Radio Better Offline: Allison Morrow, Paris Martineau & Ed Ongweso Jr.

Also media.

Hi everyone. Before we get to the episode, I just wanted to lead in and say we are up for a webby. I'll be including a link. I know it's a pain in the as to register for something. I'm sorry. I really want to win this. Never won an award in my life. It will be in the links. And while you're there and registered, look up the wonderful weird little Guys with Miss Molly Konga. Vote for both of us. I'm in the Best Business Podcast, Episode one. She's in

I'd be happy to. I wrote a piece. I think the headline we ended up with was like Apple's AI is not the disappointment. AI is the disappointment. Yeah, And this was inspired by credit to where It's due. I was listening to hard Work with Kevin Ruse on My husband and I were driving out to the country and

So such a funny argument, given that it was a mistake being made by Apple that resulted in the whole hoo. They PC small group situation that was specifically how the editor in chief of the Atlantic ended up in a secret military signal chat.

Wait I missed what? How?

How online? I? Oh, gosh, I should leave.

I've been reading scrolls.

So basically, The Atlantic came out a couple of weeks ago with a article about how their editor in chief one day was suddenly added to a.

Signal group chatigny signal game. But how did this?

So the apple thing was I'm forgetting who exactly reported this? This was in the last couple of days. That how it happened was the con like you know that thing that comes up in your iPhone where it says like, oh, a new phone number has been found. It was a

Well, in this particular instance, I take it all back. Apple. AI is a mea. It gave us one of the best journalism stories.

Of the year.

You also made a really good point in here as a message you this on the way, and you said you made this point that there's a popular adage in part of the Sea Circles that the party can never fail, it can only be failed. It is meant as a critique of the ideological gatekeepers who may be, for example, blame voters for their parties failings rather than the party itself. The same fallacy is taking root among AI's big as back as AI can never fail, it can only be failed.

Yeah? He said, what that you have to prove your job can't be replaced by AI?

Yeah, else it will be.

And he also said that now it's going to be Shopify policy to include in all of the employee performance reviews, both for your self assessment and for your like direct reports and colleagues assessment, how much this person use AI. And obviously what's going on there is if you are not reporting that you use AI all the time for everything, you could get fired.

Whyn't I just tried to overhaul hiring presses so that they could have AI first or AI only, and then roll it back because they realize you can't replace yes all these jobs questions?

Is I mean, this is something that's brought up on the show all the time. But who are these people that are encountering the AI assistant suddenly plugged into every app and being like, yeah, this is actually beneficial to my life and this works really well done because it sucks every time I use it.

Or you made the point in your article Allison's words like, if it was one hundred percent accurate, it would be really useful. If it's even ninety eight percent accurate, it's not right.

I think that was the point that you know, to his credit, Casey Newton made in the episode, which is that AI is fundamentally an academic project right now. And it's like, yeah, you can have all the kinds of debates about its utility, but ultimately is it a consumer product? And no, it's just like it's failing as a consumer product on all fronts.

And what's crazy as well, is I'm surprised he would say that considering everything else he's ever said, because he's he quite literally has had multiple articles recently being like consumer adoption is up in He had an article the other day where it was like, data provided exclusively from Anthropic shows that more people are using AM. It's like, mom, man, it's twenty thirteen. Again, we're past this. You can't just

I mean, in one way, that's great, right. The only place you should encounter it is maybe the team building retreats, you know, that's the utility of this shit.

This reminds me a lot of like Media twenty twelve twenty thirteen, where it was all pivot to video and what's our video and our vertical video strategy? And it's like, okay, now, what's our AI strategy? How are we injecting AI into everything we're doing? And it's like, well, to what end? Yeah, this is just the point.

This is something that has been driving me mad, especially with partnerships we're seeing between media firms and these AI firms. You know, these are these are firms in the same sector that keeps lying to verbs about how if you integrate artificial intelligence this time it will optimize your ability to find an audience or to get revenue, and we can include you in some esoteric revenue share program where we'll be able to claw back some of the eyeballs

Right.

What's insane is this company has a multi billion dollar market cap and I'm just going to read point two. AI must be part of your GSD prototype phase. The prototype phase of any GSD product should be dominated by AI exploration. Prototypes are meant for learning and creating information. AI dramatically accelerates this process. How fucking how Like That's

Yeah, yeah, because it feels I mean, we've had so many of these hype cycles, kind of back to back to back, from even the horizontal video days of Facebook, to vertical video to whatever the hell the metaverse was supposed to be. Literally in an ominous moment, as I was walking into record this, I saw a guy wearing a leather jacket with board apyacht club on the back, and I was like, god, well, yeah, I was like that that guy rocks.

What a cool dude.

But it's like, how long is this going to last?

I have been actually looking at the numbers recently, and I don't know either, because for soft Bank to fund open AI might require them to destroy soft Bank, like S ANDP is downgrading the credit rating potentially due to yeah, yeah, I know, like we're really at this point where it's just like we've got we've gone so much further than like the Meta US and Crypto did, because those weren't really like systemic things, but this one, I think it's

I mean, in some elements it kind of reminds me at the near the end or near the real peak is when we started also to see Metaverse and Crypto Sustainable Refine Ship where they're actually, you know, we can fight climate change with crypto, yeah, putting carbon credits on the on the blockchain. And so there was a there was a moment where the frenzy and the speculatve frenzy led to like world transformative visions that were bullshit, And

Yeah, I was thinking about this in a kind of like fundamentally cultural way, where I think at some point in the last thirty years, there was a time when everything coming out of Silicon Valley was cool. Yeah, whether it was like useful or world transformative, it was cool, and there was like an edge to it, and people were like.

Ooh, that's neat, it's disruptive.

Yeah, disruption was everything, And like I think posts like Facebook, Cambridge Analytica era, like twenty sixteen tech has just stopped being cool and edgy. It's very corporate, and like, I don't think the rest of corporate America has kind of figured out that Silicon Valley's not the cool thing anymore.

And that they can't lot. They're fully capable of being wrong and lying, like that's that's the other thing.

They've gotten very good at fundraising and marketing.

But they're also not like kids anymore. Like we talk, I still see people referring to Opening Eye as a.

Startup, former Lucky as a kid.

Palmer lucky as a kid who looks like leisures Larry and sells arms, which is.

Gy. We refer to them as startups. But also I think one of the most accomplished parts of AI marketing has been, like we always refer to them as labs, so they seem like so academic and like good fundamentally, and it's like these are companies, Like some of them might be part of, uh, you know, a research institution or a university, but a lot of them are startups.

Lateral companies.

Yeah, they are companies.

Like anthropics, public benefit, I believe, And it's it's just remarkable, And I think what's happened here is that the narrative has gotten away to the point that we're really dunce mask off moment I mentioned is people like mister Ludke from Shopify. It's very clear he doesn't do any work. Like I think that anyone who is just being like, yeah, AI is the future and it's changing everything without specifying anything, doesn't do any work. I just don't. Bob Iger from

Walking around Diagram is increasingly becoming a circle.

Yeah, but he's onto quantum now, which is a bad sign. That's a bearish sign. When you've got one of the Forbes guys moving on to quantum, we're cooked.

What about thermo, Yeah, there's thermodynamics.

Become a thermodynamics influencer. I know what I know what that means. I also know what that means. But if anyone could tell me real quick. But it's I think the most egregious one I've seen. I sent this all to you, Ed, I think you and I have talked about this the most. There was one of the stupidest fucking things I've read in my worthless life, and it's called AI twenty twenty seven. Now, if you have not run into this yet as a listener, it will be

It is a little thing.

Throughout it I was like, is this fan fiction? This is fan fiction? Oh, this is interactive fan fiction.

It is, and you can hit the bund says what is this? How do we write it? Our research on key questions what goals will future AI agents have can be found here. The scenario itself is written iteratively. We wrote the first period up to mid twenty twenty five in the following period, et cetera, until we reach the ending Yeah, otherwise known as how you write stuff like you've wrote it writing anl in their fashion. We then scrapped this and did it again. You should have scrapped

Over the next decade. It go basically says it's going to be have superhuman like catastrophic or world changing impact in the next five years, like by twenty twenty three, we're either going to be completely overtaken by a robot overlords or like at a you know, tenuous piece.

And it's insane as well, because it has some great headlines like mid twenty twenty six, China wakes up.

And then I love that China was so far behind, you know, it's like when.

Did this come out?

This came out like a week ago, and I've been sent it by a lot of people. If you're one of the people who sent it. Don't worry, I'm not mad at you. It's just like got sented by a lot of people. This thing is one of the most well written pieces of fan fiction ever in that it appears to be like a Manchurian candidate situation for idiots, not saying this is about the same thing, but anyway,

Even an accurate description of all of the storms it predicts.

And the long and short of this, by the way, I have read this a few times because I fucking hate myself. The long and short of it is that a company called open Brain. Who could that be? Yeah, it could be anyone, anyone open Brain. They create a self learning agent somehow unclear how all layout is just that, how many are terror flops it's going to require, and it can train itself and also requires more data centers than ever, how they get them, how those are funded?

I do also want to add in here that you know behind the AI in twenty twenty seven is you know, one of the people connected to it, if I remember correctly, is Scott Alexander. Who's this guy that's part of the rationalist community, which is one of the what are the groups that overlast with the effect of altruist acceleration this Yeah,

And it's written to be worrying. Yes, well it's written.

It's written to be worrying. But it also in the productions for the next two years keeps talking about how the stock market is going to grow exponentially and so well is going to be making all of these wise and formed decisions and having really deep conversations with the leader of Open Brain, and I was like, are you That's why I asked when did this come out? Because I was like, maybe this was written a couple of years. No, but no, it is why Like literally.

It's like a Benjeser kind of planning gonna be born. He's gonna lead us to the promise.

It's so good as well, because the people who sent this to me have been very concerned, just because they're like, this sounds scary, And I really want to be clear. If you read something like this and you're like, that doesn't make sense to me, it probably doesn't make sense to anyone because it's nonsense. This thing is. Let me read you one cut from it, the AR and D progress mark player. What do we mean by fifty percent

I do really love all of the incredibly, So.

This guy is just I'll put the link in there for this.

He got seventy five pockets on his truck.

Yeah, my man is read.

Oh that's that's where it is.

And it's just him Like this guy's sitting with his hands classed like, staring mournfully into the distance. This is what you're spending your time on ket And I'm just going to read some Kevin Ruce. The II prediction, well, this tone between uptimism and gloom. A report released on Thursday decided lands on the side of gloom. That's Kevin Ruce's voice. But my favorite part of this by far, I'm gonna take a second to get it, because Ed I sent this to you as well. Where is it?

So?

Also a lot of this is oh, here we go. If all of this sounds fantastical, well it is nothing remotely like want mister Cocker to tell Joe and mister Lyffland are predicting is possible with today's AI tools, which can barely order a Buruto on Dowdesh without getting stuck. Thank you, Kevin. I'm so fucking glad the New York Times is on this.

And that was at the end. Yeah, right, Like you set up this whole article. And it's like these guys have these doom predictions. And that's the other thing about the like the altruistic AI guys all have told themselves this story and they all believe it, and they think they are like the Prometheus bringing fire to the people and like warning the people.

And it's like, you.

Guys have sold yourself a story I with no proof.

I don't know. I feel like they just scam out to that. Nothing about this suggests they believe in anything. You can just say stuff.

Look it literally the second sentence in this is that in two months there will be personal assistance that you can prompt them with tasks like order me a brito on DoorDash. He'll do great stuff. There are so many things that go into ordering me a burrito on DoorDash. What restaurant do I want? What burrito do I want? How do I want it to get to me? Where am I? It can't do any of those things, nor will it.

He gazed out the window and admits that he wasn't sure. And the next few years went well, and we kept aie on the control, he said, referring to one of the writers of the piece. He could envision a future where people's lives was still largely the same, but where nearby special economic zones filled with hyper efficient robot factories would churn out everything we needed. And if the next few years didn't go well, maybe this guy would be

You know, one of the things I really really love about I don't know, It's just it's so frustrating because we're we're constantly fed these you know, sci fi esoteric futures about how AI powerful AI, superhuman AI is around the corner, and we need to figure out a way to accommodate these sorts of futures we need. And part of that accommodation means restructuring the regulations we have around it. Part of that accommodation means entertaining experiments, grafting them onto

Yeah. One where you need to believe. You need to believe it.

You need to do your part, you know, to summon God.

And that's the thing. It goes back to what you're saying. It's like you've failed AI by not believing.

Yeah, and if you're bad at it, it's your fault and not the machine's fault. In to Ed's point, I think like all of this like predicting of the future and this like revolution is like they have told themselves a story that is this is inevitable and that there are no choices that the human beings in the room

And on the subject of human choice, if this shit is so powerful, why have them mighty human choices not made it useful yet? Like that's the thing. It's and you make this point in your piece as well. It's like a I can have a fail. It can be failed failed by you and me, the smooth brained bloodites who just don't get it. And it's like, why do I have to prove myself?

And listen, you know the ladites they have more goo on their brain that coverant. So he needs to I think it's it's worth embracing a little bit, you know.

Which has already been around us and is generative.

Yeah, and it isn't generative, but they but the firms that are offering these generitive products are spending actual you know, the stuff that they're actually putting their time and energy into is you know, the sort of demonstrably destructive tools under the guys in the kind of murky covering of it's all you know, artificial intelligence, right, it's all inevitable, it's all coming down the same pipeline you should accept it.

Yeah, and it's I think the thing is as well. It's those guys really think that's the next big money makeup. I don't think anyone's making any money off of this. No one wants to talk about the money because they're not making any like no one. Look, I think I've read, I've read the earning schools. I'm not going to listen of every single company that's selling an AI service. At this point, I can't find a single one that wants to commit to a number of the Microsoft and they'll

No, that's the classics.

I refuse to mark client's asses to the ground with that one, because it's like, you can't just fucking make up a number unless you're in AI than you absolutely can. It's just frustrating because and the reason I bag on Newton Rouse other than all the others I've listed, is the I feel like, in their position and in the position of anyone with any major voice in the media, skepticism isn't like something you should sometimes bring in. It's not.

I mean, I think part of it is also just because investors have poured a lot of money into this, and so of course they are going to want to back what they have spent hundreds of millions or billions of dollars on. And much of the tech media involves reporting on what those investors are doing, thinking and saying, and whether or not what those people are saying or doing is it's often not based in reality.

H yeah, I say, as not a member of the tech media, so I have like kind of a general assignment business markets econ. That's kind of my jam. And when I come when like AI first started becoming the buzzword, like chat GPT had just come out, I was like, oh, this sounds interesting. So I was paying attention like a lot of journalists were, and you know, like we've hit limitations, and I think part of the reason it's gotten so far is because the narrative is so compelling. Curing cancer,

Nice, okay?

How how Also, the problem of hunger in the world is not that we don't grow enough food. It is a distribution problem. It is a sociologic It is a complicated problem. What actually is AI going to do? Also that you're going to need human beings to distribute it? It just like you push them one step.

If you read the twenty twenty seven AI thing, it explains that the AI is going to give the government officials such good advice they'll be actually really nice and caring deportment.

And what's crazy is here's the thing. And I'm glad you brought up one thing I've learned about politics, particularly recently, but in historic historic means, when the government gets good advice, they take it every time, every time. Every time they're like, this will this is economically good, like Medicare for All, which we've of course had forever and never and came close to numerous times decades ago, versus now when we

Nah?

Okay, it's going to cull hunga Can it do that? No? Okay, it's easy. Then perhaps it could make me an appointment. Also, No, can you buy something with it?

No?

Can it take this spreadsheet and move stuff around?

Maybe it can write a robotic sounding script for you to make the appointment yourself.

Wow, you know, I mean I would even say that I could give the benefit of the doubt to researchers who are really working on the scientific aspects of this. Like I'm not a scientist. I don't know how to cure cancer, but if you're working with an AI model that can do it, like God bless. But businesses actually do take money making advice and money making technology when it's available. And I think about this all the time. With crypto, which is another area I cover a lot.

And it's the same with.

Crypto and a lot of other tech where it's just like, this is not a product that people are are hankering for.

And I think part of the notable thing is when we do see examples of large businesses being like, oh yeah, we're gonna change everything about our business and integrate AI. We're going to be an AI first company. The products that end up coming out of that are there's an AI chat bot in my one medical app now cool

The same product every fucking It's.

Just an a It's a chatbot that isn't super helpful.

And it's great. I remember back in twenty fifteen twenty sixteen, I had an AI chatbot company. They took large repositories of data and turned it into a chatbot you'd use. I remember pitching reporters at the time and then being like, who fucking cares, who gives a shit? This will never be Like a decade later, everyone's like this is literally God. I cannot wait to go to the office of a guy who wrote fan fiction about this and talk to him about how scared I am. Now I can't wait

Guy who was like the chat but is real and I love it.

Yeah, had that come out three years later, he'd be called Kevin Roos because that's exactly what Kevin Rooss wrote about being Ai Savria is like being AI told me to leave my wife. And Kevin, if you ever fucking hear this, man, you're worried about me dogging you, I'm gonna keep going do your fucking job. Mate. Anyways, it's just insane because I am a gadget gizmo guy. I love my dude, dad's I love my shit, I really do. If this was gonna do something fun, I'd have done it.

Who cares because you haven't tried claude yet.

I've tried. I've tried Claude so much, And it's just I don't know. I feel like people's response to the media is going to be negative too, because there's so many people that boosted it. There was a Verge story. There was a study that came out today. I'll link it as well in the notes where it was a study found that most people do not trust A like regular people do not trust AI, but they also don't trust the people that run it and they don't like it.

Yes, yes, well I think that gets to you if you just push the proponents. And this is like I've I've come into AI skepticism as a true like I'm interested in this. I'm interested in what you're pitching to the world. And when I hear and I hear like CEOs of AI firms get interviewed about this all the time, and they talk about this future where everyone just has a life of leisure and where you're lying around writing poetry and touching grass and like everything's great. No one

Get what it jobs? They get oddly specific on some things. Then the meaningful things they're like, yep, they'll be jobs.

Yeah, and the stock market's just gonna go up and it dude.

Numbill go up all the time as it is right now as we report.

Yeah, I believe they say. In twenty twenty eighth, Agent five, which is the super AI's deployed to the public and begins to transform the economy. People are losing their jobs, but Agent five instances in the government are managing the economic transition so adroitly that people are happy to be replaced. GDP growth is stratospheric, Government tax revenues are growing equally quickly, and Agent five advised positions show an uncharacteristic generosity towards the economically dispossessed.

You know what this is We failed to uphold like public arts education in America, and a bunch of kids got into coding and know nothing but computers, and so they can't write fan fiction.

Yeah, No one's fighting is bad.

Not enough people spent time in the minds of fanfiction dot net and it's no one's anyone. Yeah, Like this is.

Clearly this is just like someone wanting to have a creative like vision of the future and it's like it's not interesting or compelling.

Joy looks.

I mean, that's why they brought him on. That's why they brought Scott Alexander on because to write this narrative, right, because that's what he spends a lot of time doing in his blog, is trying to beautify or flesh out

I just feel like a lot of the AI safety guys at Grift is too. I'm sorry they love saying alignment just say pay me like I know that we should have. I get the occasional email about this being like, you can't hate AI safety. It's important. It is important generally. If AI isn't AI, it's just trying to fucking accept it. They're not. If they cared about the safety issues, they'd stop burning down zoos and feeding entire lakes to generate

Do you think that one criticism of the AI safety initiatives that is incredibly politically salient and important right now is that they are so hyper focused on the long term thousand, one hundred years from now future where AI is going to be inside all of us and we're all going to be, you know, robots controlled by an overload, that they are not paying attention to literally any of the harms happening.

Right well, they deliberately not talking about the harms today because then they'd have to do something at work when they men each that.

You know, It's like when nine seven two mag reported on how its reel was using or trying to integrate artificial intelligence until generating gets killed thiss and targets so much so that they started targeting civilians and use that

But that's that would require the AI safety people to do something. And what they do is they get into work. They're making quarter of a million dollars a year. They get into work, they load Slack, they load Twitter, and that's what they do for eight hours, and they occasionally posting being like by twenty twenty eight, the AI will of fuck my wife, and everyone's like, God, damn it, no,

It's just like the like also sick that they renamed themselves after it's so good.

Yeah, it's bad.

I can't go about Yeah, like they should get METAI. They should just change Oh.

They should add an eye, they should just add and I.

At the end, it's just if anyone talks about what's actually happening today, which is boardline identical what was happening a year ago. Let's be honest. It's April twenty twenty five. April twenty twenty four was when I put up my first piece, being like, hey, this doesn't seem to be doing anything different. It still doesn't even with reasoning.

It's just not just a way for Q three. Agent Force is gone.

Yeah, agent the Agent zero is going to come out. Yeah, Actually the information reported that Salesforce is not having a good time sailing Agent Force. You never guess why. Wow, turns out that it's not that useful due to the problems of generative AI. If only someone had said something which the information. I've begged on the information a little bit. But they are actually doing insanely good work like Corey Weinberg, dur Alisa Gardizzi Paris of course, but I'm specifically talking.

About AAI team as best.

Of course, and like it's great because we need this reporting for when this just collapsed, so that we can say what happened, because it's going to if I'm wrong, and man, would that be embarrassing, just gonna be honest, Like, if I'm wrong here, I'm gonna I'll look like a huge idiot. But if I'm right here, like everyone has overleveraged on one of the dumbest ideas of all like silly, silly.

What if we have an Ubertionnario where maybe they abandoned the dream of like this three trillion dollar addressable market that's worldwide. They abandoned the dream of like being monopoly in every place and and focus on a few markets and some algorithmic price fixing so that they can figure out how to juice fares as much as possible, reduce the wage as much as possible, and finally eke out that profit.

What do we see?

You know, some of these firms they pull back on the ambition or the scale, but they persist and they sustain themselves because they move on to some smaller.

Vision like Oham's razor. That's the most likely situation is that you know, AI tools are useful in some way for some slice of people and make a lot of maybe maybe let's be optimistic, it makes a sizable chunk of a lot of people's jobs somewhat easier. Like it's great, it was that worth spending billions and billions of dollars and also burning down a bunch of trees, saying that could be I think best case scenario.

No, I'm not saying you're wrong. I'm just saying like, we haven't even reached that yet. Because with uber it was this incredibly lost and remains quite a losty business, but still delivers people to them from places and objects from to them from places.

You know you don't have to, you know, as much as I hate them, I'll go, okay, you know you don't have that. Not less drunk driving, you know, and some transit in parts of cities where there's there's a there's there's not much in the way of public transit. Right.

This is like if uber, if every ride was thirty thousand dollars in every and every car weighed one hundred thousand times in.

Manu factor in the externality.

But that's the crazy thing. I think general if AI is so much worse as well, polution wise. But even pulling that back, it's like, I think open AI just gets wrapped into co pilot. I think that that's literally they just shut the ship. They're like they absorb Sam Moltman into the hive mind and he. I think my chaos pic for everyone is satching the dellar is fired in Amyhood takes over. If that happens, I think it's Prometheus, the one who can see stuff. I'm fucking not done.

I mean the impact on the educational sector, especially with public schools, it's really terrifying, especially because the school districts and schools that are being forced to use this technology, of course, are never the private, wealthy schools. It is the most resource starved public schools that are going to

And that's the thing. And the school could, of course, I don't know, spend money on something that's already being spent, which is teachers have to buy their own fucking supplies all the time. Teachers have to just spend a bunch of their money on the school, and the school doesn't give them money, but the school will put money into chat. JP say, it's just oh, they should ban at universities as well. Everything I'm hearing there is just like real fucking bad.

Like the I mean the issue is from talking to university professors. It's like impossible for universities to ban it. I guess professors are The obvious example is like essays, Like professors get AI written essays most of the time, and they can't figure out whether they are AI written or not. They just notice that all of your students seem to suddenly be doing worse in class well having

But on the.

In K through twelve especially, it's been kind of It's been especially frustrating to see that some of the biggest pushers of AI end up being teachers themselves because they are overworked, underpaid, have no time to do literally anything, and they have to write god knows how many lesson plans and i EPs for kids with disabilities and they can't do it. Also like, well, why don't I just plug this into what's essentially a chat GPT wrapper and that results in worse outcomes for everyone.

Probably, So I have some personal experience with IEP. I don't think they're doing it there, but they're definitely doing it elsewhere. And if you've heard, if you've heard IPA, that fucking kills me.

That's one of the things that these tools often pitch themselves as I want to make sure.

Something, I want to put my hands around someone's.

Fucking Can you describe what an IP is? Is A?

I forget what it stands for Individual education plan?

I don't wrong, but that's it.

Is generally the plan that's put for a child with special needs, so autism being one of most obvious one. It names exactly what it is that they have to do, like what the teacher's goals will be, like socio.

They legally have to do all the things in that document.

And it changes based on the designation they get, and so like it's different if you get there's like an emotional instability one, I believe, and nevertheless, there's like separate ones, and each one is that the goals of where the kid is right now, where the kid will be in the future, and so on and so forth. The idea that some would use chat GIPT and if you listen to this and use chat GPT for one of these, I fucking hate yours so bad. I understand you're busy,

Can I have one thing to what you said about people who don't work? I think there are people in Silicon Valley and I don't I'm gonna get a million emails about this, but there are a lot of Silicon Valley men who are white men who don't really socialize yep, And I think they are kind of propagating this technology that allows others to kind of not interact, Like so much of chat GPT is designed to like subvert human interactions, like you're not going to go ask your teacher or

I think there's also a class aspect to that because totally don't grow up particularly with like add food on the table. But one thing I grew up was I don't trust any easy fixes. Nothing is ever that easy. Is something that kind of a if something seems too good to be true to accessible, there's usually something you're missing about the incentives or the actual output. So no, I wouldn't trust a computer to tell me how to fix something because I don't fucking like that you made

Well, I mean that kind of gets back to what you're saying, where it's like, you know, if we entertained the fan fiction for a little bit, what is the frame of mind for these agents, if they're autonomous or not, how are we thinking of them? Are we thinking about like their persons or if they're you know, the bottomized in someone, do.

They have opinions?

You know?

And I think really it just gets back to like, you know, part of the old hunt for like you know,

You know.

I think similarly, you know, the hunters automatons for certain algorithmic systems is searching for a way to figure out how do we replicate you know, human labor without the limitations on extracting and pushing and coercing.

As much as possible, Yeah, with you know, and there's no agent or something else. And the thing is, yeah, sure, the idea of an autonomous AI system would be really useful. I'm sure it could do stuff that sounds great. They're all massive, as you've mentioned, like sociological problems like do these things feel pain? If so, how do I create anyway? But in all seriousness, like sure, an autonomous thing that could do all this thing would be useful. They don't

It's like they don't even speak to how we're going to get to the point where Agent one knows truth from.

Falsehood less inevitable of course.

Yeah, you know, we just need to give it all of our data and everything that we've paid money for, required other people to pay money for, and then it will finally be perfect.

And it doesn't even make profit of any kind. That's the other thing. It's like people saying makes profit it is the profit seeking? Is it profit seeking? It doesn't seem like we've sought much profit or any Yeah.

That's also I think a good point of comparison to what you were talking about earlier, ed with the comparison to Uber and Lift, these companies that achieved massive scale and popularity by making their products purposefully unprofitable, by charging you five dollars for a thirty minute Uber across down so that you're like, yeah, this is going to be

And how much of that is dependent on what is clearly naked the subsidized compute prices. How much of this is because Microsoft's not making a profit on a Zeo compute, open Aiye isn't making anthropic is then what happens if they need to? What if they need to, they're gonna chuck. That's the Supperme Mai crisis from last year.

It's just it's it's well, that's when you get the venture capitalists insisting that that's why we need to you know, do this air capex rollout because if we build it out like infrastructure, then we can actually lower the compute prices and that subsidized when yeah, that's.

The thing, but that's the other thing. So the information reported the open AIY says that by twenty thirty they'll be profitable. How Stargate Yeah, And you may think what does that mean? And the answer is Starguate has data centers. Now you have to I just have one little question. This isn't a knock on the information. This is their reporting what they've been told, which is fine. A little question with open AYE though, how how does more equal

I can't wait until all of those companies announce that because of the tariffs yep, that they have to delay their timeline and it's like completely out of their hands.

But no, the tariffs.

You understand the tariffs.

I've got a fungth of dine. I got a full roasted pig. I'm going to be tailgating Microsoft earnings April twenty third. Cannot wait. Yeah, you should go to like a data center to have a marching band severance. Yeah, it's but that's the thing like that, I actually agree. I think that they're gonna there's gonna be difficult choices. Sadly, there's only really two one capex reduction, two layoffs or both because they have proven willingness to lay off to

And I talked to an analyst just last week who's like a bullish AI tech investor, and he said, he said, already you're seeing investment pull back because of expectations in the market that there was these stocks were overbought in the first place, and now there's all this other turmoil external macro elements that are going to kind of take the feeling, you know, the jargon of like the froth out of the market. They're gonna it's all going to

What's great as well, talking of my favorite Wall Street analyst Jim Kramer of CNBC, so Court Weave's IPO went out. I just need to mention we are definitely in the hype cycle because Jim Kramer said that he would sue an analyst Da Davidson on behalf of Nvidio for claiming that they were a lazy Susan as in as in basically, what the argument is is the Nvidia funded care Wave, so the core Wave would buy GPUs, and at that point core Weave would then take out loans on those

And I before it even came out, I like downloaded the beta. I was like, I'm going to test this out because you know, I talk about the thing on my podcast sometimes and it's so bad bad. I like turns it off for most things, but I have it on for a couple of social networks, and I mean, I guess with the most recent update got marginally better, but it still constantly tells me so and so replied to your blue sky ski. I checked they didn't. That person didn't even like the ski. I don't know where

I'm like, how my favorite is the summary takes for Uber where it's like several cars headed to your location Kallennick Klannick mode activate it. No, it's it's great as well, because I usually don't buy into the Steve Jobs would burst from his grave thing. I actually think numerous choices Team cook has may have been way smaller than how Jobs sort of done. This is actually like he's going to burst out of the ground thriller style. Actually did

And I mean it's also small other things beyond just the notification summaries. The thing that every time I highlight a word and I'm trying to sometimes might want to use fine definition or any of the things that come up, I have to scroll by like seven different new options under the right click or double click.

If you hit right tools, it opens up a screenway, Yes, it opens up a.

Thing, and I'm like, who has ever, who is trying to use this to rewrite a text to their group chat.

Who is this for?

I feel like Apple, to its credit, is recognizing its mistake and it's clawing it back and like delaying Siri indefinitely.

I mean, I don't know if I agree on that one, because the thing they're delaying is the thing that everyone wanted. I think they can't make it work because the thing that delaying is the contextually a wes are right.

Yes, their quote unquote delaying it.

It doesn't exist, it never existed. Yeah, we'll see Apples washed.

I mean, but that's the thing.

It's the most brand conscious company on the planet. And I wrote like when they did their June revelation of the SERIAI is going to come out, and they said it was going to come out in the fall, and then it was coming out in the spring, and now it's not coming out ever, question Mark. But throughout the whole like two hour presentation, the letters AI were never spoken. Artificial was never spoken. It was Apple Intelligence. We're doing this,

You're saying it looks like shit, you know, And the suggestions are also really bad too. I've had like, over the last few weeks, a few people give me some bad news in their lives and the responses it gives are really funny. Oh no, it'd be like someone telling me something bad happen and it's like oh, or like

It often does not. Why would I think that the same search function in Apple that cannot find a document that I know what the name is, I'm searching for it. Why would I think that that same computer is going to be able to make wise decisions about my life, finance, and personal relationships.

Because that's Apple and this is AI.

Oh that's true. That's that's how I'll show myself out.

I don't know how much is AI versus just like a good translation app, Like I genuinely don't know.

Like well, it's because AI is such a squishy term that you really don't like in some way, Like I guess AI could be expanded to include a lot of modern computing.

Like I can see travel and like emergency situations where you need where like a good AI translator would be like a real life saver, just as the small A side I was just in Mexico and my step kids were using Google Translate and we were like kind of remembering Spanish, and you know, blah blah blah. Go into a coffee shop and I wanted to order a flat white, and so I used Google Translate to say, like, how

Like an Australian coffee.

I learned that very quickly with the help of Reddit, because I went to the barista and ordered a blanco plano and they were like, are you crazy, gringos? Yeah yeah, I mean like it's the functionality is very limited on those things, and it's just like also, it gets back to like, if it's one hundred percent reliable, it's great. If it's ninety eight percent reliable, it sucks.

And just as an aside, did any of you hear about like the latest like quasi fraudulent thing with Jony Ive that's happening. I just saw the herd so Sam Altman and Jony I've founded a hardware startup that has built nothing. There is a thing they claim, there's a phone without a screen. And Open Ai a company run by Sam Moltman, owned forrincipally by Sam Moltman, and Microsoft is going to buy for half a billion dollars this company that has built nothing, co founded by Sam Moltman.

The description so my colleagues, the information reported this journey, I've Sam Moltman news and the description for the device really makes me chuckle. Designs for the AI device are still early and haven't been finalized that people said. Potential designs include a quote unquote phone without a screen and

Like a bad philosophy class where it's like, what is a phone that's not a phone?

Semi for beginners, Jesus fucking Christ, Oh my god, And that's like that's the thing as well. This feels like a thing that tech media needs to be on as well. Just someone needs to say I'll be saying it like this is bordering on fra all, Like it seems like it must be legal, because otherwise there would be some sort of authority, right, you can't do anything illegal without anything happening. Hm hmm. But it's like, this is one of the most egregious fucking things I've ever seen. This

Yeah, a M with AI just declared bankruptcy.

On the blockchain rumored that rumored dead. Why roomber dead?

I think they did. I don't know. Actually, I remember I read the headline they were to be acquired.

By Amazon, but I think the deal fell through under Lena Khan's FDC I had assumed. And then also a one quick note on the Johnny I Have Similton thing. I guess it's notable that oldman has been working closely with the product but is not a co founder, and whether he has an economic stake in the hardware project is unclear.

Yeah, you know, it just seems to be working closely, like he's just hanging out there on taking a salary in an equity position.

I do think it's very interesting all of these different AI device startups that have popped up in the last couple of years, and my question for them is always just like, to what end? People didn't like Amazon Alexa.

And it also lost ton of money.

Yeah, and Amazon's still trying to make it work. Series never been super popular, and I just don't get like one of my co hosts on the podcast We're Gonna Intelligent Machines is obsessed with all these devices just because he's like one of those tech guys.

Leo.

Yes he's and we love to make fun of you, but he his late device is this thing called a Bee.

We just have Victoria's song on Tokio about that records.

It records everything all the time and then makes puts that up in the cloud and then I guess doesn't store the full transcripts, but does store a little AI generated description of everything you did and whoever you talk to that day. And there's no way. I mean, he's in Leo's in California, which is not a one party recording. You gotta get consent from everybody to record, and the

Record write a diary?

There's literally a Black Mirror episode about that.

I believe first episode.

Everyone has like a recording device, and and then it does. When I was when you were talking about this on the show, I was listening thinking, like in this black thing, it reminded me that like when Facebook started having like all your photos collected under your photos, and like how we started reliving so many experiences of time, like you never for all that, and like look at how happy you were like six years ago, you know, like and

I do think it's also just a night Like humans were not built socially to exist in a world where every interaction is recorded and searchable with everyone forever, Like you would never have a single friend. Romantic relationships would dissolve.

The spotless. But even then, like memory is vastly different to the experience of collecting it, like just existing like we all brainslow. I don't know, my brain just goes everywhere. But like compared to memory, which can be all crystalline and wrong, you can just remember something, you can remember a subtle detail wrong, or you can just fill in the gaps. Memory sucks, so.

It doesn't having like a device that constantly records everything a road at the impulse or maybe the drive to be as present, you know, because you're like, well, it's refer to it.

But this has also got huge privacy implications where suddenly the cops could just be like, yeah, we're just gonna take a recording, We're just gonna subpoenat everybody who was in this area's b device and then suddenly get a recording of everyone's days ever that was just happened to be in this place because we think a crime could have happened there.

But I think that there's an overarching thing to everything we're talking about, which is these are products made by people that haven't made anything useful a little while, and everything is being funded based on what used to work. What used to work was you make a thing, people buy it, and then you sell it to someone else to take it public. This only worked until about twenty fifteen. It's not just as a zero in just free era thing.

Ever?

That would be good. Then you could reflect on things. That's what people do, right And I actually genuinely think there is only one question to ask the be founder and that's are you wearing one of these now? And how often do you use this? Because if they use it all the time, I actually kind of respect them. I guarantee they don't. I guarantee they don't, and they'll they'll probably say something around a lot of privileged information

Just as also, who wants a phone without a screen. The screen's the best part. I love the screen.

They don't.

I love to hate the screens.

But they don't talk to anyone. They don't have human experience. They don't have friends that like they have friends who are all like have fifty million dollars in the bank account at any time. They just like exist in a different difficulty level. They're all going at very easy. They don't really have like predators of any kind. They don't really have experiences, so they're what they experience in life is when you have to work out what you enjoy.

Uh?

Auto McDonald's.

Can it do it?

Not? Really? But it also could take a photo of something, it could be pixelated.

That could all kind of order an uber through it.

Maybe what was great was the rabbit launch. The rabbit launch and he tried to order McDonald's live and it just didn't work. It took like five minutes to fail. It was And that's the thing, Like, I feel like when this hype cycle ends, the tech media needs to just be aggressively like, hey, look, fool me, thrice, shame on me. Like maybe maybe next time around, we can't ask the questions I was asking twenty twenty one, where it's like, what does this do? Who is it for?

Hi?

You can find me at CNN dot com, slash Nightcap. I write the CNN Business Nightcap. It's in your inbox four nights a week.

Oh yeah, Ed.

I write a newsletter on Substack, the Tech Bubble. I co host podcast This Machine Kills with Jathan Sadowski, and I'm on Twitter a Big Black Jacobin on Blue Sky too, Yeah, Blue Sky at vwudon Guasso Junior dot com.

You can read my work at the Information I also host a podcast called Intelligent Machines and You can find me on Twitter at Paris Martino or on Blue Sky at Paris dot NYC.

And you can find me at at Edzitron dot com on blue Sky Google who destroyed Google Search? Click the first link. It's me. I destroyed Google Search along with public Ragaban. Fuck you dude. If you want to support this podcast, you should go to the Webbies. I will be putting the link in there. I need your help. I've never won an award in my life. It's the best episode. It's sorry, best business podcast episode. If we

Better Offline is a production of cool Zone Media. For more from cool Zone Media, visit our website cool Zonemedia dot com, or check us out on the iHeartRadio app, Apple Podcasts, or wherever you get your podcasts.

Yes,