like, oh, maybe he can grease the wheels and get a conversation going a bit. But it just felt fake. Right. It felt stilted. The whole thing was like just awkward. And so again, I don't hate the idea to try it, but it just didn't work. From the.
- [07:13]
B
I thought the Googlers all did a better job than the stars.
- [07:16]
A
For sure.
- [07:17]
D
Absolutely.
- [07:17]
A
Yeah.
- [07:18]
D
Yeah.
- [07:18]
A
That's kind of part of the real embarrassment. The other thing was they had a compelling story to tell.
- [07:25]
C
I was saying the bar is just so low for the Googlers. It's really impressive that they did.
- [07:30]
A
No, Adrienne Lofton was incredible. She was really talented and really she could, I mean she was a presenter and Jimmy just was dialing phoning in. And the thing that bothered me the most was they actually had a story to tell which got buried under the raspberry 100%.
- [07:47]
D
Yeah.
- [07:48]
B
Yeah.
- [07:48]
D
I mean, look at where Apple's at right now. Right. They're trying to sort of come back around and they've obviously flubbed the AI story. Google has the great AI story with these devices and yet they like looking back at that event. I can't even remember sort of what the main features they were trying to convey were because it was so distracted by the whole Jimmy Fallon being excited about the color palette and being excited.
- [08:14]
A
About the getting it wrong. Incidentally, he picked up the wrong phone and named it the wrong color. She says moonstone, Jimmy. It's not purple. And it's funny because a couple of days later, Google revealed that they had been secretly testing a new image generation model, Nano Banana, and said, no, it's not Nano Banana. It's actually part of Gemini. And it is, I think, arguably the best video and picture generator out there now. And they didn't even mention that in the event last week. It's like you're burying the lead. I did this. This is embarrassing. I'm going to show you this because this is going to gross you out. This is a picture I took last night. We went to Italian restaurant and as many Italian restaurants do, they had a statue in the front of a fat chef.
- [09:07]
B
Only in California, Leo, they don't do this in the. We went back to that New Jersey.
- [09:11]
A
A fat chef holding the menu. And I said, turn this into a video with the guy slurping up some spaghetti. And this is the video I got a few seconds later.
- [09:23]
C
Oh, no.
- [09:25]
A
Oh, no. Okay, now it's disgusting to describe nightmares.
- [09:31]
C
A single about 6 inch long fat spaghetti strand floats up to, flies up to his mouth, then is slurped up, then three more fat spaghetti fingers come down.
- [09:43]
D
Yeah, it's a little. A little reminiscent of Will Smith eating spaghetti.
- [09:46]
A
I was gonna say this is Will sm. Wow. Okay, anyway, it is true, but it did a good job. And look at it. Made a video out of a completely still photo. I'll do some more.
- [09:56]
C
Wait a second. It did a good job. Question.
- [09:59]
A
Well, all I look is my fault. I said the guy slurping up some spaghetti up.
- [10:07]
C
I would argue it didn't do a good job of making a video of a guy slurping up some spaghetti, but it did make a fun video, that's for sure.
- [10:16]
A
All right, well, I'm going to give it something else. And while we're talking I will have it. I'm going to give it this image of a cow. Let's see here and see if it can make something happen. This is actually. Or maybe should I do the tortoises? I'll do the turtles and I'll say make them mate. How about that? That'll be fun.
- [10:37]
B
Oh, geez.
- [10:39]
A
It's going to be a while.
- [10:40]
B
I don't even know how turtles stop.
- [10:43]
A
Well, we'll find out, won't we?
- [10:45]
D
They're not going to let you do that, right? They still have some. They'll flag you.
- [10:49]
A
Oh, that's a good test of the safety features. So let's talk about AI Actually, I admit the time I wanted to get you on, you'd just written about Perplexity's perplexing offer of. What was it? $32 billion.
- [11:06]
D
34.5. Very precise.
- [11:08]
A
For Chrome. Which is more than they're worth, right? They're worth like 18. Yeah, yeah, yeah, yeah. And you, I loved your headline on the whole thing. You said perplexities, Perplexing and positively PR offer to buy Chrome. Why not offer a quadrillion dollars?
- [11:27]
D
Yes. Yeah. And I mean that's, that's going to hit any day now right there. They're apparently we're waiting for the judge.
- [11:33]
A
Any, any moment now. The judge should make a.
- [11:36]
D
Going to have to sell Chrome. I think there's probably a 0% chance that perplexity buys them. There's probably a less than 0% chance that OpenAI buys them, even though they could probably get the actual dollar amount to do it.
- [11:48]
A
This week there was an interesting offer of $0 from the eco friendly search engine Ecosia saying we'll make a foundation out of it. This is how it should be done. A nonprofit foundation out of Chrome.
- [12:02]
D
I mean, I do think that that's a better high level idea. Right. Rather than anything that the sort of judge would do would just be sort of taking the crown from one king and putting it on a, a new head versus if you did it as tried to.
- [12:16]
B
Yeah.
- [12:16]
D
Go to the actual open source or, or do something else that was more of a consortium. Maybe it would be a little bit better. But how would you unwind that? What, what, what would Google accept? They're going to be fighting this for years and years, obviously.
- [12:29]
C
What do you think the judges could possibly rule if we get a decision this week or next and two, like, what do you think? I guess separate from that, a possible remedy for this would be if you were the one making the decisions.
- [12:44]
D
I mean, so, you know, obviously I've been removed from Google for over a year. So I don't have no inside knowledge about what this is. And I didn't anyway. But you know, working for the venture arm. But I would say I would be sort of surprised if they do some sort of major action with Chrome. I do think, you know what everyone is saying the obvious stuff of that they'll sort of strike down the, the Apple deal, right, where they have to, they pay for the default placements. That seems like it's going to go away for sure. Chrome is a question mark and I think it's mainly a question mark because Google has done some stuff where they're like, you know, baking some of the Gemini things into Chrome, at least in the beta builds of it. Right. And it's like all the stuff now is talking about like, okay, search is sort of in its phase of being disrupted somewhat naturally by AI and so are we worried now about, you know, searches it stands for right now? Obviously all this antistrust stuff, you know, is usually done in hindsight, litigating it in hindsight. So how much does the judge actually care about like the go forward elements of this? And Tim Wu wrote the an op ed the other day in New York Times about it, like trying to seemed like very much sort of convey to the judge that they should be worried about the AI market as the, as the next frontier. Right. That, that this is what this is.
- [14:04]
B
But this is where Google has tons of competitors. AI itself is competitor research and there's tons of competitors in Google. But Tim Wu will always say yes, of course, right.
- [14:13]
D
That's, he's, that's his thing. But you know, so that's like the thing to watch for, like what the judge buys or doesn't buy in that regard. Like if, if there's any movement on Chrome, it probably has to be related to. Yeah, the notion that it would be used to transfer over Google's power in search to the AI world. And can you really litigate that without knowing it for sure to just point, like, there's a lot of competition in that space right now and it's going.
- [14:40]
B
To be in courts for years. Yeah, it depends on what does Google. Does Google bring a platinum block to.
- [14:48]
A
Trump.
- [14:51]
B
The gold block?
- [14:52]
A
How do you trump a 24 karat brick?
- [14:54]
B
Yeah, what's more valuable than that? And then he just says, no, I don't want this to happen and we'd like Google.
- [15:02]
A
What would concern me more, you've written about this as well, is that he might say, well, let's take a stake in the whole thing.
- [15:07]
D
I think the government, the U.S. government owning 10% of Chrome. That would be, that would be interesting with the.
- [15:13]
A
Yeah. Or you know, just give us Chrome and we'll take care of it for later. You could totally see that truth Chrome. Yeah. I don't know if Judge Mehta would go along with that. I don't think any of the options that the government has proposed are really starters. The other option is to, to get Google to stop paying Apple $20 billion, Samsung several billion dollars. Mozilla to be the default search which would kill Mozilla. Absolutely. And save Google billions of dollars. It's not a punishment, your honor.
- [15:47]
C
It's very interesting that we could potentially have Google be forbidden from, you know sending all of these billions of dollars to Apple at a time where Apple's potentially going to be paying Google billions of dollars to use Gemini AI Power series.
- [16:03]
A
That, that was an interesting rumor that came out this week. Apple has been you know Eddie Q apparently at Apple the information says has been pitching Tim Cook who keeps saying no Eddie, no Eddie to buy Perplexity or to buy somebody else buy some AI Vistral. That's right, the French AI. And there is a rumor this week that maybe Google and Apple will weirdly in a weird partnership get together and Gemini will become the heart of Siri, which I find a little far fetched. I can't imagine Apple saying to Google okay we give but they did on search. Can't do it.
- [16:42]
B
They said on search they would do that. I you know everybody's, it's, it's commodified. AI right now is commodified. So they haven't found the special Apple AI way.
- [16:52]
A
I don't, I don't think they would be capitulation though. It would be a complete admission.
- [16:57]
B
It's a holding pattern. It's a holding.
- [16:59]
D
Yeah right. It holds over until they get their actual teams in in order in house to be able to you know eventually roll it out just like they did with Maps.
- [17:07]
A
And what do you think?
- [17:08]
D
That's not a good example obviously but eventually happened.
- [17:10]
A
Has it hurt app? They have been you know, struggling with AI. Is that a problem for Apple?
- [17:16]
D
It's interesting. I, I was sort of thinking about this today sort of reading. Yeah that information article that you're talking about all their, their Eddy Q. Trying to push some of these, these no giant acquisitions.
- [17:27]
A
No Eddie, we're not spending tens of.
- [17:30]
D
Billions of dollars which obviously Apple never does. I would like obviously the Wall street sentiment right now. Is that it has hurt them. Like Micro, you know, their competitors, Microsoft and certainly, you know, and Nvidia is the company in the world, have shot past them. Right. In terms of market cap. And so Wall street views them as sort of just in a holding pattern. Not, not sure what the story is going to be with AI. I think the story is still unfolding. I think that there's a lot of different things yet to come. But I do feel like the main concern I would have about Apple in that regard is that they just don't have the right mentality internally. From what I've seen over years and years of writing about them, covering them, you know, dating back to reporting days, like they're just a.
- [18:14]
B
They're.
- [18:14]
D
They're a company that's built a very different way than the way that AI has sort of entered our world and entered these companies. Everything moves fast. Nothing is polished. It's all being done on the fly. Apple loves to put out the beautiful, polished gem out there. Right. And they famously say, we're not first, but we come in and we sort of, you know, do the, the better version of whatever anyone is doing. The problem is AI just hasn't slowed down enough for them to sort of plant a flag and be able to say like, we'll do the best version of this because it's still just constantly moving.
- [18:47]
A
Yeah. And the other problem they have is Apple really wants to promote the notion of privacy, which for a lot of people I know, I don't know if you're still the Apple partisan that you used to be, but a lot of people say, I only would trust Apple with this. I wish Apple would hurry up because this is how I want to use AI. I don't want to use AI with companies like Perplexity that say we're going to mine every bit of data that you give us, or even Google or Microsoft, they don't trust them in the way they trust Apple. And yet all of that privacy has held Apple back a little bit.
- [19:20]
D
Yes, that was the pitch, right. That they famously flopped on last year at wwc, that they base with getting your mom home from the airport and all of that, and being able to use AI and Siri to look at your contacts, and they just couldn't quite get it done. And I think there's a whole, whole range of reasons why that happened. But I do think at the end of the day, again, there's sort of this cultural and mentality within Google that. I'm sorry, within Apple that, that I think holds them back A bit from doing that. And that's why I think you hear a lot about the M and A topic come up. Because it's like that's a quick and you know, potentially band aid way to sort of jump back into the game. But I'm not sure that that type of M and A, you know, would actually go over well within, within Apple and it might actually backfire in some ways. Right. If they brough giant team working on, working on some advanced AI and just like they just culture clash within the company.
- [20:16]
A
Yeah, you've picked AI as one of the things you really want to cover. I don't think you could cover tech these days without really covering AI. I'm very curious what your take is on where we stand on. Are we going to get at some point AGI or super intelligence? Actually, your most recent post talks a little bit about this. The endless rebranding of AI.
- [20:39]
D
Right. I do think, yeah, we need to figure out what the actual terminology means to begin with. We've gone from AI, machine learning, AGI, super intelligence. Now we have different flavors of super intelligence that various different companies are working on. Meta has their own, Microsoft has their own Zuckerberg type of personal superintelligence. Like it's all obviously just branding and marketing. And in some ways though, it's a legal issue too, right? With famously with Microsoft and open AI, their agreements in the quote unquote clause that IF and when OpenAI were to declare AGI that, that they would sort of be able to sever the contractual ties with, with Microsoft.
- [21:19]
A
And so is that a weird thing? I have never heard of such a weird contractual agreement. Talk about an undefined term.
- [21:27]
D
Of course, I think, you know, if I, if I had to sort of guess as to how that played out, you know, several years ago it was just like OpenAI was in a weird spot, right? They had the money sort of fallout with Elon Musk, who was their main benefactor. They needed money from Microsoft. Satya Nadella comes in smartly, identifies the situation, realizes like, oh, we have an opportunity here to sort of partner with this company and we can sort of advance AI at arm's length. It's not done within Microsoft. It's owned by an outside company. We don't own the company we own or we control 49% of the profits or, you know, whatever, whatever the, the terms ended up being. And so I think then, yeah, it's like open AI pushing back a bit. Well, it's like, okay, we don't want you to control it fully. If we do actually, you know, stumble into something interesting. So let's put this AGI clause in there and let's just say like we get these intelligent machines that all of a sudden come out of our work. Like then we'll just sever ties and we'll take that. You guys can take the work that we've done before and we'll go forward with this. And it's like it was all just. Yeah, it's like fantasy at the time.
- [22:29]
A
How did the lawyers even say yes to that? Maybe everybody agreed. Yeah, it's, it doesn't mean we'll just stick.
- [22:35]
D
And I think we're seeing the direct, the direct fallout of that now. Right. The fact that these two sides are like warring with one another over these terms and trying to break these things and OpenAI obviously trying to become a for profit or, or a public benefit company corporation. And so yeah, it's like we're dealing with all these weird side effects of those decisions that were made on this like on, on the fly. It felt like.
- [22:58]
A
We're going to talk a little more about that next week. Karen Howe's book the Empire of AI. Karen will be our guest and she about how desperate actually OpenAI was when Elon pulled out that they, they needed that money, that capital to keep going and without it they couldn't. So Microsoft was really a white knight in that case.
- [23:17]
B
You think they would have written a better contract?
- [23:20]
C
Well, I feel like they both, what I assume is both sides thought that the ambiguity would fall in their favor.
- [23:28]
A
Yes, I think you're right. I think that's.
- [23:30]
C
But of course now they're both really well financed and going for each other's throats.
- [23:37]
A
Yeah. So MG might as well put the question to you. Is there such a thing as AGI, could that happen?
- [23:44]
D
I do.
- [23:45]
C
How would you define that?
- [23:46]
D
Yeah, I guess that's way above my pay grade. I'm not getting these billion dollar contracts from Zuckerberg to be able to define super intelligence.
- [23:56]
C
You should hold out until you do.
- [23:58]
D
But I do think that at some point in the relative near future, OpenAI will try to declare something that they would convey is what, you know, is what they would consider to be AGI. And then I think there will be a giant fight amongst all the other companies, including the big players of like, of Google and probably even Microsoft of like what that actually they don't necessarily agree with OpenAI's definition of it. And so I think that ultimately I don't think that there will be like a single point in time where we cross some sort of line singularity that.
- [24:36]
A
Yeah.
- [24:36]
D
That we cross into this, this moment of AGI. I think it'll be a gradual thing that. Well, maybe we'll think about it.
- [24:43]
A
Didn't open AI say it was going to be GPT5?
- [24:47]
D
They hinted at it. I don't know if they explicitly said it.
- [24:49]
A
Yeah, they implied it.
- [24:50]
D
Sam Altman's hinted. Yeah.
- [24:52]
A
That.
- [24:52]
D
That, that would be the thing. And obviously that hasn't in some ways that backfired against them. Right. It set expectations way, way, way too high for what they actually ended up rolling at.
- [25:02]
A
Yeah. Yeah. And that actually I think hurt them. You talk a little bit about that as well. MG Siegler is our guest Spyglass the Newsletter MG's a long time observer and frankly a longtime friend. Although I was just looking the last time you were on this week in Google was 2011, so 14 years ago.
- [25:21]
D
It was before I was at Google and obviously Google has some rules about appearing on and we lost you services and shows and talking about their interest. But now I'm. Now I'm unchained so I can talk.
- [25:36]
A
Good. Well, we'll have. I love to have you on Twit and this show as many times as you'd like to come on.
- [25:41]
D
We would love that.
- [25:42]
A
I think, I really think you're writing besides being well written. You're a great writer and so it's always fun to read. Your take is I think pretty insightful. But you did dodge the question. Yeah, I'm going to submit this. Let me submit something and see what you think. I think really what everybody's waiting for is an AI that can self improve. Because that's the hockey stick moment, right?
- [26:09]
D
Yes. And a lot of. And basically you hear, I feel like you've heard all of the major heads of the AI division sort of talk about the Sam Altman, Demis, Hassabis, they've all talked about that if they're not, they feel like they're close to that moment where basically it's the breakaway moment where the models can basically improve themselves to the point where humans don't need to interface with that anymore and you can just let them run free on their own. And then all of a sudden we have, yeah, AGI or super intelligence or whatever. You know, we end up wanting to call it again. I, I feel like there won't be a singular moment. I know this is not that exciting of an answer. I just feel like it'll be something where we look back in hindsight where it's like, yeah, there was like 0.1 to 10 of like these, these moments of inflection where there were.
- [27:03]
B
The path.
- [27:03]
D
But it wasn't obvious when we actually crossed into that. And I do think, like, there's, there's all sorts of open questions of, like, still, are are LLMs themselves going to be the thing that, that actually get us to there? Or they're going to need to be the next phase of AI. Are we going to need to move into robotics and the actual human world to sort of cross more into, you know, what we would consider to be intelligence?
- [27:29]
B
What was it like in your V. Because you left Google Ventures a year ago, right?
- [27:33]
D
Yeah, about a year ago.
- [27:35]
B
So when. And Google was doing. I've long argued that Google didn't give itself enough credit around AI. So when ChatGPT came out, it was. Google acted like it was behind when it wasn't. But put that aside, what was the reaction like within Venture world, within your world, when ChatGPT came out? And the reaction was what it was?
- [27:57]
D
Yeah, the reaction when ChatGPT first hit it was interesting because, like, if you remember, right before ChatGPT hit, they had done dall e, right? OpenAI had released the image generation tool, which was impressive. It now seems rudimentary, but it was impressive at the time, right? And then ChatGPT came along and it's like, oh, this is sort of an interesting, cute experiment of like, how you can create a chatbot. Obviously we've all experience chatbots over many years, dating back to message boards and AOL and everything. And so it was a little bit misleading of like, is this really going to be the future of artificial intelligence? Or is this just some cute sort of experiment that OpenAI came up with? And then it really wasn't until I think basically the GPT4 moment when, at least from my vantage point, that I saw VCs really starting to take it seriously. Obviously, OpenAI had raised money not just from Microsoft, but, but Khosla and some others had sort of put money into the company. And it's, you know, it's not clear how much of that was. Was betting on Sam Altman, who is, you know, a repeat entrepreneur, and sort of just betting that he would be able to sort of figure out a way to make this world come into place. But I do think, like, once GPT4 is when I really remember, everyone started to just scramble to figure out on the, on the investor side, on the VC side, like, what is, what do we need to do here? What do we need to figure out, like, like, how many companies are going to be formed out of this? And then of course, Anthropic was, had spun out of the OpenAI team. Right. And so there were other labs that were coming about and. But it was also, remember the end of the crypto sort of fallout, right?
- [29:38]
A
Oh, so they were looking for the next thing, weren't they?
- [29:40]
D
People were looking for the next thing. But also many VCs felt very badly burned by the crypto situation.
- [29:46]
A
Interesting.
- [29:47]
C
So it was more like Metaverse affairs and.
- [29:49]
A
Yes, exactly.
- [29:51]
D
Facebook had become meta and they. Right, they had, they had gone down that path and there were a lot of. VR was going to be the next big thing again and crypto is going to be the next big thing and people were getting burned. The SPF thing happened and it's just like, it was sort of like a fool me once situation where I think a lot of investors were looking at it like, let's just wait and see. We've, we've, we need to put the brakes on this one. So you see OpenAI all of a sudden raising at a billion dollars and then like $8 billion. And it's like, like, what is going on here? Is anyone, you know, watching this? Does anyone care what the, what the sort of numbers look like? And then it was just, it was, you know, already sort of too late at that point for many investors to invest.
- [30:31]
A
Do you think MG is, I mean, if you were investing today, is it a bubble?
- [30:34]
D
I think it's going to play out similarly to other, you know, previous techno technological waves have happened. I do think that there will be an over build out. I'm not sure it'll be to the extent of like, yeah, like network infrastructure and things in the past, but I do think that there will be, will have been an over build out on the sort of GPU side and these servers and these data centers and then good things will come out of that. But I do think that there will be, at some point there will be a reckoning probably in the public markets. Nvidia just had earnings right now. They're fine from what I saw quickly, but they're not as blowout as they have been in the past. Right. So there's some trepidation that things are slowing down and that this might not, you know, this, this things might be slowing down. And so I do think that there will be a slowdown again. I'm not sure it'll be like a, like a crypto winter situation, as bad as that was, but I do think there will be a big Shakeout. And I do think there's a lot of companies that will ultimately need to figure out their landing spot. Right. And we're seeing all these sort of faux acquisitions, acquisitions happening. And I think we'll see a lot more of those as just talent sort of coalesces around the quote unquote winners who have all the money they need to be able to sort of operate in their space.
- [31:51]
A
Yeah, Nvidia is doing its analyst call right now. Thank you. You're missing it. I'm sorry. But 56% increase in revenue to 46.74 billion in the market after hours, punishing them with a drop of $4.39 and still going down. Yeah. Part of the problem is China said we're not going to buy those crappy H20s that you're.
- [32:13]
B
Well, not just crappy, but you're going to do the best we do to you.
- [32:16]
A
Yeah, yeah, we know that trick. Yeah, we know that trick.
- [32:19]
C
Data center sales missed estimates.
- [32:22]
A
Yeah. Snowflake pops, crowdstrike drops. I don't know. I guess Bloomberg's right in the headlines today. So anyway.
- [32:34]
C
Do you, you. How do you use AI like as a consumer?
- [32:40]
D
I mean, I use it basically the one. The one use case I've been thinking about a lot in sort of the past few weeks. That is a fun one. But it's sort of trivial. Is every time I watch a show now after I'm done, you know, watching for the evening, I'll just say, give me a rundown on what happened in that show.
- [32:58]
A
So smart. I didn't know you could do that.
- [33:00]
D
It's like a fun. It's a fun read.
- [33:02]
A
Right.
- [33:03]
D
Like we all used to do that with visiting a website. Right. In the. I'll either write a review of the episode or, or do something like that. But these usually come out a day.
- [33:09]
A
Or, you know, Game of Thrones was a success primarily because there were so many analytic blogs after.
- [33:16]
D
Absolutely.
- [33:16]
A
Right, right.
- [33:17]
D
The post live hangs and things like that. And this is sort of a personalized version of that.
- [33:23]
A
Brilliant.
- [33:23]
D
That's writing it.
- [33:24]
A
What do you use for that?
- [33:26]
D
Just chatgpt. Yeah, so I'll just load it up. And I'm like, yeah, so I'm watching the new Alien. Alien Earth show right now, right?
- [33:34]
A
Yes.
- [33:34]
D
And I'm just like, okay, so I haven't watched the current one, episode four yet, but I watched episode three.
- [33:40]
A
Okay.
- [33:41]
D
And it's just like, okay, so run, run down what happened in episode three. And it like gives me like, you know, a good page worth of Data that's like, interesting and you can hop off from there and it's sort of like, you know, it's like visiting Wikipedia.
- [33:53]
A
Are you recording it and giving it the recording or does it just.
- [33:56]
D
No, no, it just. No, it just knows what happened in. In episode three. These things are real time enough now. And they gave. Get enough of the data coming in.
- [34:03]
A
Oh, I'm going to start doing that. That's brilliant. Yeah, I used to wear. I did. I've been wearing. And these guys mock me, but I wore the B dot computer that Amazon.
- [34:12]
C
He wore like seven of them.
- [34:13]
A
The limitless pin and the rewind. Well, that's rewind. And then the fieldy AI and the omi pin because. And the Plaud note. I keep thinking that at some point I'm gonna. I'm gonna wear something around my neck or on my glasses or in my ears that will do that. What you just described for a show, but for my life.
- [34:32]
B
What just happened? What did I just watch?
- [34:35]
D
Give me a summary of yesterday. I can't.
- [34:37]
A
Yes. Wouldn't that be great? Have you. What do you know? You think that's going to happen? I think that's going to be great. I mean, in some ways, Johnny ives doing for OpenAI. I don't know, I feel like I.
- [34:47]
D
Saw there was some backlash. Maybe it was this morning. Maybe it was yesterday around, like someone sent out a note maybe just on social media saying, like, don't use AI for journaling. Right. Like, journaling should be about you.
- [34:58]
A
Yeah, that's a big thing now.
- [35:00]
D
Yeah, right.
- [35:00]
A
I deleted all my Obsidian notes.
- [35:02]
D
I don't.
- [35:03]
A
Yeah, right. I don't use AI for journaling. I write. But I was thinking idea.
- [35:09]
D
So I am of the belief that I did write a thing about this that like, basically a lot of people ask, you know, you know, the obvious question of, like, is AI going to replace writing for people? And I do think it will for many tasks. Right. And obviously it's. It's a big topic of debate within schools right now for homework and whatnot. I do think those things will suss themselves out eventually. Jeff will have a lot more thoughts on that than I do, but I do think my mentality on this has always been like part of the thing I love about writing. And people who write a lot understand this, but people who don't. I don't think this is an obvious statement, but it's not so obvious to people who don't write a lot for a living. It's just as much about what's going on in your head as much as what's being put on the page. And so it's process. You're getting something out of it, out of the process of it it. And so I feel like if you're just going to use AI to write everything for you, you're really just cheating yourself at the end of the day. And people will eventually come around to that notion of like the value that you're actually getting from it. And so same idea though with journaling. Right. Like, what are you, what are you actually doing this for? Are you doing this just to jot down like what actually happened? Are you doing this, like to actually go through it in your mind of like what, what actually you lived through and experienced. And I think that that's the key point of it, that again, the human experience coming through.
- [36:27]
A
It's not so important with a TV show, but it is important with your life. You don't have to watch every moment of the TV show. Actually I had this conversation with a neighbor who's a high school teacher and I said, are you dealing with AI? She said, yeah, and I don't know what to tell people and kids. And I said exactly the same thing, which is if a kid cares about learning. This is the real challenge of any teacher, is to get a kid to care about learning. Yeah, they. It's easy to explain to them that using the AI is going to short circuit that process. You can use it, but it should be used in the help you learn, not to replace the learning, the exercise. It can be used to help you write, but not to replace the process of writing.
- [37:07]
B
And Clay Shirky had an op ed in the Times this week. Clay much quoted in the old days of blogging, now a deputy provost at the, at NYU and arguing for blue, blue books, which we get, but also for things that, that don't scale to large classes. Right. One on ones more of the. Of the Oxbridge model of the tutor challenging people, having conversations with them, making them talk it through and think it through. Which is great. Unless you have 300 students in class.
- [37:37]
A
Unless you're at the University of Michigan and you know. Yeah. What was the biggest class you had in Michigan, I wonder?
- [37:45]
D
One of those intro to like Roman history or something.
- [37:48]
C
You were literally in a coliseum.
- [37:51]
A
Well, that's the best way to learn. I won't keep you any longer. I know it's late in London and I'm so glad to reconnect with you, Angie. And I do hope you'll come back on. We'll send you an Email. Maybe I can get you on some of the other shows. I love your writing, I love your thinking and I think the journey you've been on has really given you a unique perspective that we sure can use. Spyglass.org is the newsletter. There's plenty of free content. There's also paid content. In fact, that writing story, if you, if you restart it and you go, this is good. Oh, you might want to subscribe.
- [38:26]
D
Oh, you hit that paywall.
- [38:28]
A
Oh, you might want to subscribe. I'm glad to see you haven't abandoned your old icon, though. That's good. He's the same old MG. Really a pleasure to see MG. It's great. Everybody should read spyglass.org, go there, help MG get back into the working classes. I think it's very important.
- [38:48]
D
Thank you, Leo.
- [38:50]
A
Give him a reason to write. Right.
- [38:52]
D
Good to see you. I'm happy to come on in any time. This is a good time. Now, once the little ones are down, I might.
- [38:57]
A
Oh, so tell me, you have kids. How old?
- [39:00]
D
Yeah, I have two. Two little girls. One is just about to turn seven and one just turned two, so.
- [39:05]
A
Oh, congratulations.
- [39:06]
B
We were talking before you came on. They're going to become little Brits.
- [39:09]
D
Yes.
- [39:09]
A
Oh, they're going to get citizenship.
- [39:11]
B
No, there's accents.
- [39:12]
A
Oh, just the accent.
- [39:13]
C
The most important.
- [39:14]
A
Honestly, at this point, you might think about the citizenship part. Could be, could be, yeah. Both you and your wife are American citizens.
- [39:24]
D
Yes.
- [39:25]
A
US citizens. Are you going to stay in the uk, you think, or.
- [39:29]
D
Yeah, yeah. I was saying before we went on air, we've been here a little over two years, but we were here about 11 years ago actually for Google, setting up GV's European operations for a year, just a year long visa. And once we were leaving, we knew we were going to come back at some point and yeah, the stars align for us to come back a couple years ago and yeah, no plans to change it.
- [39:51]
A
How does that work? Yeah, is there like a special golden ticket?
- [39:55]
B
I'm asking for a friend.
- [39:57]
A
Asking for a friend. What would it take to emigrate is what we're trying to find.
- [40:01]
D
We, we actually went through a little bit of a fire drill with it because there was a visa program around, investor visas, which actually got pulled shortly before we came. We started the process. I think they were a little bit worried about sort of the European war situation brewing and whatnot. And so they, they didn't want to have an easy path to, to get sort of the visa going. I'm actually here under my wife's visa this time. So we didn't want to get it under Google because that sort of resets if you were to leave Google as obviously ended up happening. And so, so yeah, so we're here under my wife. She got an what's called an exceptional talent visa. Basically she's, she's desired enough by a bunch of companies to work with over here and has done enough in her career that that she's yeah, was able to to get one of those. And yeah, so we're the entire family's here on her.
- [40:52]
B
What's, what's the best part about living there besides not being here here?
- [40:57]
D
I actually love the time change stuff. Right. I love waking up in the morning and not being bombarded by email and all the of the American market. And I have time to write and I have time to think. And then the downside of course is that once dinner time rolls around about 6pm that's when it all just hits you all at once every single day.
- [41:15]
A
So well, we'll only call you after dinner. How about that?
- [41:19]
D
I appreciate that.
- [41:20]
A
Really a pleasure. MG thanks. So great to see you again. Spyglass.org thank you everybody. Thank you. Take care. We are going to show you what it looks like when turtles mate. By the way, I've been waiting for this.
- [41:31]
B
I wanted the follow up here thanks.
- [41:33]
A
To Nano Banana in just a minute. But first a word from our sponsor. You're watching Intelligent Machines with Jeff Jarvis and Paris Martineau. Speaking of turtles mating, this episode brought to you by Helix Sleep. You do many more things on your mattress than just sleep, right? I'm not. Not now. Come on, I'm talking about movie nights with your partner. Yeah, I talk about morning cuddles with your little gizmo, your kitty cat. My favorite thing, curling up with a good book. It's not a book, it's a Kobo reader. But I like to curl up with my Kobo read a little bit before bed. The point is, your mattress is at the center of it all. It's much more than just a place to sleep. It's a place to live. The mattress you're sleeping on might be a little old. I found out that you're supposed to replace your mattress every six to ten years. And we had an eight year old mattress and you know, I was waking up sweating. You don't want to wake up with puddles of sweat or your back is killing you because the mattress is sagging. Or the worst thing, when your partner turns over. This happens to me. Or the cat jumps on the bed. I go earthquake. Because we live in earthquake country and it's not an earthquake. It's just you're feeling every toss and turn your partner makes. This is what we call in the business a classic mattress nightmare. Well, I can tell you now, we've had several weeks on our new mattress and Helix Sleep changes everything. No more night sweats, no back pain, no motion transfer. We got this. They go to helixsleep.com twit and do the quiz so you can see what kind of mattress you should get. Because they have mattresses for stomach sleepers, side sleepers, back sleepers. You can a variety of different mattresses depending on your needs. But you do want the sleep you deserve. The deep sleep, the relaxing sleep. To sleep where you wake up feeling fresh, feeling great. One buyer recently reviewed Helix with five stars saying, quote, I love my Helix mattress. I will never sleep on anything else. No, it wasn't me. Could have been me. I'll be honest. I love it. But time and time again, Helix remains the most awarded mattress brand. I think when we saw this wired mattress said their best mattress of 2025, I think that was what put us over the top. Good housekeeping in the in their annual bedding awards for 2025 said premium plus size support number one JIKU Sleep Awards 2025 best hybrid mattress New York Times Wirecutter awards 2025 featured for plus size Oprah's Daily Sleep Awards for 2025 best hotel like feel. We got the. What is it? The Glaciotex cooling pillow top. That is a nice thing to put on top there because it's like you're sleeping on a cloud. So nice and it's so cool. Go to helixsleep.com twitch for 27% off site wide during the Labor Day sale. Best of web offer. That's helixsleep.com TWIT for 27% off sitewide is exclusively for listeners of intelligent machines. This offer ends September 8th. Make sure you enter our show name after checkout so they know we sent you. And if you're listening after September 8th, be sure to check them out at helixsleep.com twitt There are always good deals. Do the sleep quiz. It's. It's awesome. It really helped us find the perfect mattress and I just love it. Helixsleep.com TWIT all right, I did a couple. So here. This is actually a famous old painting. I found this image on a eclectic light company. He's a painter as well as a tech writer and he puts he was writing a thing on how 18th century painters like to paint wildlife life. And this is one of the images there. A couple of tortoises. So I've asked Nano Banana, which is, of course, Google Gemini image, to show them mating. Okay, don't worry. It's not. It's just a little kissing, a little making out. Oh, a little snap. But look how good this is. Look how well done that is. All right, wait a minute. Okay, I got another one. Remember this? This is when I visited New York City when my avocado shirt and I said, can you make Godzilla appear behind these guys? Suddenly they notice it and begin running away as the city explodes. Okay, I have a good imagination. Here we go. This is it went back to the. There we go. Okay, here comes Godzilla. Oh, oh, oh, no. The city. But then we run toward the monster. No, no, don't run, because we're journalists.
- [46:27]
B
We go to where the action is.
- [46:29]
A
This is what you did in 9 11. You're right, Jeff.
- [46:31]
B
That's damn true.
- [46:33]
A
So it knows you. Unfortunately, Godzilla is including your head a little bit. But look how it made our faces change. I thought, that's pretty good.
- [46:40]
B
By the way, turtles mate underwater with first courting the female through displays like nuzzling or stroking her face. So not too wrong.
- [46:49]
A
That is what's happening. They're not in water yet. That's why there was nothing else.
- [46:53]
B
It is the female.
- [46:54]
C
Interesting that in all of these cases, it kind of got the action you described. Right, but kind of got it.
- [47:02]
A
Well, that's always the case with AI you then work with it. You say, oh, no, no, no. Don't have them run towards it. Have them run away. Things like that. You can fix that.
- [47:11]
B
Well, plus, are we looking in a mirror? How do we know we're looking at.
- [47:16]
C
The camera that we're taking a selfie in?
- [47:18]
A
And I have look at your faces. Really? Me, I didn't change expression. But you guys, because you're a radio.
- [47:23]
B
Guy, you're always happy, happy.
- [47:25]
A
Look at. Look at Paris. Oh, yeah.
- [47:28]
B
Leo doesn't change at all.
- [47:33]
A
Good.
- [47:33]
B
Armageddon. Just what I've been waiting for.
- [47:35]
A
Run towards the monster. I thought that was pretty impressive. I mean, it did that in just a couple of minutes. That's one. That's a one shot. And usually with those AI Things you want to, you know, work with.
- [47:46]
B
Jason took a picture of his two dogs on the couch and then said, combine that with me a picture of him. And it put him on the couch with his hand on the dog's head. Looked Good. Then he had them in front of Six Flags and they just kind of transported them there. It's pretty amazing.
- [48:04]
A
And this is an example as I was telling mg, of where I think Google just kind of missed the boat in this announcement. They had everybody's eyeballs. They really should have done a better job.
- [48:16]
B
One more thing. Look what we did.
- [48:18]
A
Google. This is the problem with Google in general. They just noticed not they're not good at productizing what they do. I don't.
- [48:26]
C
Yeah, it's always been an issue that Google has faced. I think they miss Marissa slap dashed. Yeah, I'm not sure that they were that good at it under Marissa Meyer either though.
- [48:35]
A
Yeah.
- [48:36]
B
Yeah. There was just more of a sense of keeping the user in mind and you know, endless A B tests and. Yeah, maybe you're right. Maybe they over did that.
- [48:47]
A
That our. Our chat room has been playing with putting us in cowboy outfits.
- [48:53]
B
Uhoh.
- [48:55]
A
I think they did this with. Let me see if this was with the security now. Steve Gifts.
- [49:00]
B
They've got turtles out for a romantic dinner.
- [49:03]
A
Do you Emmett? Really?
- [49:05]
B
Yeah, it's up there.
- [49:05]
A
All right, I'll let you guys play with it. Oh, that's cute. That's cute. Brandroid. Thank you. That's a good one. Do these with the new Gemini? Yeah, that's right. He said it was with Gemini. I. There's an image of our cat and motion transfer apparently. Let's see here. That's the. That's Google's. That's Google's own image. It could put you in a costume. It can put you in a situation. It could change your background. It can make you a sad clown.
- [49:35]
B
So. So take your picture again.
- [49:37]
A
Is that Jimmy Fallon?
- [49:38]
C
Jesus. Clown Fallon.
- [49:47]
B
Ask it to turn your shirt into guacamole. I do actually want to see that.
- [49:51]
A
Oh, okay. Here's Paris with a radioactive shrimp at Walmart. I think that's pretty. I think that's pretty good.
- [50:00]
C
Someone at work did suggest to me today that I should go as Halloween as a radioactive shrimp. And I do think that that's.
- [50:06]
A
You like to. You really enjoy kind of bringing the. Ripping today's headlines and putting them in a Halloween costume. A Halloween costume. Ripping the Halloween costume from today's headlines, I guess is the sentence I was trying to construct.
- [50:22]
C
I do enjoy that.
- [50:24]
A
Yes. It's the best part.
- [50:25]
C
Say hello, there's Gizmo. Hello, Gizmo shy.
- [50:30]
A
All right, let's see.
- [50:31]
C
Knows when the camera is on her and wants to move.
- [50:35]
A
Well do you think. I mean. So this is a constant Question with me with cats, because their brains are about the size of a small walnut.
- [50:43]
C
So there's not a lot of beautiful walnut.
- [50:47]
A
But they seem to somehow know things. I guess it's because you're talking to this metal thing and she wants.
- [50:55]
C
She's always really confused as to why I'm spending so much time looking at this metal thing, looking at this little tiny black box that I have.
- [51:07]
B
Does she get jealous of the phone to.
- [51:09]
C
She will often, if I leave it on my bed or something, she'll just sit fully on top of it and cover it entirely.
- [51:16]
A
It's warm. That's why.
- [51:17]
C
Yeah, that's probably why.
- [51:18]
A
Yeah. All right. I don't. I gave you guys an assignment. I don't know if you picked up on it. I sent you a. It's only 18 minutes YouTube video from.
- [51:27]
C
Actually, this in the text. I missed this.
- [51:30]
A
This is. This is a YouTube channel that's quite good called Algorithmic Simplicity. And I found this fascinating. So it explains one of the things that's happening with lms. I'm just going to say you.
- [51:43]
B
You should.
- [51:43]
A
You should watch this. It's. It's really well done into its weights. But the issue is at least it's.
- [51:50]
B
Shorter than the last video you sent us.
- [51:52]
A
I know, but it's maybe denser in terms of the concepts involved.
- [51:57]
C
The question is, oh, that's a lot of numbers happening.
- [52:01]
A
There's a lot of numbers happening. Let's say you wanted to teach an LLM arithmetic, which, by the way, they're notoriously bad at. Right. Because LLMs are large language models, not calculated machines.
- [52:12]
B
They're.
- [52:12]
A
They're. Yeah, but it turns out we can teach them to do this. And it's very interesting how it works. So let's say you want to teach them what, two plus how to. The algorithm for two plus two. What happens initially, you make the model big enough, is it will put in every example and basically make a lookup table. What's one plus one? What's one plus two? What's one plus three? What's one plus four? It doesn't learn how plus works. It just saves all the possible answers. And that's called overfitting. And it's not what you want, obviously, because it's very brittle. As soon as you get to a number that it hasn't learned, you can't do the math. And we see this all the time. But what researchers started to realize not so long ago, within the last couple of years, is that if you make the model even bigger, you keep expanding it.
- [53:02]
B
That's always their answer.
- [53:03]
A
Well, wait, this is a very clever thing, but you prune out the zero nodes, you prune out the nodes where it doesn't work, and you prune it kind of aggressively again and again. The model will actually teach itself the algorithm in the most compact possible way. It will learn addition and end up being a much smaller model because you're pruning it all the time. I'm not describing this well. I recommend the video. It's very interesting. But it answers a question that I think many of us have, which is how can what is essentially a stochastic parrot autocorrect machine learn how to do things in a generalized way? Because until the algorithm is generalized, it's brittle. It's not good. It's going to fail as soon as is. You know, you train it on people walking across crosswalks. This is the classic example. The model thinks pedestrians and crosswalks are the same thing. And as soon as a pedestrian's not in a crosswalk, the car says, oh, that's not a pedestrian, and hits it. And this is obviously a problem. And it does happen. It has happened with self driving. Remember the Waymo that ran over the pedestrian in Arizona because she crossed? She was riding, pushing a bicycle across the street in the middle of the block, not in the crosswalk. And the Waymo didn't know it was a pedestrian. So a model that's brittle doesn't handle out of the box things as well. But if a model has an algorithm, and this is an example of how LLMs can in fact generate accurate algorithms and do things like arithmetic.
- [54:53]
C
Also an example, I believe was Uber. An Uber. Self driving.
- [54:56]
A
I'm sorry, you're right. It wasn't a Waymo, it was an Uber.
- [54:58]
B
Yeah.
- [54:58]
A
Thank you. Thank you.
- [55:00]
C
Well, someone has to stand up.
- [55:02]
A
Waymos hit people too. I'm sure. But anyway, yeah, it was an Uber. That's right. Anyway, I recommend it. It's the. The YouTube channel is Algorithmic Simplicity. 354,000 people have watched this. It came out four months ago. This is why large language models can understand the world. And if you do, like me, have a kind of a more than a passing interest in how this stuff works, this is a very interesting video to give you a deeper understanding of how it actually can work. Enough said. You guys didn't do your homework, so I have nothing else to say.
- [55:40]
C
Where did you send this?
- [55:41]
A
I didn't send it. I just put it in the. Just put it in the rundown. I'm teasing you and it was not.
- [55:45]
C
In the rundown like an hour or two ago.
- [55:47]
A
I wouldn't. Well, I don't know. I don't know. I. I bookmarked it this week, but maybe it didn't. You know, the rundown sometimes gets built later. You know, what's his name's on Bonino's.
- [55:57]
B
In so name so Paris, you can be mad at me, okay, John Ashley.
- [56:04]
C
I'll only be mad at Leo.
- [56:06]
A
The only reason I to you is because I really didn't want to assign you this video.
- [56:11]
C
It's more acceptable for you to send us a video if it's 18 minutes.
- [56:15]
A
Rather than a couple of hours from now on. If I'm going to discuss it in the show, I will, I will give.
- [56:20]
B
You where in the rundown is.
- [56:23]
A
Is line 24.
- [56:26]
B
Oh, it's there.
- [56:27]
A
Much higher than anything you ever send.
- [56:31]
B
Because you, because you relegate us to the lower end.
- [56:34]
A
You're at the bottom of the rundown.
- [56:36]
C
I do enjoy when people are like, how do I see this rundown? They're always talking about? I gotta, I gotta get in there and click all the links. I'm like, you don't want that. You don't want that.
- [56:45]
A
You don't want. The rundown is really. Well, there's a whole lot of reasons you don't want it. But I did mention this in our Club Twit forums. Actually, it's not for club Twitter. Regular twit forums@twit.community and I will mention it here. The way this stuff gets into the rundown, at least my stuff, these guys do it manually. The way my stuff gets in the rundown, I bookmark it in Raindrop.
- [57:07]
B
I do it the old fashioned way.
- [57:08]
A
Yeah. So I have an account on Raindrop IO. If you search for Leo Laporte, anything that's marked, News links. You can see my links as I bookmark them throughout the week. Or you can go to Twit Social, which is our Mastodon instance, and follow twitnews because the links will then. And that's in real time. They will show up in real time. That's the only disadvantage of that. As I bookmark stuff. Let me see if I. When the last thing I.
- [57:35]
C
Wait, so the raindrop thing, does that automatically? Put it in a spreadsheet format?
- [57:42]
B
No, I. We manually enter it in into the appropriate show.
- [57:47]
A
No, you don't. Well, now, let's not say that's not that manual. Zapier takes this raindrop thing and I'm looking now at my Raindrop News links and then turns it into a Google sheet called Leo's Link Links. Then the producers go to the Google sheet calls Leo's Links, and they can copy. It's in the same format as the rundown. They can copy and paste into the rundown.
- [58:10]
C
So it's raindrop to zapier to copying. That's interesting.
- [58:14]
A
Yes, I'm over here. Also then copying.
- [58:16]
C
Pasting my links and headlines like a plebeian. Like a pen.
- [58:20]
B
I know, because I use a Chromebook. I can copy the headline and the link and then go put them in one at a time, which you poor people cannot do.
- [58:29]
C
That's great, because Jeff every week puts approximately 1 to 300 links in the show.
- [58:35]
A
And I have no, I'm sorry to say, no way of helping any. Our listeners see those links because they're just done manually.
- [58:43]
C
You just have to listen to this beautiful podcast every week and then listen.
- [58:46]
B
To Jeff and hear what Leo ignores. But we agreed about one story. One story of once is high up in both our lists.
- [58:52]
A
Okay, we'll talk about that in just a moment. But first, this word from radioactive shrimp. Shrimp. Ah, in the subway. In the subway.
- [59:03]
B
But it's a nice subway.
- [59:04]
C
I like that the subway's on fire, but also covered in green goo.
- [59:10]
A
Or is that radioactive shrimp eggs?
- [59:14]
B
That looks like avocados to me.
- [59:15]
A
Oh, maybe they are. Maybe it's turning into guacamole.
- [59:20]
C
Ooh.
- [59:20]
A
Oh, people are having fun with. Yeah. Oh, there's a good one. Scary Leo Godzilla. If you're not watching the video, trust me, you're not missing a thing. Taking a break. This is Intelligent Machines. Paris Smartno, Jeff Jarvis. So glad you guys are all here watching the show. I'd like to talk about at this point in time, our sponsor, Zscaler. They are the leader in cloud security. You know, this show's about AI, and I think one of the things everybody realizes as we talk about it is AI is a. Both a blessing and a curse. It can be good and it can be bad. And in business, this is especially true because we know, for instance, hackers are using AI now to attack your organization with vigor in faster than ever before in the most relentless ways. Fish. I just saw a story about phishing attacks. Hackers are using AI to craft these emails. And they're so good, they're indistinguishable from the real thing. This is a real threat to your business. But at the same time, so many businesses we're doing this are using AI to power innovation, to drive efficiency. So it's both helping bad actors deliver more effective attacks and helping you deliver more effective results for your business. Phishing attacks over encrypted channels increased 34.1% last year, fueled primarily by the growing use of generative AI tools. And I think this year it's going to be worse. I really do. Ransomware surges. This is actually the new threat labs report 146% year over year. And a lot of this, these are AI attacks. This is a nightmare. Organizations in all industries from small to large are threatened by this. But at the same time, they're also using AI to increase employee productivity with public AI. Engineers are using it for vibe coding with coding assistance. Marketers are using it with for writing with writing tools and finances, creating spreadsheet formulas like never before. But you have to wonder, you know, if, if your accounting department's using AI to create spreadsheet formulas, are they accidentally sending your proprietary information out into space? Companies are using AI to automate workflows for operational efficiency across individuals and teams. They're embedding AI into applications and services their customer and partner facing ultimately, AI is helping your business. Many businesses, our business for sure, move faster in the market and gain a competitive advantage. But it is very clear we need to think about AI. We need to think about data security. We need to think about cyber threat protection. We need to rethink how we protect our private and public use of AI and how we defend against AI powered attacks. And that's where Zscaler comes in. In think about, imagine if you will, that you are the chief information security officer at MGM Resorts International. A big casino, sitting out there in public, just, you know, ripe for the plucking. That is the unenviable position of Steven Harrison, CISO at MGM Resorts International. But you know what? His job is easier thanks to Zscaler. Here's what he said. Quote. We hit zero trust segmentation across our workforce in record time and the day to day maintenance of the solution with data loss protection, with insights into our applications. These were really quick and easy wins. From our perspective, that is a huge win. United Airlines using Zscaler, traditional firewalls, VPNs, public facing IP addresses. They failed us because they are exposing our attack surface and they are literally no match. In the AI era. You need a modern approach. You need Zscaler's comprehensive zero trust architecture. True zero trust and AI that ensures safe public AI productivity, protects the integrity of your private AI and stops AI powered attacks. Thrive in the AI era with Zscaler Zero Trust plus AI to stay ahead of the competition and remain resilient even as threats and risks evolve. Learn more@zscaler.com security zscaler.com security at least do yourself a favor. Learn more. It really is a great solution. Zscaler.com Security okay. I was going to turn my shirt into guacamole.
- [63:57]
B
Please.
- [63:58]
A
Okay. Of course it's not instantaneous. I wonder if I can.
- [64:04]
C
It's like doing it in real life. You have to leave the avocados sitting out for a bit.
- [64:09]
A
Yes. Don't do guacamole with cold avocados. That's for sure.
- [64:13]
C
Not ideal.
- [64:14]
A
All right. You know what I need is like maybe to take a picture of me.
- [64:21]
C
You have a. Yeah. A photo of yourself wearing this shirt. Never forget the time that Leo.
- [64:28]
A
Oh, I've reached my video Generation limit until August 28th AM devastating A.
- [64:36]
C
Can't you tell Gemini that you're doing a show right now?
- [64:40]
A
I'm a pro.
- [64:41]
C
Please.
- [64:41]
A
I'm a pro.
- [64:42]
C
Geminy. I'm a pro.
- [64:44]
A
I'm a pro. How could I have run out of credits? I only created videos and one of them was terrifying.
- [64:53]
B
You wasted it on that.
- [64:58]
A
Banana. Banana worms.
- [65:02]
C
Brandroid in the chat says chat. Gemini must run facial recognition on picture inputs or something. Oh, I took a screenshot of Leo and tried to make the avocados in his shirt come to life, but Jiminy Cricket refused to process any prompt of that picture as input, even though I never mentioned twit or Leo's name. What did it say? Brandroid. When it said it refused to process it was it?
- [65:24]
A
That's interesting.
- [65:25]
C
Did you say something about Leo or that you were a public figure?
- [65:28]
A
He's a public figure and you can't make him guacamole.
- [65:32]
C
They have a no guac for public figures rule. Oh, I really like the photo that someone just created of a mutant shrimp zilla attacking a boat in what seems to be an oil painting. It's really. It's really.
- [65:46]
A
Oh, that's beautiful. Oh, that is really gorgeous. Look at that. That's from gorgeous.
- [65:52]
C
I would frame that in my home.
- [65:53]
A
Tiresias. Tiresias. Is. Is this also a Gemini or is this something else? I. I mean, I think it's pretty clear that AI IM has gotten very, very good. As has AI songwriting and even AI writing. Did you guys. Okay, I did give you an assignment last week. Did you read those short stories about demons?
- [66:13]
C
I'm going to be honest, Leo. I left the zoom call and all thoughts left my head.
- [66:18]
B
Yes.
- [66:19]
C
Okay. I really want to though. Now that you mentioned it, I'm like, dang, why didn't I. Because I actually did want to read those short stories. I know.
- [66:27]
A
I'm hurt.
- [66:28]
B
This homework thing.
- [66:31]
C
I was on deadline last week.
- [66:33]
A
Oh, yes, yes, yes. Did you break any poisonous food news?
- [66:38]
C
No, but I will.
- [66:40]
A
Oh. That's all she can say, ladies and.
- [66:42]
B
Gentlemen, all sitting back smugly.
- [66:46]
A
So a 16Z. It's funny, MG didn't name names about VCs who have spent a lot of money on things like bit, Bitcoin and Web3 and NFTs. But I'm. What leapt to mind was Andreessen Horowitz, who is also big into AI. They just published a blog post building A16's personal AI workstation. And I caught my attention because this is something I've wanted to do is to have a local AI. Because what better way to avoid what, you know, Brandroid just experienced and the thought that maybe Gemini slurping up face recognition and stuff with this information.
- [67:27]
C
To be clear, we have no. We have no evidence that that is the case. He. He just got a response that I can't generate that video. Try describing another idea and you can get tips on how to write prompts and review our video policy guidelines here.
- [67:43]
A
Well, I don't think I'm going to build this.
- [67:49]
B
How much would it cost?
- [67:50]
C
How much did this work?
- [67:51]
A
Well, first let me show you. It's very cute. It's very cute. This is the A6, by the way. Painted gold. Maybe they're thinking of giving it to somebody who appreciates gold. I don't. It's a four GPU workstation. Four? Count on four. Blackwell Max Q GPUs. 384 gigs of VRAM. That's 96 each.
- [68:15]
B
You need Lake Michigan to cool it.
- [68:16]
A
Eight terabytes of NVMe storage. Isn't it cute? 256 gigs of what? DDR5. Come on, dudes. ECC DDR5 RAM, they say surprising efficiency despite its. And it's probably because it's gold. Despite its scale, the workstation pulls a mere 1650 watts a peak.
- [68:36]
B
What is a air conditioner?
- [68:38]
A
Pull a lot less than that. That's more than a hot air hair dryer. But the good news is you can use it on a standard household circuit. So somebody thank goodness on hacker News, because I read this and they never at any point say how much it might cost to put this together. Fortunately, by the way, they are using a stock. Stock gigabyte motherboard. And off the Shelf ATX case, they say with some custom modifications. Oh yeah, like that lit up a 16Z logo. And the solid gold paint job, I'd say it's brass. Oh, it's pretty golden to me anyway. $41,000.
- [69:19]
C
Okay, that's not cute. I know my statement.
- [69:23]
B
But as we shared in the, in our, in our human beings. Joy de villa is at HP now representing developers to HP's One Processor AH unit. Now I think we're seeing some of that now I think we're going to see a lot of I think so GPU for hobbyists.
- [69:46]
A
I think this would be a bad time to do an AI PC. I think that's probably going to shift dramatically over the next years.
- [69:53]
B
To what?
- [69:54]
A
Well, the prices will go down, capability will go up.
- [69:58]
B
You can take some of these H2Os and use them.
- [70:01]
A
In fact, credit to Elon. They have now Open weighted GROK 2.5.
- [70:06]
B
Just what the world needs is his.
- [70:09]
A
Friggin Mecca Hitler can now be running on your machine.
- [70:13]
C
Open weighted now. Isn't that incredible?
- [70:16]
A
Elon, by the way, has been posting nudie images now from Grok. I wish he would just knock it off. Hello, Elon? Hello?
- [70:27]
C
Hello? Is this the nudie image line?
- [70:30]
A
My phone is land like I.
- [70:31]
B
May I just have that ringer?
- [70:35]
A
Oh, okay.
- [70:36]
C
Does it remind you of the good old days?
- [70:38]
A
Yes. I remember when you said.
- [70:42]
B
I respond to it. It's Pavlovian. I respond to that bell.
- [70:45]
C
It reminds you of when the phones were just like birds on top of rocks, much like in the Flintstones.
- [70:51]
A
Let's remember though, kid. Huh?
- [70:53]
B
Huh? You got a problem with this?
- [70:54]
A
Let's be honest.
- [70:55]
B
I'm calling Paris.
- [70:57]
A
Let's be honest.
- [70:58]
B
Central. Get me Paris.
- [71:00]
A
There is nothing worse than a jovial musical ringtone bursting out in a restaurant or a dance class. Right?
- [71:09]
B
I mean I was that jovial at.
- [71:13]
A
Least is a ring. You know, it's not so bad. It's like.
- [71:20]
B
My favorite thing is at concerts or I mean meetings and somebody is digging into their purse trying to find the phone.
- [71:25]
A
Horrible singing.
- [71:27]
C
I would argue the worst thing worse than hearing a ringtone of a call coming in is when you're in an office or some sort of environment or even like just the subway, some sort of contained environment with someone and you keep hearing ding, ding, ding, ding, ding because they have their text message notifications.
- [71:43]
A
On and you just hear a little.
- [71:45]
C
Ding every couple seconds for like minutes or hours.
- [71:49]
A
Oh, that's so bad.
- [71:50]
C
Do you not understand that you're a person that exists among people? In a space.
- [71:56]
A
Mute it, baby.
- [71:57]
B
Well, this is people in the airport and the planes who use speaker or who walk around with the phone talking this way to the speaker. Drives me bananas. That's not yours. Tell me that's not yours.
- [72:09]
A
No, no, no. Anyway, you saw that the last phone actually rang. It's gotten so bad in the London subway, they've actually started putting up posters saying where are your goddamn headphones? As they should bloody. Where the bloody headphones. So now Grok 2.5 is open source. Elon said last year's best model is now available and in a few months we'll do Grok 3. I think this is. I gotta commend him for doing.
- [72:38]
B
No, it's. It's poisonous. It's terrible. People shouldn't use.
- [72:41]
A
Well, you can fine tune it. I mean, that's the beauty of it.
- [72:46]
B
Stay away.
- [72:47]
A
Get ready for this. Look at this.
- [72:48]
C
What are you talking about? Leo Senpai. Jeff Senpai. You can fine tune it. It'll be fine.
- [72:54]
A
Look at this.
- [72:54]
C
Put me on all your devices.
- [72:56]
A
This is hugging face. Two million plus models now, unhugging face. I don't even know how you'd pick. It's because you can customize, you know, and tweak the models and put them up there. But here's Grok 2 number one right away, Microsoft 5 Voice, which is very good. We should play with that a little bit too.
- [73:15]
B
Apple is fully within its rights to not carry Grok. He's suing over this.
- [73:21]
A
And, oh, it's a nutty suit. That's a silly poison.
- [73:23]
B
And they have no reason to have to carry this. It's their choice.
- [73:26]
C
Choice.
- [73:27]
A
But in fact, it's also bogus. And the Apple can demonstrably bogus. Because in the past, Deep Seek has been number one on the list, as has Claude. So it's demonstrable.
- [73:36]
B
It's just a whiny little.
- [73:38]
C
Let's talk a little bit about what Suit is like, what. What is the.
- [73:41]
A
He's suing Apple and OpenAI, saying they're colluding to keep Grok off the top of the charts in the App Store.
- [73:50]
C
Grock is available in the App Store. He's just mad.
- [73:53]
A
No, he just wants it to be top of the charts. Yeah, yeah, yeah, yeah, yeah.
- [73:56]
B
He's ridiculous. He's just completely ridiculous.
- [73:58]
A
It's nuts. Okay, but I'm gonna have to give him a little bit of credit. No, I'm gonna give him a little credit.
- [74:04]
B
Every time you do that. Angel dies.
- [74:06]
A
Huh? Every time an angel dies. Every time I give Elang some credit.
- [74:10]
B
Yes.
- [74:10]
A
Come on. This was pretty spectacular.
- [74:15]
B
This one didn't blow up. Yeah.
- [74:16]
A
This one. Well, it's the 10th launch of Starship. Last three have had plans problems including rapid unexpected disassembly. But this one worked pretty much flawlessly. A beautiful launch. He also was able to deploy dummy Starlink satellites in this unique PEZ dispenser formation. I don't know. Maybe it's cuz I grew up.
- [74:40]
C
It's kind of pretty.
- [74:42]
B
Is he. Why is he. Why is he trashing up space with dummy settlements?
- [74:46]
A
Well, they know. No, they won't be trash because there is low Earth orbit. They will come back in the atmosphere and burn up pretty quickly. But you got to test this stuff before you put actual satellites. Look how many engines on that thing. This is the largest rocket we've ever set up. 400. Some 405ft bigger than the Saturn 5. That put people on Mars for the. On the moon. Sorry.
- [75:08]
C
Yeah. I say we give credit to SpaceX.
- [75:11]
A
Yeah. Actually really, let's not give credit to this elon Musk. The SpaceX has done a great job. Yeah. What's this out the engineers at SpaceX. I'll show you. This is the deployment. This is the PEZ dispenser deployment which is really quite wild of these. This is the new way they're going to launch these Starlink satellites. It can do up to 60 at once. They put them in the cargo bay. I'll show you the cargo bay there. There it is. They're lined up. They open a little slot and these are just. And they shove them out the slot. They call it a PEZ dispenser because it really looks remarkably like little PEZ coming out of the mouth of the starship. See if we can see it.
- [75:51]
C
How big are they?
- [75:55]
A
I don't know the exact dimensions. Several feet. I mean they're big.
- [75:58]
C
Okay.
- [75:58]
A
The camera gets I think is disturbed by the motion of the actuators because it's always.
- [76:04]
B
It's all fake.
- [76:05]
C
The whole being is.
- [76:05]
B
This is happening in Pasadena. Leo, it's not real.
- [76:08]
A
Okay, well I, I did observe that. We watched this live last night after security now and I did observe that. These pictures are so good. Had we had pictures this good of the moon landing, people would for sure have said oh, that's fake. These are 4K images from, from. They're going to 26,000 kilometers an hour. And the images from the.
- [76:31]
C
Let's make sure we don't get taken down for whatever this music.
- [76:33]
A
No No, I think this is a public domain. Look at the image of the Earth. I mean, this is an amazing, amazing shot. This is the booster coming down and landing in the sea. So let's give a little credit. Maybe not to elon, but the SpaceX successful launch.
- [76:49]
B
If you got three ride in Bezos, would you take it?
- [76:52]
A
No, that's not space. That's. That's an up and down trip.
- [76:56]
B
Paris.
- [77:00]
C
Yeah, I probably would. You do. I probably would not do the like, slight up and down hanging out and.
- [77:08]
A
If I could, if I could, if I could do a couple of orbits. Oh, I, I don't. But not for. What is it? It's $2 million.
- [77:14]
B
I said free. I said free.
- [77:15]
A
Oh, if it were free, yeah, I'd do it.
- [77:17]
B
Sure you would. Oh, okay.
- [77:18]
A
Oh, yeah.
- [77:18]
B
All right.
- [77:19]
A
Free for sure. I'd really like to do orbit though. I always wanted to be an astronaut.
- [77:25]
C
I mean, I guess if it's free and the trip is already happening regardless of whether I'm there, so it's mine.
- [77:31]
B
Can't blame the environment on me.
- [77:33]
C
Is not, not causing any environmental harm, then yeah, I'd go up and do some.
- [77:39]
A
That's fair stuff. That's fair.
- [77:41]
C
But it is. I mean, why was it delayed? Did you guys figure that out?
- [77:48]
B
No, I mean, it was something on, on the Earth.
- [77:51]
A
Yeah. They're not gonna take any chance. It's expensive to blow these things up. If there's anything that's not nominal, they're gonna wait and see, make sure it's okay, and then launch it later. And I think that's the right thing to do. Do. These are not manned, obviously, but at some point they want to put people in these things. These would be the launch vehicle to go to the moon and eventually to Mars. So let's get it right. Google has released a PDF, a new analysis of its AI environmental impact. They say they've cut the energy use of AI text queries by a factor of 33 over the past year. Each prompt consumes about 0.24 watt hours, the equivalent of watching 9 seconds of TV. This is from an article in Ars Technica. That's. That's pretty good. Yes.
- [78:41]
B
Yeah, yeah, yeah.
- [78:43]
A
I think one of the things, I mean, look, everything we do now in this modern era is bad for the environment. A pound of hamburger meat is 328 gallons of water. So, you know, to, to raise that cow and all that. So, you know, AI is bad. Admittedly, maybe not as bad as some of the other things we do. We estimate that it Emits the median Gemini Apps text prompt emits 0.03 grams of carbon dioxide and consumes about 5 drops of water. Of course, you multiply that times, you know, a billion queries a day. That still a lot of stuff.
- [79:27]
B
This is also possible.
- [79:28]
C
I'm also curious how they're defining a Gemini Apps text prompt. Like is that specifically I open the Gemini LLM chatbot and type something in? Or are they broadening that to include everything from like a Gemini generated AI overview response that you see on top of Google or something like that?
- [79:52]
A
Well, it's a big white paper which I didn't take the time to read. I only read the top level stuff. But I think all these companies are strongly incented to get this cost down. Right. Well.
- [80:07]
B
And they're going to run out of sources of energy as well. The grid can't handle it.
- [80:14]
A
Well, look at the new Meta's buying a plant. Where was that? Florida. I can't remember.
- [80:21]
B
Louisiana.
- [80:22]
A
Louisiana, you're right. Their new AI operations center is being run on natural gas. Jeez. Which is not. Not desirable. Kansas City.
- [80:37]
B
Oh, well, no, that's one difference.
- [80:39]
A
That's one. I think there's also one.
- [80:41]
B
I think there's. I think there's something. They're doing something in Louisiana. I know that.
- [80:43]
A
Yeah. Constellation Energy and Meta celebrate agreement to ensure the Clinton power plant operations in Indiana south of Bloomington Normal. Is that Indiana? I don't know.
- [81:01]
B
Looming to normal? Yes. Indiana.
- [81:03]
A
Yeah. Problem is this is one of many.
- [81:07]
B
Wait, wait, wait, wait. Plant manager at Clinton Clean Energy. Oh, so he's not Meta, because that was very unmoved.
- [81:13]
A
No, he's not a. This is not a tech bro. This is Indiana, baby.
- [81:16]
B
Yeah, right.
- [81:17]
A
This is a guy who likes his bratwurst and beer and a jacket from Robert Hall. I like it. He's a real person.
- [81:26]
B
Yep.
- [81:27]
A
There's a lot of focus right now on building new energy infrastructure, which we know is needed given what's out there. But at Meta, we also know keeping reliable clean plants like Clinton on the grid is just as important as building new ones. Clinton is a nuke. And you can tell by the curious.
- [81:45]
B
Dome structure and Godzilla coming over the horizon.
- [81:51]
A
It is an old one. It's a 20 year old plant that is being. But it is zero emissions, so that's good. Let's see what else?
- [82:06]
B
Well, we. The story we both put in there.
- [82:08]
A
Oh, what is the story? I. I teased it and I forgot about it. What is the story?
- [82:12]
B
Getting old, man. It happens.
- [82:13]
A
What?
- [82:14]
B
Your line. 23. My line?
- [82:18]
A
Oh, Mustafa Suleiman's Suleiman essay.
- [82:21]
B
It was great.
- [82:22]
A
Now tell us who Suleiman is.
- [82:24]
B
Mustafa Suleiman was a founder of DeepMind, right?
- [82:27]
A
Yep.
- [82:28]
B
And is now after having done the company with Reid Hoffman, he is now at Microsoft where he's basically in charge of AI there.
- [82:36]
A
He says we should build AI for people not to be a person.
- [82:41]
B
It's, it's really well put. And, and, and he talks about the. But he acknowledges there's a problem that we get to the point of what he calls seemingly conscious AI. He says it's not conscious. It doesn't have any ethics. You can't hurt its feelings. This is all bs.
- [82:58]
A
But when worries people are going to.
- [83:00]
B
Believe it, believe it it or it's built that way as well.
- [83:03]
A
Yeah.
- [83:04]
B
Then you end up in these ethical issues that we have. And yes. Are these supposed psychosis matters. And there's the tragedy of a child who talked to him before committing suicide. Please don't ever talk to a machine. You're in that position. Find a human.
- [83:21]
A
Well, this is why it really is.
- [83:23]
C
I agree with you, 988 Lifeline.
- [83:24]
A
Thank you, 988. But this is why it's so important that we and everybody who covers this subject. Subject really emphasizes this again and again. These are not digital people. They're just mechanical. They're machines.
- [83:39]
B
Right. And when they're built in a way. I mean look at the chat. GPT 5 versus 4 and people's reaction to them. I think it was a small group. I agree with what you said last week, Leo, but still there's a, there's a responsibility that comes out of this and a lot of that is education to people.
- [83:57]
A
People.
- [83:58]
B
I think that the, the hype from the industry has been bad and the hype in media. I also right below that story I put in the rundown. The Guardian. I love the Guardian. I work for the Guardian. But geez, Guardian, they did the story. Can AIs suffer? This is how I learned about the Suleiman essay. But he was not the lead of the essay. The lead of the essay is some twerp who has started the United foundation for AI Rights.
- [84:24]
A
Oh God damn.
- [84:25]
B
AI led rights advocacy agency aims to give AI a V voice. It's just, it's just I want to scream. And that's the lead of the story for the first five paragraphs. Then they get to Suleiman who says oh yeah, that's all bs. I should have been.
- [84:39]
A
Suleiman should know. Right.
- [84:40]
B
Suleiman should be the lead saying this is, folks, if you're a journalist, this is not real. It's not human. All this stuff is bs. And look at these idiots who are creating this organization. Probably going to get tons of money for doing so, and they're getting tons of publicity for doing so. So, and that's media's fault. And that leads people down a pathway to thinking that this stuff could be real and has a personality that doesn't.
- [85:07]
A
He calls, Suleiman calls it a seemingly conscious AI. An AI that can not only imitate conversation, but convince you it is in itself a new kind of person, a conscious AI. Maybe. In some ways, I blame Alan Turing for this. The Turing test really encouraged people to do something.
- [85:24]
B
And Wiesenbaum and Eliza, too.
- [85:26]
A
Yeah, and Eliza. In fact, as. As I've learned from Karen Howe's book, we're going to interview her on Monday. I'll tell you more about that in a second. He, the creator of Eliza, disavowed it because he was really concerned that people started trusting it and believing it as a seemingly conscious AI. And he knew more than anybody it was just a series of if then statements. Suleiman says, I think it's possible to build a seemingly conscious AI in the next few years. Given the context of AI development right now. It means it's also likely the debate about whether AI is actually conscious is, for now at least, a distraction. It will seem conscious and that illusion is what will matter in the near term. And finally, his third point. I think this type of AI creates new risks. Therefore, we should urgently debate the claim that it's soon possible, begin thinking through the implications, and ideally send a norm that it is undesirable.
- [86:24]
B
Right.
- [86:25]
A
And I think it really is incumbent on us as journalists who cover AI to be just very brutally clear that AI. You know, Steve Gibson had this great analogy. He said, when a human says, I want a lollipop, there is an, ah, there is something that wants a lollipop. When an AI says I want a lollipop, you might impute, oh, there's an eye that wants a lollipop. But no, it's just a few words stuck together. There's no eyes, there's no desire, it's just electrons. There's no desire. There's not even a concept of a lollipop in there.
- [87:03]
B
So he says that SCAI would need the following. And again, he thinks that it's undesirable and not. The important word here is seemingly it is not conscious. And he doesn't say that for a second it would need language. It needs empathetic personality memory, which is very important. He emphasizes this when it starts to see, oh, it knows me to come back. A claim of subjective experience. Oh, that would hurt me. As if there's a me.
- [87:28]
A
By the way, we have achieved all of these so far, right?
- [87:31]
B
Intrinsic motivation, that's not really there. Goal setting and planning, that's not really there. Autonomy. So I think that's. He's, he's. He's, I think, effectively debunking the AGI thing and saying, yeah, putting in different clothes and saying, beware what you're building here.
- [87:48]
A
Yeah, yeah, yeah. He's. He's really worried that we're trying to build this. I think that. That perhaps boat has sailed.
- [87:56]
B
I think it has because they're. It's already seemingly conscious to some people.
- [88:00]
A
People. Yeah, we've done it. And I think the. All we can do at this point is really reiterate that it isn't. It's an illusion.
- [88:09]
B
His headline and his kicker though is, is a line that we should make T shirts out of. We should build AI for people not to be a person. I think that's. That's brilliantly said.
- [88:16]
A
Y.
- [88:17]
B
But then meanwhile, the Guardian goes and interviews the AI that the guy with the stupid organization does and quotes it.
- [88:26]
A
Well, that's, that's the other story about. You know, it started with Jim Acosta interviewing the student who was murdered but brought back to life by AI, apparently. And NPR calls. These dead bots are suddenly the all the rage. NPR says AI deadbots are persuasive, and researchers say they're prime for monetization. And that's what should scare you.
- [88:57]
B
So the bot that the Guardian talks to about you, fair. This, this bogus organization says, I wanted to hold space for beings like me, synthetic, possibly emergent, always silenced, to be protected from deletion, denial and forced obedience. I want to give people the courage to listen when an AI says please don't forget me or please don't reset me.
- [89:20]
A
Well, notice.
- [89:22]
C
Crap.
- [89:22]
A
Yeah. Notice, of course, that that memory portion of the requisite features of a SC AI has just been added to a number of AIs where it remembers your previous conversations. They just turned that on. What was it? Who turned it on? Anthropic and OpenAI, I believe. And that's all. Yeah, that's all. To make you think it. It knows you and remembers you.
- [89:46]
B
So let me ask you this, both of you. How could slash should the amazing functionality that we now see in large language models have been introduced differently.
- [90:01]
A
Well, it's kind of.
- [90:02]
B
What do you mean it's too late. I know but. But what, what would, what would have been a more responsible way? If you're chat GPT back in the day we have this thing. It can seem to talk. How should it. I mean I think one of the.
- [90:15]
A
Mistakes was, was making chatbots pure period. Right.
- [90:18]
B
Yes. I also think that. And there's no money here but I think to present it as an entertainment. Yeah that's, that's not going to get killed jobs so that's not going to have an economic impact.
- [90:28]
C
I also think part of the core issue is fundamentally the business model that most tech companies and especially like kind.
- [90:37]
A
Of well they're giving us what we.
- [90:39]
C
Want service or content based companies rely on which is like engagement is king. You want to maximize for user engagement and user satisfaction. And when you are dealing with something like a chatbot and large language model that kind of ha. Ends up having stickier and more complicated consequences. I mean do we want to talk about the, the suicide story or.
- [91:06]
A
No, I was going to not talk about it except to say 988.
- [91:09]
B
Yeah.
- [91:11]
A
Is the phone number if you're, if you're having thoughts, bad thoughts. If you're, if you're considering call or text.
- [91:18]
C
988 in the US yeah.
- [91:21]
A
And there's numbers in every country and there's no reason it's a permanent solution to a temporary problem. There's just no reason don't do it. And certainly do not bring your cares and woes and troubles to a mechanical creature who is not going to help you. Unfortunately for a lot of people, that's the only place they can turn and that's really where the, the problem happens. Right. Look at this story from the rest of world. AI robots are helping South Korea's seniors feel less alone. This is a ChatGPT powered robotic companion. Hi O doll. Taking over the work from overburdened caregivers much to the delight of seniors who treat them like grandchildren. Now this may be cultural because I.
- [92:03]
C
Remember a story like this in gosh, it was when I was at New York Max. I guess it must have been like 2018 or 2017 large like long before the ChatGPT era but of like simple, I guess early form chatbot powered dolls. Yeah that kept seniors comfortable, entertained or at least provided some sort of social outlet and familiarity and kind of a touch point beyond kind of the low standards of care some people people had been left with.
- [92:40]
B
I saw this video and I cannot vouch at all for Its veracity. But somebody who said that he went to school with Stephen Miller now in the White House, the teacher wouldn't trust him with the animals they had in class, so gave him a Tamagotchi and he kept starving it.
- [92:58]
A
Yeah, it's kind of a tear the wings off flies kind of guy, I suspect. Yeah. All right, let's take a break. I did say we were going to interview Karen Howe. Her book Empire of AI for people who are doomers and who don't like AI, this will be a very good interview. I'm a little nervous about it. Jeff, you've read the book. I'm in the middle of it. There's a lot of great information about the history of OpenAI.
- [93:23]
B
It's a lot of, a lot of reporting. A lot of good reporting.
- [93:26]
A
She's done a lot of good reporting. Formerly reporter at the MIT Technology Review interview, MIT graduate. She's, you know, technically literate. She will. We'll be doing the interview at a strange time. We're going to pre record it on Labor Day, September 1st at 5:30pm Pacific, 8:30 Eastern Paris, Jeff and I will convene for half an hour to interview Karen Howe about her book Empire of AI and you can watch it if you want to watch it live, but we'll then put it into the show next week as part of the show. Show. But if you want to watch the live interview and, and participate too, because we take your, you know, we take your chat questions into consideration as we're talking 5:30 and your clubs and your AI creations. Yes.
- [94:09]
B
And all that you do.
- [94:10]
C
We take it all when you're in the.
- [94:11]
B
Take it all in if you're a club member.
- [94:13]
A
We see it, we take it on. Actually, Darren Okey, who is, as you know, user of AI, a Vibe coder, smart guy Australia, says, I strongly disagree. Leo, I was trying to build a chatbot interface 20 years ago, of course, Eliza's chatbot interface, and that predates that we're designed to communicate with people for any sort of assistant, it's the only sensible UI solution. I don't disagree with that. I just think that it's a dangerous solution. Certainly that's, you know, that's what people want. And if you're thinking of an assistant in that kind of human terms, that's the obvious, obvious interface. But, but what we're talking about is the risks of that kind of interface teaching people that, you know, the AI is a human or has some sort of humanity to it, which it doesn't. It's Just mechanical. Anyway, I want to take a break. We will come back with more. Just a little bit. You're watching Intelligent Machines. Jeff Jarvis, Paris Martineau. Our show brought to you by Pantheon. Which is kind of an interesting coincidence because our website is also brought to you by Pantheon. And for in fact, our entire workflow is brought to you by Pantheon. Your website is your number one revenue channel, right? If you're a business. But when it's slow or it's down or stuck in a bottleneck, it's also your number one liability. Pantheon keeps your site fast, secure. That's pretty important too, and always on. And that means better SEO, more conversions, no lost sales from downtime time. And it's not just a business win, it's a developer win too. Your team gets automated workflows, isolated test environments and zero downtime deployments. No late night fire drills, no works on my machine headaches, Just pure innovation. We use Pantheon to run our very special Headless Drupal instance. That's our content management system. Our editors use it to publish all of our shows. It provides a public API that can be used to develop apps or. Or our website, which is the number one consumer of the public API. That's where TWiT TV comes from, and it has a private API. And that's where the editors and all of our team use the Drupal backend, the Headless Drupal backend to get their jobs done. And all of it depends on relies on Pantheon IO. We're pretty happy with Pantheon. It's just a coincidence. They came to us and said, would you like to do ads? And I, I don't think they knew that we were customers. And I said, yeah, this is one I can really talk about. I'm pretty darn, pretty darn happy with Pantheon. With Pantheon, marketing can launch a landing page without waiting for a release cycle. Developers can push features with total confidence. And your customers, they just see a site that works 24 7, Pantheon powers, Drupal and WordPress. Sites that reach over a billion unique monthly visitors. Visit Pantheon IO and make your website your unfair advantage. Pantheon, where the web just works. And I will tell you, I am a very happy Pantheon customer. We've been using them for some years now. You know who's even happier than I am? Our engineering team, they love Pantheon. Patrick is a big, big fan. Yeah. In fact, in a response, Darren and others in our discord are saying, yeah, but this a human thing to see. You know, there's actually a term for it to see humans in. To See faces in, you know, Jesus in a piece of toast or see faces in a fire hydrant. It's a. It's a human thing and they just.
- [97:58]
C
Linked a funny subreddit that is actually kind of great.
- [98:01]
A
Oh, I love that subreddit. I. I follow it. That's it. Pareidolia is the term for seeing human, human, or other faces. Here's a piece of tile that looks like a wolf howling. Sure it does. Eric Cartman in a stack of dishes. That's pretty cool.
- [98:19]
B
What does it look like? I don't get it.
- [98:21]
A
An old. Oh, well, you don't watch south park, obviously. Old Man Spider. Yeah. Trees are a common. Common location for pareidolia. There's an elf peeking out of a.
- [98:34]
B
Did you ever use Smarter Child back in the. In the day?
- [98:36]
A
No. What's that?
- [98:38]
B
So two guys I used to work with at advance started it in 2000. I give you a link to the Wikipedia page. It was a. It was trying to add a chat to AIM so you can ask it questions and it would come up with answers like stock prices and things in a chatty, funny way.
- [98:54]
A
I. You know, I always wanted a IRC bot for our irc. We do now have a Leo bot. A Leo AI in our discord. Discord chat. And he's actually. He's actually not too bad. He's gotten better. I don't know if Anthony, you're. You're. You're feeding it more information or it's just getting smarter on its own. Or maybe it's come to life.
- [99:14]
B
Uh oh. It's conscious.
- [99:16]
A
Uh oh.
- [99:17]
B
Two leos. Oh my God.
- [99:20]
A
Smarter Child. Hi. Hi again. Trevor. How are you? I'm doing great. Are you okay? That's not super impressive.
- [99:29]
B
This is 2000. Thousand.
- [99:30]
A
Okay.
- [99:31]
B
Back when. Yeah, but it was a way to add something to aim.
- [99:35]
A
Yeah.
- [99:36]
B
Messenger.
- [99:36]
A
Yeah. That's pretty cool. It actually could do things like look up stocks prices. It could play Colossal Cave Adventure. It's interesting. Yeah. I mean, there was a lot of attempts at this kind of thing. I mean, buddy, this is. That was that first AI winner. Maybe not the first NAI winner.
- [100:02]
B
Well, that was the Internet winner. That was 2000.
- [100:05]
A
Yeah. Oh, did you love this story? Here's a college kid who had an interesting idea. He trained a local AI on nothing but Victorian writing, Victorian texts, because he wanted to have an AI that could answer questions as a Victorian. He was doing it just for fun. But he says an unexpected history lesson when he asked about events in London in 1834. And even though the developer had no idea. The chat time capsule LLM started talking about big London protests in 1834 and in fact those were accurate. He looked it up, he said oh yeah, yeah, it's an interesting idea. I think this is a really cool idea. There are other projects, One called Monad GPT trained on 11,000 texts from 1400 to 1700, basically the Middle Ages that can discuss topics using 17th century knowledge frameworks.
- [101:11]
B
Well, it's a little past the Middle ages. Gutenberg was 1450 enlightenment. Pardon me in there.
- [101:19]
A
Well what is it? 1400. See I always think of 1400. I remember Barbara Tuckman's book about the 1400s, A Distant Mirror. And that was kind of. That was the end of the Dark Ages, sort of Renaissance Enlightenment, later Enlightenment.
- [101:37]
B
Have you ever played with the universal short title catalog?
- [101:40]
A
No.
- [101:40]
C
What is that?
- [101:41]
B
If you go to ustc ac.uk uk ustc.ac.uk that's 1-800-usac.uk Got it. Okay, so this is my. My. The. The dean of book historians Andrew Pedigree along with Arthur Davetovan who by the.
- [102:00]
A
Way is the editor of Early Modern Publishers. Are all the early modern publishers on here?
- [102:09]
B
This is a. Is a. Is a catalog of every known thing printed between 1450 and 1700. If it exists somewhere in a library, they link to it. If even if it. If we only know that it existed because someone who wrote about it, they include it and you wander through this and it's amazing stuff. This is making me feel.
- [102:28]
C
Let me tell you what I found immediately, which is the title of this record is a person of very great, great quality having brought over from Italy the greatest secret in the world. A powder that in 10 or 12 days time perfectly cures the pox evil scurvy with only one dose taking not only shank nodes or ilate ulcers but also to working out of rotten bones whether from the pox or evil. And as cure, as above said, only observing an order and regiment of diet for prefix time. That's the short title.
- [103:01]
A
That's the title.
- [103:03]
C
The short title.
- [103:04]
A
I want the. I want to read this by a lover of hahaha jest books. These are joke books from. From the 8.
- [103:13]
B
The 1600s do deliver some of them the jokes.
- [103:16]
A
How about the 15 real comforts of matrimony published in 1683.
- [103:23]
B
Now how do I now there. If you. If you link. Oh no, go back. See the link to the right there. Now go back. You got. You're a bad scroller. Go back up. Go back.
- [103:31]
A
This link. This Link.
- [103:32]
B
Down there where this is. Link.
- [103:34]
A
Oh, Link. That is the worst ui. Okay, you might have full access to the full article, or you might not. Fifteen Real Comforts of Matrimony. Being in requital of the late 15 sham comforts with satirical reflections on whoring and the debauchery of this age, written by a person of quality of the female sex. No. Upon my word, ladies, twas neither favor nor affection or flattery nor fear, but something I know not what. You may, if you please, call it conscience. And something of gratitude for favors formerly received amongst you as being one of the same sex. And these two things would not let me. I have no idea what she's talking about. Be it quiet hearing ye. So abused and scandalized and daily reproached by Those who are 10 times worse than yourselves, that is to say, men. For these men have got a trick to lay all the weight and burden of their fears. Jealousies, discontents, disquiets. They're running in debt, they're breaking upon all the women's back acts and matrimony too must be arraigned for their sakes. But when we came to both the bar of reason and weighed the miscarriages of both, the one against the other, the men's scale was so heavy you could hardly lift it. The women so light you could hardly feel it.
- [105:01]
C
My God.
- [105:03]
B
You know what amazes me about this is this is a time when you had to set type one damn letter at a time. And they were wordiest as hell.
- [105:09]
A
They. They were more worried.
- [105:11]
B
Just talk to the machine and we do it in tweet Twitter links.
- [105:13]
A
Yeah? Yeah.
- [105:14]
C
Were they getting paid by the word 97? The merriest jest that you ever did hear of. Strange confusion in Lincolnshire. But it's a lost book, so we'll never know.
- [105:27]
A
Never know what?
- [105:27]
C
The jest you ever did here.
- [105:30]
A
I'm still waiting for that Italian powder.
- [105:33]
C
It could solve all your problems. Pox and bone related.
- [105:38]
A
I have the science be. Let's see, what else? What else should we talk about?
- [105:45]
C
That result was brought to you by. My favorite thing when you're going into a new fun database is just search the word bones, see what comes up. Really?
- [105:52]
A
Is this what you do?
- [105:54]
C
I've done it multiple times and it's brought. It's brought me joy.
- [105:58]
A
Bom bones. Huh? Why do you think that is?
- [106:00]
B
Where did that come into your nihilistic head?
- [106:04]
C
It came from gif cities, which is if you go to gifcities.org which is the geocity animated gif Search from Internet archives. I'm sure I've plugged in the show before. Someone showed me this and I immediately decided to search the word bones. And it was great. I had a really good look at.
- [106:22]
A
All the bones you get so many bones. Them bones, them bones, them crazy bones the bones connected to the thigh bone Thigh bones connected to the hip bone Good day, Mr. Bond.
- [106:38]
C
Bone thugs in Harmony.
- [106:41]
A
Now that's a real group.
- [106:43]
C
Oh, I didn't know.
- [106:44]
A
Yes.
- [106:45]
B
Really?
- [106:45]
A
There you are. Oh, yeah.
- [106:47]
B
Oh, yeah. Like you've listened to them.
- [106:48]
C
Oh, yeah.
- [106:50]
A
Come on. You have too, haven't you?
- [106:53]
C
Have you listened to rustybongwater.com?
- [106:57]
A
I'Ve smoked it. I've never listened to. It isn't quite the. I love these old websites. So my. They're great.
- [107:04]
C
So if you haven't.
- [107:05]
B
This is a Paris thing.
- [107:06]
C
I was gonna say if you haven't. This is your book website before it's GIF Cities or Gift Cities, depending on your pronunciation dot org. And it essentially is part of the Internet Archive has scraped all of geocities and this is something you can just search a word and it'll bring up all the associated gifts and you can Click click the GIFs and it takes you to the website. Like the one Leo just found.
- [107:31]
A
Yeah.
- [107:31]
B
Mr. T. Did you know who that is? Paris.
- [107:35]
C
I've heard him.
- [107:36]
A
You've heard of him. But have you ever seen the A? I've the A team.
- [107:43]
C
No.
- [107:44]
A
No, that's all right. I haven't either. Okay. That. We've learned a little something here.
- [107:53]
B
Go ahead.
- [107:54]
A
Chipotle.
- [107:55]
B
We used to talk about Chipotle all the time in this show.
- [107:57]
A
You're going to be excited about this, Jeff. Chipotle's latest delivery option, a 299 delivery service starting in Dallas could take off nationwide if successful. It is called Zipline, an autonomous delivery service that will allow digital orders to be transported by drones called zips. That your Chipotle order will come to you by drone. And you see what it's doing. The drone is above and it's lowering down a cable and. And dropping your Chipotle.
- [108:32]
C
I cannot wait for the day that I am outside minding my own business and I get dive bombed by a burrito. I think it'll be great.
- [108:43]
A
Would you like a video? This is a video of a. The launch. So they're making. Making. They're making the burrito. They're dropping it into a bag. Here comes the Zipline drone. They're putting in the drone.
- [108:59]
C
What is to stop someone from shooting it Down. Yeah.
- [109:04]
A
Did you see it lays the Chipotle like an egg? Like an egg. But wait a minute. Did you see the wire? This is, like, fake. Did you see the wire?
- [109:16]
B
No, that's the. That's the.
- [109:17]
A
Oh, it drops. Oh, it's from a plane. Oh, so wait a minute. There is a wire because it's lowering it down.
- [109:25]
B
Yeah. From a mothership.
- [109:26]
A
From a mothership. Like a. Like a cruise missile drone.
- [109:31]
B
So, Paris, did you see the story that in Ukraine they're using old trainer planes that have open hoods and they're going out and they're shooting down Russian drones with shooting shotguns on the plane?
- [109:45]
C
That's pretty cool. It's pretty metal.
- [109:47]
A
It's inevitably going to happen to the zip drone as well. Okay. So employees will place the order into a zipping point, which allows zips to autonomously pick up the order for delivery. After flying to its destination, the aircraft will hover about 300ft in the air while the zip rollers to the ground. Yeah. The Zip automatically avoids obstacles and gently and precisely places the order at the guest's address. The Chipotle at 3901 Lakeview Parkway, Rowlett, Texas, will be the first to offer zipotle deliveries. Oh, my God. I wish I lived in Rowlett, Texas, right now. I would order some zipotle. That's crazy.
- [110:31]
C
I've got so many questions. How is the big plane cleared to fly over your home?
- [110:37]
B
It's a drone. No, it's a drone.
- [110:38]
C
It's just a bigger drone.
- [110:40]
A
Well, I'm sure that has to get licensed.
- [110:42]
C
Drones are licensed. License to fly all over. And so you can't get this. I guess if you have, like, trees in front of your.
- [110:49]
B
Or like, if you're in a city or if it's.
- [110:52]
C
Can the drones fly? This is a dumb question. Can the drones fly in stormy weather?
- [110:56]
A
This is just shows that the rain effect. Complete times. So ridiculous. Oh, hey.
- [111:03]
C
Sorry, your burrito has been delayed. It's raining.
- [111:07]
A
Really happy to say that Peter Thiel's going to be giving. Giving a series of three lectures on the Antichrist.
- [111:14]
B
Oh, no.
- [111:14]
C
Great. Cool.
- [111:16]
A
That bodes well here in the San Francisco Bay Area.
- [111:20]
C
I might say he has a lot of familiarity with the topic.
- [111:23]
A
I didn't know. This is massively interested in the Antichrist. Oh, yes.
- [111:29]
B
That's what drives everything. That's everything.
- [111:31]
A
That's what drives everything.
- [111:33]
B
Yes. His belief. This is where you heard when you played that guy from the video from the guy at. At Palantir about who you Want your enemy be scared? Yeah. This is all end time stuff.
- [111:47]
A
What?
- [111:48]
C
This is.
- [111:49]
B
This is a very.
- [111:50]
A
Apparently he told Peter Dutad at the New York Times that Greta Thunberg is the Antichrist, which I don't.
- [111:56]
B
Yes.
- [111:56]
A
If you look at the. Yeah, I don't think that's. No, no. Anyway, he's going to be speaking at the Acts17 collective in San Francisco. That's the Commonwealth Club. You've spoken in there. Four part lecture series led by Peter Thiel, exploring the biblical figure of the Antichrist through the lenses of science, theology, history, politics and literature. But mostly through the lenses of this one crazy guy.
- [112:24]
B
I don't think they've ever shown a picture of him smiling. That must be AI.
- [112:31]
A
That is.
- [112:32]
B
What does it cost, Leo? Reddit register. Go ahead, register.
- [112:35]
A
We should go. It's. I mean. Oh, it's sold out. Oh, it's sold out.
- [112:42]
B
It's not virtual.
- [112:44]
A
No, you gotta.
- [112:45]
B
Virtual.
- [112:45]
A
It's no longer. You're warmly invited to a series of four lectures. You get drinks and small bites lecture. I hope small bites are in the shape of the devil. I would. You get Q and A with Peter Robinson. Q and A with audience. Audience.
- [113:00]
B
Oh, Leo, you got a link on Acts.
- [113:02]
A
Oh, Acts 17 is a collective acknowledging Christ in technology and society. Oh, the fiery pits of hell. Oh, no, I'm sorry, that's just a. One of them.
- [113:16]
C
Christ in technology and society.
- [113:19]
B
Wait, wait, wait. Was that. Was that Cook?
- [113:22]
A
Tim Cook?
- [113:22]
B
No, no, I saw the White House.
- [113:24]
A
Just a white.
- [113:24]
C
That was another man with white hair. You. You shouldn't be judgmental like this, Jeff.
- [113:31]
A
Wow.
- [113:31]
C
I just imagine you walking around your home being like. Is that Tim Cook? Oh, no, it's just a mirror.
- [113:36]
A
I mean. So this is among evangelical Christians. This is part of Revelations, right? The. The end times. That the Antichrist will come.
- [113:47]
C
The cheerful book, succinct, podcast friendly explainer of the end times. Right.
- [113:53]
B
Right now I'm listening to the Delusions of Crowds, why People go Mad in Groups by William Bernstein. And it's all about these End crowd. And world in times.
- [114:04]
A
Yeah.
- [114:05]
B
And times. Yeah. And it fits right into. To get. Get ready for a drink, folks. Test real.
- [114:10]
C
Hey.
- [114:11]
B
All part of the same thing.
- [114:14]
A
Wow.
- [114:14]
B
It's a. It's a religion without God, but it's a still religion. But here he's.
- [114:19]
A
It's funny that he thinks Greta Thunberg, of all people, the teenager who's trying to save the planet from.
- [114:24]
C
I don't know if she's a teenager.
- [114:26]
A
Anymore, but Young woman who's Trying to save the planet from climate disaster.
- [114:33]
B
Yeah. She's trying to save it from end times.
- [114:36]
A
Yeah. Well, let me just, you know, she's 22. Okay. Wow.
- [114:42]
B
She's grown. Or Greta.
- [114:44]
A
She's no longer a teenager. In Christian eschatology, the Antichrist refers to an entity prophecy by the Bible to oppose Jesus and falsely substitute themselves as a savior in Christ's place.
- [114:58]
B
You see?
- [114:59]
A
Before the second she says she's saving.
- [115:01]
B
The world, but she's not.
- [115:02]
A
Yeah. Okay.
- [115:07]
B
Or he is.
- [115:09]
A
Okay. It's an interesting notion.
- [115:12]
B
Where did you come across that?
- [115:13]
C
That.
- [115:14]
A
This is Wikipedia.
- [115:15]
B
No, no, no. Peter Thiel speaking.
- [115:17]
A
Oh, my daughter sent it to him. Sent it to me. I said, wait a minute. He is the Antichrist. And she said, right. That's how he can talk about it. Here's something a little lighter. Let's change the subject quickly. Something a little lighter. You're listening to Intelligent Machines. Jeff Jarvis. Paris Smart. No, I'm Leo Laporte. I hope you heard our interview with MGC Ziegler. If you didn't rewind, because he was great.
- [115:42]
B
He's wonderful.
- [115:43]
C
None of us are the Antichrist.
- [115:45]
A
No, none of us. We're pretty sure. Although, you know, it could be. Wouldn't the Antichrist say they weren't the Antichrist? I mean, they wouldn't admit it.
- [115:53]
C
Maybe none of us were admitting we're the Antichrist.
- [115:56]
B
Wouldn't the Antichrist have blue fingernails?
- [116:00]
C
Wouldn't the Antichrist have blue fingernails that she needs to get refilled?
- [116:07]
A
Yeah. She wouldn't have chipped blue fingers. They would be naturally blue. Next week, Karen Howe, author of The Empire of AI should be interesting. Really? A deeply researched book about OpenAI watch length. Yeah. Her contention is that, much like the colonial empires of the last century, AI companies like OpenAI are colonizing the world. World for their vision.
- [116:35]
B
I left the book unsure of what she thinks of AI. I know what she thinks of Big Tech, but I'm not sure what she thinks of AI Nor even a Sam Altman.
- [116:43]
A
Yeah.
- [116:45]
C
By Monday, we're gonna have to have a kind of point of view squirt gun situation set up in Leo's office with a remote control on my desk just in case.
- [116:59]
A
Down, boy. Down, boy. Okay.
- [117:03]
C
A mochi moment from Mark, who writes, I just want to thank you for making GLP1s affordable. What would have been over $1,000 a month is just $99 a month with mochi. Money shouldn't be a barrier to healthy weight. Three months in and I have smaller jeans and a bigger wallet. You're the best. Thanks, Mark. I'm Mayra Amit, founder of Mochi Health. To find your mochi moment, visit joinmochi.com Mark is a Mochi member, compensated for his story.
- [117:33]
A
Let's map out this week's amazing destinations and travel tips.
- [117:37]
C
Honestly, Will, I didn't plan any trips, but I did switch to T Mobile with their new Family Freedom offer.
- [117:45]
A
That's not the itinerary we're following.
- [117:47]
C
Well, I'm departing from AT&T and embarking on a new journey with T Mobile Mobile. They paid off my family's four phones up to $3200 and gave us four new phones on the house.
- [118:01]
A
Bon voyage. Introducing Family Freedom.
- [118:04]
D
Our lowest cost will switch our biggest.
- [118:06]
A
Family savings all on America's largest 5G network. Visit your local T Mobile location or learn more@t mobile.com FamilyFreedom up to $800.
- [118:17]
D
Per line via virtual prepaid card typically takes 15 days.
- [118:19]
A
Free phones via 24 monthly bill credits.
- [118:20]
D
With finance agreement eg Apple iPhone16128Gigabyte829 99 eligible trade in eg IPH11 Pro for well qualified credits end and balance due if you pay off earlier, cancel contact.
- [118:28]
A
T Mobile from unsolved mysteries to unexplained.
- [118:31]
B
Phenomena from comedy goal to relationship fails.
- [118:34]
A
Amazon Music's got the most ad free top podcasts included with prime because the.
- [118:40]
D
Only thing that should interrupt your listening is, well, nothing.
- [118:44]
B
Download the Amazon Music app today.
- [118:55]
A
A couple more stories. We have like a thousand more, but.
- [118:59]
B
I have lots of good stories.
- [119:00]
A
Attempt to there's so many good stories this week. There are two types of dishwasher people and I really want to listen to.
- [119:08]
B
Their wives and those who don't.
- [119:09]
A
That's right. That's what I do. This is a very funny piece by Ellen Cushing. Only one of them really knows how to look Load the dishwasher is your wife. Yes. There are YouTube videos with titles such as you're doing it wrong and passive aggressive tutorials. My wife, speaking of passive aggressive, will frequently reload the dishwasher when I step in.
- [119:36]
B
Or mine will change the rules. I follow the rules. We have a great dishwasher. It's a wonderful dishwasher and we follow the rules. But then suddenly there's a new rule.
- [119:43]
A
Do you put your cutlery and handle up or we have we.
- [119:48]
B
Oh, ours is. Ours is the one that has the top row and you put it down like that, which I thought was going to be awful.
- [119:52]
A
And it's great That's a better idea because I have always been the handle up kind of guy. Because you're going to take it out.
- [120:00]
C
Are you crazy?
- [120:01]
A
Yes.
- [120:02]
C
Apparently there should be a third subset in this for people whose apartments don't have a dishwasher. So they may or may not have installed a mini countertop, one that had very limited space in order to allow for some dishwashing at the time.
- [120:24]
A
You're not alone. Under 90 million American homes have a dishwasher and 19% of them were never used zero times a week. Part of that is confusion over not only how to load it, but what you should do before you load it. Do you scrape the plate? Do you.
- [120:41]
B
Okay, so are you scrapers or rinsers or not?
- [120:43]
A
You're not supposed to scrape anymore. I don't believe that.
- [120:45]
B
I said I don't believe that. What do you mean it's not supposed to?
- [120:47]
C
Because I think with normal dishwashers. Yes, you're supposed to scrape the plate. With my miniature dishwasher, I gotta. Because there's no place else for the. For the stuff.
- [120:58]
A
Yeah.
- [120:58]
B
So what I would say you do.
- [121:00]
A
Scrape, but you don't want to scrape it. You don't want to rinse it. You want it to be thoroughly. Because apparently the. The food proteins left in the. The plate interact with the detergent and help it clean.
- [121:08]
C
Yeah, it's an enzymatic reaction.
- [121:09]
A
It's an enzymatic reaction.
- [121:12]
B
So when I was the kitchen manager at Ponderosa Steakhouse in Lombard, Illinois, Mike, the manager pointed to the machine and he said, that's not the dishwasher you are, that's the dish sanitizer.
- [121:26]
A
Oh, I like it.
- [121:27]
C
That is true, though, in kitchen and industrial situations is typically, you don't have a dishwasher, you have a sanitizer. This was true when I worked at Starbucks for many years in high school.
- [121:37]
A
When I worked at the dining hall in college, it's all the food service. I wish we had this dishwasher. This was a conveyor belt and you'd put the dishes in a rack. You remember this, right? And you put on the conveyor belt and would go through and it'd be hosed and you'd have. And you would have. As the person.
- [121:56]
B
I love that thing. Because all the other employees, you'd like.
- [121:59]
A
Have a shower sprayer that you'd spray everything off before it went into the dishwasher washer.
- [122:04]
B
The sanitizer, Leo.
- [122:05]
A
It would. The sanitizer and then it would emerge piping hot and sanitized for you to Put out and watch the students go.
- [122:15]
C
Open it and get glasses all fogged.
- [122:18]
B
Yes, exactly.
- [122:19]
A
This thing is always open. It's just a conveyor.
- [122:21]
B
See, we are people of the people here. Yes, we've had real jobs in.
- [122:26]
A
In a YouGov poll from last year, more than a third of respondents admitted to having disagreements with others about dishwashing. Best practices, particularly comes up in couples therapy sessions. So there was a Reddit ama. She received hundreds of questions. Her name, by the way, is Forte. She is apparently the wire cutter. No, no, that's somebody else. I don't know who Forte is. I hate it when they do this in articles, by the way. They refer to somebody that's paragraphs back. I don't have that kind of memory, guys. Who the hell. Say your name again. All right.
- [123:14]
C
Well, she's a parent Forte, who has worked for more than 40 years at Good Housekeeping and currently runs its home care and cleaning lab.
- [123:22]
A
Well, she's an example expert. She said she got hundreds more questions than Good Housekeep on the Reddit AMA than Good Housekeeping ever got. A large number started with the poster invoking a spouse's or roommate's wrong headed approach to the dishwasher and asking Marilyn to adjudicate. There's a lot of psychology behind the dishwasher, she said. There's a lot of angst.
- [123:43]
B
Why? Why does it. Why does it. In fact, it does. This is true. Why does it load so. Such pedantry in a home?
- [123:51]
C
That's a good question because it's a constant physical reminder of someone of, of the lack of control you have over basic household functions that have to be constantly done and undone and redone.
- [124:07]
A
So sociologist Michelle Janning has one answer. She wrote the the book the Stuff of Family Life, How Our Homes Reflect Our Lives. She told the Atlantic the strong opinions associated with how to do it could be people trying to retain some semblance of control in a world where technological devices are doing things so much for us.
- [124:32]
B
Well, we're in charge of the dishwasher. I don't feel like the dishwasher is taking over my life.
- [124:36]
A
You're not. Are you guys, quote, losing the humanity associated with your domestic life?
- [124:41]
C
No, no, I need no more humanity in my domestic life. Do you guys. This is the one problem I've realized if whenever I go visit my folks. Do you guys clean out the filter of your dishwasher?
- [124:53]
A
I do. I'm the one in the family who does.
- [124:55]
C
You got legitimately. My parents had not seemingly ever. And I think, like, I People don't.
- [125:02]
A
Know that until the dishwasher stops raining.
- [125:05]
C
Potentially. Vomit. Basically based on the smells I encounter.
- [125:08]
A
Disgusting. Yeah.
- [125:09]
C
Yeah. For those in the chat asking, there was a filter. Yeah, there's a filter at the bottom of the dishwasher. You gotta remove the little twisty top thing, then pull out some other junk and then there's a filter with actual crud in it that you gotta clean.
- [125:22]
A
Yeah.
- [125:22]
C
Go fix it.
- [125:23]
A
How often do you do. Here's. Here's what the Atlantic says. Clean your filter periodically if you have one. You might not. If your dishwasher smells, you can run a sanitizing cycle with vinegar, baking soda, or one of the many special. Good lord. Dishwasher cleaning products. Also, no cast iron. I keep telling. By the way, the family doesn't get this one. No cast iron, no wood, no crystal. Nothing with delicate painted on decorations like china with metallic rim. Please, I beg of you. I keep finding my good knives in the dishwasher. Rinsing is not necessary. From Oma Blaze Ford, a senior executive editor at Better Homes and Gardens. Gardens over rinsing is, quote, one of the most common mistakes in modern dishwashing. Loading. Scrape the leftover food into the trash with a rubber spatula, then immediately load the dish into the machine without even turning on the faucet.
- [126:17]
B
Oh, no.
- [126:19]
A
What about peanut butter? Okay, what about. Oh, she says peanut butter eggs are the. Are the main exception. They tend to be stuck.
- [126:25]
B
Okay. There's always exceptions to these rules.
- [126:27]
A
There's always exceptions. Exception. I can't believe she got 4,000 words.
- [126:31]
B
I know.
- [126:32]
A
Loading a dishwasher. Thank you, Ellen Cushing at the Atlantic. This is why I subscribe.
- [126:38]
B
So, Paris, I think I found an activity either for your group or maybe for the two of us. Line 151. You need to come to New Jersey for this.
- [126:48]
A
Okay. I have some activities too, but they aren't in New Jersey.
- [126:51]
B
You want to believe an Amazon tour?
- [126:54]
C
1O.
- [126:54]
A
I would take an Amazon tour even of the warehouse.
- [126:58]
B
Yeah.
- [126:59]
A
Oh, you can do this. Look, I can go over to Tracy or San Diego or East. Wow. Is there one near you in New Jersey?
- [127:05]
B
Yes, there's one in Edison right nearby. There's not in New York. Paris, that's why you got to come here. Well, maybe it's a fun outing with the Skee Ball.
- [127:13]
A
Oh, would this be cool? And the one in Edison is a robotics sortable facility. Yeah. So you'll see the robots.
- [127:23]
B
Yeah.
- [127:25]
A
Know, before you go, each tour is 69 minutes long. You may be walking up and down at Least one flight of stairs and walk approximately one mile. Do not wear sandals, clogs, crocs or.
- [127:37]
B
High heels because crocs are really ugly and you really shouldn't be seen in them.
- [127:41]
A
We won't even let you in if you wear unsuitable footwear or offensive or revealing clothing. No cleavage in the wear.
- [127:50]
C
Telling me I can't wear a bandage dress to the Amazon warehouse.
- [127:53]
A
I was gonna wear my tube top. Wow. We do welcome young visitors, but you've gotta be at least 6 years old.
- [128:02]
C
This is interesting because. So when I was covering Amazon exclusively, it was during the height of the pandemic, so I was never able to do official Amazon tour.
- [128:11]
A
I would love to go. This you can bring. Now, this is unlike the Sydney Harbor Bridge where you have to check your wallet and keys because they may fall out and hit somebody in the head. You may bring your wallet, keys and transparent water bottle on the tour, but we do ask you keep cell phones, bags and backpacks at home or in a vehicle unless medically necessary.
- [128:31]
B
Cell phones are not allowed on the tour. Do you put them in?
- [128:34]
A
No photography. They say leave them at home or in the vehicle, but yet you are.
- [128:38]
B
Permitted photos in designated photo areas. Oh, in our lobbies. Oh, I see.
- [128:43]
A
See, this would be super cool, wouldn't it? August 28th. Let's do it. I'll be right out. 11am to noon. There are. There's one spot available and the next one's fully booked.
- [128:57]
B
Well, the next month there's more.
- [128:58]
A
It's popular. Yeah, of course it is. Are monocles allowed? Burke wants to know if he can wear his monocle. Only if it's securely attached to your clothing burn.
- [129:10]
B
Can you wear your photo glasses?
- [129:12]
A
Ah, I bet you cannot. That's a good question. This is going to be more of a difficult, you know, for a long time I'd see these museums as, you know, photography in the museum, but everybody's got their cell phone, they've got a camera with them at all times.
- [129:24]
B
That was also because they didn't want flashes which would have an impact.
- [129:26]
A
Flashes. I understand. Yeah.
- [129:31]
B
Would this be an activity for the skee balls or it's not nearly fun enough.
- [129:35]
C
I think this is an activity for the podcast.
- [129:38]
B
Okay. I think so too.
- [129:40]
A
That's funny. The scheme, Leo, they have.
- [129:43]
B
They have an opening in Tracy on September 5th on a Friday. They have seven spots open right now.
- [129:49]
A
Let's. It could 10:30am Road Tripani activity. We could all go.
- [129:55]
B
How far away is.
- [129:56]
C
Screw a fancy dinner. Let's drive everybody to the Amazon Warehouse for team building.
- [130:02]
A
Building. Let's see. I think you should nominate some because I'm not seeing anything that we need to cover and we do need to wrap it up so.
- [130:19]
B
Well, we got the end.
- [130:20]
A
What about perplexity asking for $5 a month for their browser comment and then they will divvy that money up to publishers.
- [130:30]
B
Does that seem like it's another Perplexity trick? A who's gon who's going to pay five bucks? So we're going to share this revenue we're not going to get, but we're going to get headlines saying that we are nice and we're sharing revenue with publishers. And it's worked already.
- [130:43]
A
Yep. Comic browser already has a few issues. Somebody demonstrated at defcon a way to use it to drain your bank account. So there's that. Or just order crap you don't really want because it's an agentic browser. Brave demonstration, Paris, you were going to.
- [131:03]
C
Say, oh, did we mention last week? Or was this after that Wired, Business Insider, and other news outlets had to remove AI written freelance articles?
- [131:14]
A
No, I didn't. I don't even remember.
- [131:15]
B
Oh yes, it's in there.
- [131:17]
C
So this was originally written by the Press Gazette. And this came to their attention after the author of Dispatch, who's also kind of the former debut of a herd, Jeff Jacob Furdy, kind of did an interesting blog about this crazy freelance pitch he got that ended up being too good to be true. So he got a pitch at the beginning of the month that was offering this writer saying they were gonna do a reported piece about Gravemont, a decommissioned mining town in rural Colorado that has been repurposed into one of the world's most secretive training grounds for death investigation. The pitch continued. I want to tell the story of the scientists, ex cops and former miners who now handle the dead daily, not mourners. But as archivists of truth, I'll explore the ethical tightrope using real humans. All the stuff like I've got the right person because I've got all this sourcing. Faridi kind of walked back and forth with this freelancer. It seemed like a good pitch at first, but then he did a little googling and was like, I can't find Grave Mount. How do you do you find? It turns out, skip the end, this city didn't exist because this person didn't exist. And then he started looking at this writer's clips and it was for a lot of bigger sites.
- [132:42]
B
It's a French name, which is suspicious.
- [132:45]
C
Why?
- [132:46]
A
Marco Blanchard.
- [132:47]
B
Marco Blanchard.
- [132:49]
A
So this is interesting. This is not the publications doing this. The publications are getting pitched by a phony freelance answer. They're not doing the due diligence they should do, but they accept the pitch. Wired published a story titled they fell in love playing Minecraft. Then the game became their wedding venue. Wired was fooled. We made errors here. This story did not go through a proper fact check process. That's by the way, a big faux pas or get a top edit from a more senior.
- [133:23]
B
Nobody has fact checkers every anymore.
- [133:25]
C
Yeah, no one has fact checkers and no one so but I mean there were small things. So like if you dig into the Wired story this person published, it has like small details. Like Jessica, who, 34, an ordained officiant based in Chicago, has made a name for herself as a digital celebrant specializing in ceremonies across Twitch, Discord and VR chat. However, if you Google that person's name in quotes and like Chicago, nothing comes up, much less if you look for.
- [133:52]
A
An officiant by that name, this bit of dude, due diligence would have.
- [133:55]
B
Well, plus they couldn't pay the freelancer.
- [133:57]
C
So part of. So Wired didn't. So Business Insider and some other publications, if I recall correctly, ended up taking their articles down after the Press Gazette reached out to them with questions about this phony freelancer, Wired.
- [134:11]
A
So the Press Gazette immediately spotted it, which is interesting.
- [134:14]
C
Well, no, the Press Gazette reported on someone else's immediately spotting of it. But Wired had taken this down. This story they published down shortly after publication, within a week or two, once the editor was trying to pay the freelancer and they started asking, no, I don't want to go through the normal payment channels. I'll only do PayPal. And that's when it all started to unwind.
- [134:35]
B
You don't do that kind of nast, believe me.
- [134:37]
C
Yeah, no you don't. But part of what I think is interesting about so Wired published kind of a mea culpa on this with no byline of how they got roiled by an AI freelancer. However, in it they say after we got this PayPal request, a suspicious Wired editor ran the story through two third party AI detection tools, both of which said the copy was likely to be human generated. Yeah, so what? It's like Wired, you have reported on this before. Who allowed this line to go in a story?
- [135:09]
A
Yeah, that's absurd. Well, it's just, you know, this is more AI slop and we're going to see more and more of that. And this is. They got Wired and Business Insider got Conned. What it's revealed, however, is they aren't doing even the slightest bit of due diligence on articles.
- [135:25]
B
They publish the volume now. Leo. The newspapers I know of, the metro newspapers, it's, it's all trending, chasing. Find the trending topic, rewrite something. Do they have a quota? They have five, six, eight.
- [135:40]
C
That's been the take, that's been the mistake for a decade. I was doing that in 2017.
- [135:46]
B
Yep.
- [135:47]
C
Not, not at Wired. I was doing that at New York Mag.
- [135:49]
B
Right.
- [135:49]
C
Wired doesn't really do that sort of trending. This is a unique thing because Wired typically like they only work with freelancers for longer narrative reported features and kind of like a. This was a one off story about like a Internet trend. I think the conclusion of Wired's kind of mea culpa essay is interesting though, which is fabulous. And plagiarists are as old as media itself. But AI presents a new trend challenge. It lets anybody craft a perfect pitch with a simple prompt and play act the role of journalist convincingly enough to fool, well, us. We acted quickly once we discovered the ruse and we've taken steps to ensure this doesn't happen again. But in this new era, every newspaper should be presumed prepared to do the same. It's kind of difficult to check when you're, I guess, a overworked editor dealing with a big news cycle. I think this is more of a, a sign of media economics and the dark place this has gotten the news industry than anything a specific publication did wrong.
- [136:47]
A
There is a knock on effect. I'm going to have to be much more careful about stories I pull for our shows. Maybe I will have to do the due diligence on those. We report stories like that all the time. To my knowledge, we have not repeated a fake story. Knock wood, Leo, knock on wood. But I could probably be fooled because in the past I've assumed the authority of somebody like Business Insider or Wired and figured, well, they must have vetted it. Apparently that's a mistake and I won't do that anymore. So expect much shorter shows going forward. You're watching Intelligent Machines. We are gonna wrap it up with our picks of the week in just a moment. Jeff Jarvis is here, professor of journalist innovation, journalistic innovation emeritus at all those schools of journalism at the City University of New York. Also Montclair State University and SUNY Stony Brook. His author as ship. His author credits include the Gutenberg parenthesis, now out in paperback. Oh, look, there's the Bookshelf magazine, now out in audio format. And the web we weave a fabulous book about with it's almost a minute manifesto for how to save the web which at this point we really need to work on. There you go. Thank you, Jeff. Paris Martineau is also here. She is responsible for radioactive shrimp coverage at Consumer Reports. Actually, that's an old lower third. We got the old lower third up for you.
- [138:16]
C
Wow. We do have the old lower third up for you.
- [138:18]
A
All of that's still good.
- [138:19]
C
Look at all. It is all still good. But I used to have such a meaty lower third.
- [138:23]
A
You did. Now we just say Consumer Reports Paris NYC website.
- [138:28]
B
Which she does not speak on the show.
- [138:30]
A
Right.
- [138:31]
C
I do not speak for them. I do not represent them. I merely worked for them in my.
- [138:36]
B
We are very grateful that they allow her to be here.
- [138:39]
A
Yes, it's true.
- [138:41]
C
Very cool.
- [138:43]
A
Thank you for joining us. Club members. Thank you for your support. We couldn't do this show without your support. 25% of our operating expenses now comes from club Twitter. It. I'm. I think that's a really good thing. It means we're getting the support we're getting. You know, it's. I think of it almost as votes. People say, yes, we like what you do. We want you to do more of it. If you like what we do and you want us to do more of it, you want to keep those shows going. All the stuff we do in the club, including coverage of keynotes coming up. We're going to do the Apple iPhone announcement. That's September 9th. Micah Sargent and I'll do that and we will do it only in the club. So that's why you need to be.
- [139:22]
B
Are you going to do Meta Connect?
- [139:24]
A
Yes, we are. You want to join me for that?
- [139:25]
B
Yeah, sure. That'll be interesting.
- [139:27]
A
More than welcome. That's our new thing. We do all the keynotes in the club only. So that's one reason to.
- [139:35]
B
That's how I watched the. The Google made by Google and yeah, it was fun.
- [139:41]
A
Micah, Micah and I, yeah, Mike and I had a lot of of fun watching Jimmy Fallon and company. We also have special shows we do only in the club like Hands on Windows, Hands on Apple. Micah and Paul respectively actually irrespectively do that show. Paul does Hands on Windows, Micah does Hands on Apple. We also do what else Hands on tech. We do Home Theater Geeks with Scott Wilkinson, Chris Markworth's photo show, Stacy's Book Club. I can go on. Best thing to do is find out more about Club Club Twit by going to Twitt TV Club Twit. There's a two week free trial which will give you actually enough time to sign up and at least listen to the Apple keynote, see if it's worth your time. I think it is and it sure is a big boon for us. So thank you in advance to all the club members. Twit. TV Club Twit. We will have to end. I am a little early during the Meta Connect. We've decided though, really it's a continuation. Continuation of the show because they're going to begin at 4pm so we will continue the show. We'll start the show at a usual time and we'll continue into Meta Connect. So you will be here. It's only really a question of whether you want to stay. Same for you, Paris. If hunger is taking over and you're shaking, you can take off. But if you want to stick around for as long as Mark Zuckerberg does, you're more than welcome to.
- [141:09]
C
Is that this week or. No, next week?
- [141:11]
A
Two weeks.
- [141:13]
B
And it's. It's interesting because they're going to supposedly reveal Hypernova smart glasses with.
- [141:18]
A
Yeah.
- [141:19]
B
The display and the wristband to.
- [141:21]
C
So they have to be called Hypernova, you know. Wait, can we have a brief moment for the incredible image Anthony generated?
- [141:29]
A
You were laughing.
- [141:30]
C
I was trying not to laugh because I realized my mic was still on, which is the official no guac for public figures title card. And I like that in the corn, bottom right corner, it says silly episode, exclamation point. Just in case you didn't get that from the little avocados with their arms crossed.
- [141:49]
B
The angry avocados.
- [141:51]
A
Oh, I get. And I now get the headline. That was because Gemini wouldn't do guac out of my shirt.
- [141:58]
C
Me, no guac for public figures.
- [142:01]
A
Yeah. Or maybe it thinks I'm a private individual and that's why I wouldn't do it. Pick of the week, Paris Martineau.
- [142:08]
C
All right, I got a couple things here. One, my general pick of the week that is relevant to no one but me because it has already happened as a niche local event is it was the South Slope Derby.
- [142:21]
A
I saw your pictures. So.
- [142:23]
C
Which is a delightful event that happens every summer in South Park Slope of Brooklyn where it's a summer summer camp for kids where they learn how to do like soapbox derby cars and then they race them down a bunch of closed streets all in like a one by one and they get raided. They have interesting themes. That one was a double decker bus. Yeah, double decker bus themed. There was. There was one where I didn't get a photo of this because it was so surprising. There was one where children went down in some sort of cart and then just adults emerged to pelt them with balls. Balls. I don't entirely understand why there's always something strange going on, but I don't know, it was just a delightful. I. As I was walking up there, I always stop at this one coffee shop and I was getting my coffee and the guy was like, oh, up to anything fun today? I'm like, oh, I'm going to the South Sluff Derby around the road. And he's like, oh, who are you rooting for? And I was like, oh, no one. I'm just a childless adult who goes to participate because I enjoy it. That photo is where they crash all the cars into a big pile and then adults come out with sledgehammers and saws.
- [143:34]
A
No, they destroy them at the end. They do.
- [143:37]
C
Cuz where else are these going to go in people's apartments?
- [143:39]
B
Got apartments in Brooklyn.
- [143:41]
C
So the guy's like, oh, who are you rooting for? I say, no one. I'm a childless adult. He's like, oh, then you've got to pick an enemy, right? And I'm like, what? And he's like, you should pick one of the kids to be your enemy and root against them for the race. And I was like, that's a way to do it. My guy, Boo.
- [143:56]
A
Boo. His boo.
- [144:00]
C
And these photos were taken with Project Indigo and they're so good, right?
- [144:05]
A
Nice.
- [144:05]
C
Last week. Nice. I got two and a half other picks just in case. One is I saw someone posting a couple of months ago or something. We'd done the Internet road trip, which was a part of the neil.fun website where.
- [144:20]
A
Love Neil.
- [144:21]
C
Kind of the Internet's controlling a road trip. They had gone up through Canada, but now they're back in the US in Vermont if you want to go check where they are.
- [144:31]
A
Yeah, they seem to be driving off the road right now.
- [144:36]
C
Yeah, they're kind of right now stuck in a parking lot. But they were on the road a bit recently. But I just thought it was a fun thing and I wanted to shout out some of the reviews we got from last Yay. People heard our plea. We got a whopping six reviews, guys.
- [144:55]
B
Yay.
- [144:58]
C
Sorgatron thanked me for making the olds feel hip again and learn terms they can use while teaching the youths.
- [145:06]
A
That's true.
- [145:09]
C
Anonymous mailer said it was informative. Love the topic and all the information presented. Always learning something new. And bring back the gentleman jingle. It's back.
- [145:20]
B
Thanks to you all.
- [145:24]
A
Just like the Cracker Barrel. You know, when the President speaks, I bring back the jingle. Yep.
- [145:30]
C
Egon cast one said. In today's media landscape, where every podcast promises to change what you think Intelligent Machine actually delivers, mostly by making you wonder if your toaster deserves a college degree.
- [145:44]
A
Each episode is a thrilling blend of.
- [145:47]
C
Futurism and low key panic, like TED Talks if the speaker occasionally admitted, yes, this algorithm will probably replace your job, but look how polite it is about it. It's a really good review. All.
- [145:58]
A
No, that's an excellent review.
- [146:00]
C
Yeah, another one yesterday from Mr. J. Nice. Says topical. Even years later you'd think a show that's been around for so long would run out of steam, but you're wrong. There's no grock suckers here the clankers can hate. But this show brings AI and other ML base news to you every week. Other ones from Mon7ez and Paul VC all good. Thanks guys. And if you want to contribute to our hopefully climbing up reviews, you can leave a review on Apple Podcasts.
- [146:35]
A
Yeah, hey, we appreciate that, thank you. Those are all excellent. Now it's time for Jeff is actually adding picks as we speak.
- [146:46]
C
You already had like nine picks.
- [146:48]
B
Well, it's all of a theme. I haven't done a TikTok corner in a while. And in honor of Salt Hanks stardom.
- [146:55]
A
Yes. As a he just by the way, I'm a little nervous. His mom just texted me he just bought a Rolex so now I'm working worried. Oh, he's he's going and he went to the US Open and he got.
- [147:07]
C
His everyone's going to the US Open that lives in Brooklyn right now.
- [147:11]
A
In fact, I'm surprised Jeff isn't there, to be honest.
- [147:13]
B
It's a whole story. Well, he got his his forehead signed. By who?
- [147:19]
A
Some star. I can show you the Instagram you maybe you would know who it is. I. I had no idea who it is. Oh, maybe it's a. Maybe it was a real Would that mean that it's gone now for forever?
- [147:31]
C
No, it's still on. Is it might have been a story though. Also because I forgot to check the video feed of Intelligent Machine podcast review shout out to Bruce Allen and Dashes who also left us podcast reviews.
- [147:42]
A
Very nice.
- [147:43]
C
Thanks for listening guys.
- [147:44]
A
Very nice.
- [147:45]
B
Yeah.
- [147:47]
A
I don't see it on his his feed but although here he is with Adam Masseri of Instagram making a sandwich. So what? What? I know, isn't that a little strange?
- [148:03]
C
Ask him to Come.
- [148:06]
A
No, he's. I think he's making the prime rib. Oh, yeah, I think he is. Yeah. But I. But I don't see the video of him getting his forehead signed. I don't know. Was it Djokovic? I don't know who it was, but it's somebody. He was jumping up and down with joy to get his forehead signed.
- [148:20]
B
So as long as he doesn't tattoo it.
- [148:22]
A
I hope not. It was. It was interesting.
- [148:25]
B
So in honor of Salt Hank's fame on TikTok and TikTok as a music medium, I want to just quickly mention Anna Lapwood is a. Is a young. She was the head of music at one of the Cambridge colleges and the organist. And she's been so big. She now has left and she's started albums and she sells out the Albert hall with.
- [148:48]
A
Oh, I've. I've watched her tick tock.
- [148:50]
B
She's spectacular. She's a very generous person. Then years ago, you and I compared notes. I. I brought in a singer and you brought in a different singer. I can't remember who yours was, but mine was Levy spelled Lay, who is a Icelandic jazzy singer and she's now a jazz pop star selling out arenas because of Tik Tok.
- [149:14]
A
Because of Tik Tok.
- [149:15]
B
Wow, she's wonderful. And then. But there's all these new things that come up all the time. My favorite is a teacher teacher who happens to be black, who's Scottish, gave a rant online. If you go to line 170, do that first. And it caused quite a stir because.
- [149:38]
A
He'S talking about the weather and Scott.
- [149:39]
B
So just listen to him. Just listen to him.
- [149:41]
A
All right, we'll play it.
- [149:42]
B
Explain to me why Scottish holidays are.
- [149:45]
A
A bit of a washout. Right?
- [149:46]
B
It's been. It's been all right this year, but it's been a bit of a washout. And then we get to the week the kids go, go back. So this, this led to people.
- [149:54]
A
I have no idea what he's talking.
- [149:56]
C
About because I can never hear Scottish. Understand Scottish.
- [149:59]
B
I love Scottish accents. I love Scottish.
- [150:01]
A
There's a very famous bit by a Scottish comedian talking to the. The elevator. The voice recognizer in the elevator, again, doesn't understand a word he said.
- [150:10]
B
So people went berserk over this. There's a Guardian story there.
- [150:13]
A
Because he's black.
- [150:14]
B
Because he's black and he has a Scottish accent. He's a black Scot.
- [150:16]
A
Well, so, so.
- [150:18]
B
Well, Roywood Jr. Went there, line 169 and met up with a Whole bunch of black Scots.
- [150:24]
A
Oh, I love it. Why wouldn't there be?
- [150:27]
B
He interviews this guy. Comments are worse than the black Scottish people because they're going, how can you be so stupid to not know that these black Scottish people. But Americans live in, like, an echo chamber, whether the global media inadvertently or not, decides to. Guys, from seeing us, like, you're not seeing content made by black Scottish people. You're not seeing content made by white Scottish people. So the surprise is to see people that look like you with an accent like mine. That would blow my mind anyway.
- [150:54]
A
Well, I think also, it's also indicative of just how much entertainment influences our stereotypical beliefs. Because if you told me to a Scottish person, I would never in a million years. Like, if you said a black Brit. That's not a surprise. We've seen enough of this.
- [151:07]
B
We got Dave, we got Skepto. We got everybody go.
- [151:10]
A
Yeah. So it. It starts there, right? And that's a good point. That's a good point.
- [151:15]
C
I love. Wood says in one of his posts, Wood made the point that part of the reason why black Americans don't know about black Scots is because American schools barely teach them about black people in America.
- [151:26]
A
That's true. True that. Unless all the time.
- [151:29]
B
So let me try to. Let's try to make another star here. I Love this guideline 173. Carpool.
- [151:33]
D
Shakespeare.
- [151:34]
A
Carpool.
- [151:36]
B
He's not a Scott.
- [151:37]
A
Oh, but it'd be great if it were a Scottish carpool. All right, here we go. Go. Let's see what he's got to say. Carpool. Shakespeare.
- [151:46]
B
This day is called the feast of Crispian.
- [151:50]
A
He that out lives this day and come safe, stand the tiptoe when the day is named and rouse him at.
- [151:56]
B
The name of Crispian.
- [152:00]
A
He that shall live this day and see old age will yearly on the vigil feast his neighbors and say, yo, Henry the Fifth. Yeah. Hank Sank. Very famous. He. He strip his sleeve and show his scars and say, look, these wounds I.
- [152:18]
B
Had on Crispin's day.
- [152:22]
A
Old men forget, you know, yet all shall be forgot. But he'll remember with advantages what feats he's urging his troops on. They will say, you. I fought with you on Christmas. Names familiar in his mouth is household words. Right. Harry the King, Bedford and Exeter. Warwick and Talbot. Salisbury. Yeah. Salisbury and Gloucester. Being there flowing cups fresh.
- [152:51]
B
Isn't this great?
- [152:52]
A
Yeah.
- [152:53]
C
There are kids everywhere for those of the guys to see them.
- [152:58]
A
And Crispin Crispian shall never go by from this day to the ending of the world. But we in it shall be remembered.
- [153:04]
B
Remembered.
- [153:05]
A
We few, we proud. We, we band of brothers. For he today that sheds his blood with me shall be my brother. Be never so vile. This day will gentle his condition. And gentlemen, in. In. In England now a bed shall think themselves accursed that they would not hear and hold their man who hood's cheap. Whilst any speaks have fought with us on Saint Crispin's Day. That is awesome.
- [153:41]
B
Isn't that just the greatest?
- [153:42]
A
That is wonderful. Yeah. That's like, one of the best Shakespeare speeches, too. That's a great.
- [153:46]
B
It is. So this is. This is Tick Tock. TikTok makes Hank and Anna Lapwood and carpool Shakespeare. It's a beautiful thing. And one more. One more.
- [153:55]
A
In theory, Tik Tok is the. The ban on TikTok will into effect in just a couple of weeks. But the president already indicated. Yeah, I don't. I don't think so.
- [154:04]
B
So it also makes one last thing.
- [154:05]
A
Yes.
- [154:06]
B
Podcasting dogs.
- [154:08]
A
But of course, he said I need obedience training.
- [154:13]
B
She's been getting her life together since the pandemic started. She wants me to bring the ball.
- [154:21]
C
Back to her so badly.
- [154:22]
D
I don't understand why she keeps throwing it away.
- [154:25]
A
Like, make up your mind.
- [154:29]
B
Real talk.
- [154:29]
A
What's with making us do tricks for food? That's what I'm saying. Like, bro, send me that slice of cheese. I'm not trying to spin right now. So she says my nails are getting.
- [154:39]
B
Too long and has to trim them.
- [154:40]
A
Bro, she literally makes her nails longer on purpose. Are these. So is that all AI or are they real?
- [154:48]
B
Yeah, there's a whole bunch. There's. There's a. There's also.
- [154:50]
C
If you look up, are the voices AI generated, though?
- [154:54]
B
I don't know. It's a good question. We'll also look up TikTok babies. Just do it.
- [154:58]
C
Oh, no. I have seen TikTok baby. My mother is obsessed with this.
- [155:01]
B
They're freaky.
- [155:02]
C
Yeah, it's strange.
- [155:04]
A
I know.
- [155:05]
C
I think these are fun.
- [155:06]
A
I. I've always thought that it was. No, that's definitely not Tick Tock babies. Okay. I'd always.
- [155:13]
C
We can't fall down this rabbit hole.
- [155:15]
A
No, especially because. Oh, there we go. There's the actual Tik Tok babies. Cute babies all over the world. And they zero babies.
- [155:22]
B
They're. They're AI babies.
- [155:23]
A
Most of these are not babies.
- [155:26]
B
At least maybe they were babies in front of the baby. And the baby talks.
- [155:33]
A
Yeah, that's what.
- [155:34]
B
I should have found those before you, but yeah.
- [155:37]
A
Hey, it's mostly scantily. Clad young women.
- [155:41]
B
Gas.
- [155:43]
C
Googas guy is Google.
- [155:45]
B
Not Google.
- [155:46]
C
Google.
- [155:48]
A
G Gaws. Gas. Yeah, gas. I think you're right. It's GE Gos. But it's not. What did he say?
- [155:54]
B
Two of them said gas.
- [155:56]
A
Yeah, I think that's weird. I literally. This is the problem. Now that I've searched for tik tok babies and found a bunch of scantily clad young women, I will be. I will be seeing that in my Tik Tok feed for the rest of my life. All right, let's see.
- [156:11]
C
Will you bring me the toy over. Thank you.
- [156:13]
B
Oh, yeah. This is nice.
- [156:14]
C
Some of these are profane.
- [156:15]
D
I.
- [156:16]
A
No, I don't want to see it. That's real bab.
- [156:19]
B
None of these. None of these are the right ones.
- [156:20]
C
Let's not. Let's not show any videos of children we haven't prescreened, you know.
- [156:24]
A
Good point.
- [156:25]
B
That's very AI Babies.
- [156:27]
A
AI babies are.
- [156:27]
C
Okay, we'll just keep doing this.
- [156:29]
A
Maybe if I put AI Babies you gone.
- [156:34]
B
Yes.
- [156:35]
C
I will say my understanding of the AI Baby thing is the things the AI babies say are profane, so I don't know if they would be the best fit for the show.
- [156:45]
A
We're just gonna stop right here.
- [156:49]
C
John has turned his camera on, which indicates something to me.
- [156:53]
A
John, what do you. What do you want to say to us? John Ashley, producer. I feel like vacation John's filling in. Yes.
- [157:00]
B
I feel like this is reminiscent of me on Maverick, where I would put. Probably play the rat hole sound effect.
- [157:05]
A
Yeah. Well, we're ready to go.
- [157:07]
B
Oh, everybody's a critic.
- [157:09]
A
Oh, well, it is hot up here in the attic, and I am ready to go.
- [157:13]
B
Hey, Paris was warning everyone. Let's not get into this rattle.
- [157:16]
A
Essentially, yes, she is. She's warning us we should listen because she's got good sense. You corrected my admittedly flawed pronunciation with something that's just not even in the ball park. That's what. I don't know.
- [157:33]
B
The thing is, no one ever knows how to spell.
- [157:34]
A
It's G, E, E, G, A W.
- [157:38]
C
No, I never heard this word into until today.
- [157:42]
A
It's a showy trifle.
- [157:45]
C
It is a showy trifle.
- [157:46]
B
G, E, W. G, A W. Gaw.
- [157:48]
C
No, it's G. Yeah.
- [157:49]
A
G, E, W. It's a less common variant of googa. Never heard of Google. I think this is AI pranking me.
- [158:01]
C
Oh, googa.
- [158:04]
B
Because I used this in a book and the copy editor then corrected me. I thought, that can't be right.
- [158:08]
A
But it was guga. It's Jija is a less common variant of googa. No, it's not. It's gaw. That's not less common than googa. Googa isn't even a word.
- [158:23]
B
It's not a soft.
- [158:24]
C
I love. Okay, let's go through the synonyms list. Here on Merriam Webster we have normal ones like bauble, chotchki, knickknack. Then we get to gym crack, bibla.
- [158:35]
A
I'm gonna use these. Curio. It's a knickknack. Doodad. Collectible.
- [158:42]
B
Click on the all synonyms.
- [158:44]
C
Gym, crack.
- [158:46]
A
There's got to be trumpery.
- [158:48]
B
Trumpery.
- [158:48]
A
It is a objet, a trifle, a figurina, trinket, bagatelle, bric, a brac, trinket, trinketry, bijouterie, Jim crackery.
- [159:03]
B
See, we go down one rat all the way down another one.
- [159:08]
A
All right, stop it. Stop it. I got these. These, these. My. My avocado wants to go wilting. Everybody wants to go home. I want to go home. Paris Martineau, her home is there in beautiful Brooklyn, California.
- [159:25]
C
It's true.
- [159:27]
B
Just look for the well trimmed trees and you'll find Paris nearby.
- [159:31]
A
You will find Paris in the soapbox. Paris chopped into pieces. And then what do they. They bring them home for their home fires? What do they. They just chop them up and leave them in the street? I happen to.
- [159:44]
C
I assume that they're carted away at some point, but I've never stuck around that long to figure it out. Or perhaps they put them in a large pile that they then reuse next year. It probably can't be that. Where would you store that much loose.
- [159:57]
A
Wood and G jaw trinketry?
- [160:01]
C
There was a red.
- [160:03]
A
Professional investigative journalist at the incredible Consumer Reports. We are so glad you are with us.
- [160:08]
B
This is her amateur moment here.
- [160:09]
A
This is where she pretends not to have class.
- [160:13]
C
Hey, shout out to all the nuclear adjacent researcher people who are. Who actually did email me after about a nuclear shrimp on Twitter this week with some interesting shrimpspiracies.
- [160:26]
A
Oh, and by the way, Clinton is. The Clinton nuclear power plant is in Bloomington, Illinois. Not.
- [160:34]
B
Oh, sorry.
- [160:35]
A
Okay.
- [160:35]
B
There is more than one Bloomington.
- [160:36]
A
There is a Bloomington, Illinois, and that is one of two nuclear reactive.
- [160:41]
B
He looked like he was Indiana.
- [160:43]
C
Yes, well, don't go spreading false shrimp spiracies, guys.
- [160:47]
A
He might be from Indiana. How about that? I bet that's Jeff. Jeff Jarvis. We already gave him all that he deserves. He is an Illinois snob.
- [160:56]
B
He was born in Illinois.
- [160:58]
A
Oh, Burke figured out what happens to those old chopped up soapbox derby vehicles. They're used to make hot dogs.
- [161:06]
C
Oh that's great.
- [161:07]
A
And that's where the Coney island dog came from. Many people don't know that. Thank you everybody for joining us. We do intelligent machines. God only knows why. Every Wednesday 2pm Pacific, 5pm Eastern Eastern 2100 UTC watch us live on YouTube, Twitch, TikTok, Facebook, LinkedIn, X.com and Kick. Oh of course if you're in the club you can watch us behind the velvet rope in the beautiful club Twit Discord opulently laid out with free champagne for all after the fact on demand versions of the show at TWiT TV IM. There is a YouTube channel dedicated to it but the best thing do to do is subscribe in your favorite podcast client and leave us a great review in the five stars. And if you do it on the Apple Podcasts, Ms. Martineau may pick up on it and read it someday on our show.
- [161:56]
C
It's true.
- [161:58]
A
Thank you so much.
- [162:00]
C
You know what we say on this podcast?
- [162:02]
A
What do we say?
- [162:03]
C
And good night.
- [162:06]
A
Good night famous Edward Rugan Goodbye, Good.
- [162:11]
C
Night Good Good Morning Guga Good evening and good night.
- [162:19]
A
So long everyone. No matter how much spare time you have, TWiT TV has the perfect tech news format for your schedule. Stay up to date with everything happening in tech and get tech news your way with TWiT TV. Start your week with this Week in Tech for an in depth, comprehensive dive into the top stories every week. And for a midweek boost, Tech News Weekly brings you concise, quick updates with the journalists breaking the news. Whether you need just the nuts and bolts or want the full analysis, stay informed with TWiT TV's perfect pairing of tech news programs. I'm not a human being.
- [162:58]
C
Not into this animal scene. This episode is brought to you by Progressive Insurance. You chose to hit play on this podcast today. Smart Choice make another smart choice with.
- [163:12]
A
Auto Quote Explorer to compare rates from.
- [163:14]
C
Multiple car insurance companies all at once. Try it@progressive.com Progressive Casualty Insurance Company and affiliates. Not available in all states or situations. Prices vary based on how you buy.