🚀 Add to Chrome – It’s Free - YouTube Summarizer
Category: AI Critique
Tags: AIBubbleEconomicsInvestmentTechnology
Entities: AmazonAnthropicEd ZitronEric SchmidtGoogleMicrosoftMITOpenAISam Altman
00:00
When did making plans get this complicated? It's time to streamline with WhatsApp, the secure messaging app that brings the whole group together.
Use polls to settle dinner plans. Send event invites
00:15
and pin messages so no one forgets mom's 60th. And never miss a meme or milestone.
All protected with end toend encryption. It's time for WhatsApp.
Message privately with everyone. Learn more at what'sap.com.
This episode is brought to you by State
00:32
Farm. Checking off the boxes on your to-do list is a great feeling.
And when it comes to checking off coverage, a State Farm agent can help you choose an option that's right for you. Whether you prefer talking in person, on the phone, or using the award-winning app, it's nice knowing you have help finding
00:48
coverage that best fits your needs. Like a good neighbor, State Farm is there.
[Music] In case you were wondering, it is perfectly fine with Ed Zitron if you
01:04
call him an AI hater. I think so.
I think considering I wrote the haters guide and when you type haters guide into Google, it's one of the top two links, the thing I wrote. [Music] Ed writes the newsletter, Where's Your Ed at?
and has a podcast called Better
01:22
Offline. He's a PR guy, but for the last few years has been really making a name for himself as someone who dislikes generative AI, especially large language models.
They are economically not feasible. They require stealing from millions of
01:37
people. They require burning our environment.
They're incredibly expensive and inefficient. And also, they have now become a loadbearing part of our economy without having any means of bailing out or going into the green at all.
It's all very dangerous.
01:54
Yeah, not a fan. Which is exactly why we asked him on the show.
And I want to be clear here, Ed doesn't hate all AI. But large language models like Chachi PT and Claude, he is out there shouting from the rooftops that all that money pumped
02:10
into those companies has created a giant bubble. One he started warning about in 2024.
Indeed, I looked across pretty much every public company talking about AI. No one would talk about the revenue from AI.
They kept talking about how big AI was, how important AI was, how many
02:27
things AI could do, always could do, never what it did that day, but never, hey, we're making this much money. And then I started reading articles about it, and the quotes, and all the quotes were really vague, too.
And it became really obvious that we were doing this
02:43
again, this being another bubble after the cryptocurrency one and the metaverse one, but this one was more egregious and was already at that point running away. What kind of reactions did you get?
I did not get good ones. Pretty much the
02:59
argument I got against me was you are not being willing to accept that the future is here. You are not being open.
you're your mind is not open to the glory of AI and to the obvious future and that it was inevitable that AI was
03:15
coming and people were quite pissy with me. People were genuinely there were people who were saying I was being a a crank that I was just being a contrarian.
I don't have the ability to be anything other than I am. So that really pissed me off.
The reason I wanted to talk to Ed now is
03:31
that other people are also starting to use the B- word. some investors, a smattering of journalists, even and it's debatable how seriously he meant it, Sam Alman.
Let's say that this is a bubble
03:46
that's on the verge of bursting in one way or another. If you're a normal person, not deeply steeped in following every move of every AI company, how should you think about this?
I don't know. That's the sad thing.
I
04:04
tried to warn people a year ago because the thing to do would have been, I don't know, start re-evaluating Nvidia, having a firm market understanding of the company. Stop valuing the Magnificent 7 stocks at such a high growth rate despite the fact that every bit of growth they've had is disconnected from
04:20
AI. The Magnificent 7 is an investor nickname for Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla.
We are at this point guaranteed to suffer. There is no avoiding this.
I'm not a financial analyst. I'm not going
04:35
to tell people what to do with their stocks, but it is inevitable that something happens now. And the longer this goes on, the worse it will be, the more cataclysmic it will be.
I don't know what to tell people. I truly don't.
Other than you need to look at companies
04:51
with fundamentally more logic and reasoning than we currently are, and our markets are disconnected from reality and rationality. Today on the show, Ed's take on where we're at and what happens next.
It's a
05:08
little spicy. I'm Liz Liri and you're listening to What Next TBD, a show about technology, power, and how the future will be determined.
Stick around.
05:24
This episode is brought to you by Sakfth Avenue. Saks Fifth Avenue makes it easy to get creative with your personal style and find the best arrivals for fall.
Whether you're trying out the latest fashion trends or curating a closet that stands the test of time, Sax has you
05:40
covered. Saxs makes shopping feel personal, their trusted instore stylists are there to guide and inspire you, helping you find pieces that fit every aspect of your lifestyle.
Prefer online shopping? Saxs.com offers a curated selection tailored to your taste.
From
05:57
new arrivals by your favorite designers like Prada jackets and Gucci loafers to stylish pieces for your everyday needs. Whether it's date night, a vacation, a special occasion, or just lounging in style, Sax helps you find the perfect look for every occasion.
If you want
06:14
shopping to be fun and easy, then head to Saks Fifth Avenue for inspiring ways to elevate your personal style every day. At Sierra, discover great deals on top brand workout gear like highquality walking shoes, which might lead to another discovery.
06:29
40,000 steps, baby. Who's on top now, Karen?
You've taken the office step challenge a step too far. Don't worry, though.
Sierra also has yoga gear. It might be a good place to find your zen.
Discover top brands at unexpectedly low prices.
06:46
Sierra, let's get moving. I want to start by helping our listeners understand the AI industry kind of as it currently stands.
Um, sure. Let's start with the generative AI companies since that is I think what most people are thinking of when they
07:01
hear AI. Who are the big players we need to pay attention to?
So, there are two, OpenAI and Anthropic. Now, it's much easy to think of these as outgrowths of Microsoft, Amazon, and Google because that's where they get their money.
It's where they get everything. If you
07:17
really look at this closely, all that the generative AI industry is magnificent seven companies like Microsoft, Amazon, and Google building things and sending themselves money. So, OpenAI principally funded by Microsoft.
Microsoft runs all of their infrastructure. All those GPUs that are
07:32
needed to run chat GPT are owned by and run by Microsoft. While OpenAI is building more infrastructure, they're doing so dependent on venture capital.
OpenAI has about 700 million weekly active users, though they failed to define what that means. And I question that number at its source.
They're
07:48
expecting to lose anywhere from 8 to 12 billion this year and will I think they're expected to burn $44 billion between now and 2030. They have no path to profitability.
Anthropic principally funded by Amazon and Google with Amazon running most of their infrastructure. In
08:04
fact, I think between Google and Amazon, they own like 30% of this company, which is wild. And in both of these cases, these companies lose billions of dollars.
Anthropic has leaked that they will lose $3 billion this year. I don't think that's true.
I think it's going to be more like 5 to10 billion.
08:19
You think they're sort of softening things. I am deeply suspicious of the numbers these companies leak.
But nevertheless, the generative AI industry is effectively anthropic and open AI. Neither of them are profitable.
Both of them lose billions of dollars. Neither of them have a path to profitability.
08:35
OpenAI is known for chat GPT. Anthropic is known for Claude and and Claude Code being their coding platform.
Yes. Now, in both of these companies cases, they have a real problem in that their products do not really have firm use
08:50
cases. They don't have the kind of mass market use cases that generate the returns they need.
And that is continuing to be a problem because even in this massive three-year hype cycle, even with all of these headlines and all of this money, most of this stuff comes
09:07
down to generating text, summarizing text, generating images, looking at stuff, and generating what is most likely the thing the user might want. Generating code, which costs an incredibly large amount of money.
And also all of these things problematically
09:22
are rife with hallucinations which are when a model authoritatively states something that isn't true. This becomes a problem infrastructurally because every single time these models do anything it costs quite a lot of money.
It costs so much money and indeed Anthropic has a massive problem on their
09:39
hands where users of their clawed code they have users who even with rate limits are costing them hundreds thousands of dollars a month on a subscription that caps out at $200 a month. H and there is no way to reverse this trend because nothing about generative
09:55
AI maps onto a monthly fee. And that's there are economic problems at the center of this that are absolutely impossible to fix.
So here I think it might be time to talk about this report that recently came out
10:10
of a of MIT that reviewed about 300 different AI initiatives and found that 95% of businesses are getting zero return from generative AI. I found that to be phenomenally striking
10:26
and at the same time many enterprises are rushing hard and spending lots of money to incorporate AI into all of their products. Why?
So that report is actually fascinating because one tons of outlets misreported
10:43
it. They said that the problem with that f that 95% was that there was a learning gap and they said that it was because this learning gap was the user could not understand it.
No, the learning gap was between the models and the task. The models are incapable of learning because they're
10:58
not conscious. And that report had one particularly fascinating little fact which was the people have been saying, "Oh, we're early with the the early days.
The early days, Liz, we got to be patient." No, actually the report said that adoption is high but disruption is
11:14
low. This means everybody is trying it and it isn't bloody working.
The frustration in my voice is that this is what I said a year ago. Everything that they said in here was exactly as I suspected because I actually talk to people that make these things and run these things and
11:30
people that run businesses. And when you talk to them, there are either the fanatics who are just lying.
they just lie because it's within their interest to lie. All the people really building stuff would go, "Yeah, you know, it's kind of useful, but I wouldn't build a whole company on it because I'd just lose a bunch of money."
11:48
I mean, it is interesting when you look into the inards of that report and you look at where generative AI investment went, sending outbound emails, personalized content. It's a lot of what I would think of as um making
12:06
Microsoft Outlook a little snazzier. Yeah, pretty much.
Most of It's funny. You hear all these people claiming that, oh, everyone's using AI and it's changing everything.
Then you actually hear what it does. It's like, oh, right.
Um, yeah, you can summarize an email. You can draft an email.
Don't know about
12:22
you, Liz, I've never had trouble writing or reading an email personally. Never really been an issue.
They're usually short. I guess that there are people who might summarize documents with this, but are we really spending $400 billion in capital expenditures, which by the way is more than
12:39
contributed more to GDP growth in the US in the first half of this year than all consumer spending combined. Are we really doing that to summarize emails or documents?
Are we doing this to generate sexy Garfields? Like what is it that PE and that's the thing when you really look at the things that these things do
12:55
and even on a business level that these things do they're not exciting they're not big the what is this changing what is this disrupting what is the thing that changes and you could bring in code you could say coding I've yet to find an
13:12
actual real person who can show me that they've replaced a coder with this and also we don't hire coders we hired software engineers is software engineers do more than write code, right? I mean, I will say that I have a sibling who is a front and backend developer who uses it to code, but he checks his code,
13:28
right? It's not like you can just leave it to do the work.
The second stop on on our our road to bubble warning after the MIT report is an op-ed that Eric Schmidt co-authored in the New York Times and Eric Schmidt who's been an AI booster. And yet there
13:46
is a line that is very striking where it was written it is uncertain how soon artificial general intelligence can be achieved which to me is a big woe because AGI artificial general intelligence is what the co-founders of
14:04
open AI have always said they were pushing toward. They have told journalists early on that, you know, that was the thing, that was the loadar.
And to me, that is a another flag of,
14:19
hey, maybe this thing isn't doing what everybody said it was. I wonder how you read that piece.
I read it in the same way I've read everything Eric Schmidt has said for the last 10 years, by rolling my eyes. Back in April 2025, Eric Schmidt said that he
14:35
believed in the next year the majority of programmers will be replaced by AI. I know.
So what? Why the shift?
Lying that like it really is that simple. They never believed it was possible in the beginning and now it's become really obvious they are trying to
14:50
backpedal. Sam Alman never thought AGI was possible in my opinion.
Can't prove it but I don't know. He Alman referred Clammy Sam Alman, Clamuel Alman a couple weeks ago said that AGI was no longer a useful term after saying a month before
15:07
that I believe in one of his annoying poorly written blogs that super intelligence was in reach. Which is it Sammy?
And the answer is it isn't. It isn't.
It's all it's a fugazi. It's all just complete theater and the media has
15:23
fallen for it. hook, line, and sinker again and again and again.
That really brings us to a fundamental economic question. Mhm.
You're essentially saying all of these guys are talking their books, a Wall Street term for, you know, pumping up what you're selling. Let's talk about
15:39
how these companies make money or are supposed to make money because you very clearly say not only are they not making money, they're losing money. Um yes, if we take OpenAI, the bulk of their supposed revenue, they say they will do
15:54
about 12 billion in business. Uh almost half of that comes from subscriptions, but you have a very interesting take on how they book that revenue.
And I wonder if you could explain that. So, it's funny.
I actually think it might be more than half of their revenue comes from subscriptions. And actually,
16:11
I'm not really sure where all their revenue comes from. Everyone's been reporting and there's this term annualized revenue that a lot of these companies explain that please happily.
So what annualized revenue is meant to mean is month times 12. It is generally meant to refer to a calendar
16:27
month. You generally have to report this like bigger the company the longer it takes like a month after it because you need to let the book settle.
This is standard. The reason they use this term annualized is because when you say the actual amounts, I mean, I don't know, $13 billion in annualized revenue means
16:44
$1 billion of revenue in a month. That's pretty good, right?
Yeah. Here's the thing.
How the hell are they at a billion dollars of revenue a month? Cuz when you really break down the numbers, when you really take a look at the bits and pieces, things start getting a little weirder because they
17:00
report they have 20 million paying subscribers, 5 million paying business subscribers. 500,000 of those business subscribers are Cal State University, a $15 million one-year contract that works out about $2.50 per user per month on a
17:16
contract that is meant to be 35 per user. So they're just burning money there across the board.
It isn't clear where where OpenAI is making this money. I don't even think they're cooking the books.
I really don't. I don't think that they're doing anything illegal per
17:31
se because these are not this is not a public company and thus they are not bound by account. I think that's important for listeners to understand that if this were a public company, we could go to the SEC, we could read their filings and we would have a much clearer understanding of the
17:47
company's finances. Yes.
And that's the thing. They don't want that.
The plan the plan is somehow for this company to go public, which requires it to convert from a nonprofit to a for-profit, which is a whole other thing. But if I don't believe their accounting is good and is going to exist
18:04
after the scrutiny of a real accountant for accountancy firm, a PWC would eat them for dinner. And this is the thing.
This company has built this myth, this legend that they're going to hit $ 122.7 billion of revenue this year. How much
18:20
are they going to lose? Well, we don't really know.
They've leaked that they will lose 8 billion, but they lost 5 billion last year. So, I think it'll be more like 12 to 15 billion they lose this year.
I'm throwing a lot of numbers at you, but it comes down to something quite simple, which is they do not appear to have the means to charge enough money to
18:37
actually go into the green. And I think it's because they would need to charge at minimum 50 bucks per user per month without fail.
And I think right now they're doing the literal opposite. I have evidence that they are well I mean evidence being my own experience and tons of Redditors that if you click the
18:55
right buttons, you can make OpenAI give you 3 months of half off the 20 bucks a month rate. They are also to some subscribers and I'm seeing this in a lot of places offering one month of their team subscription for a dollar.
How many of those paying subscribers are
19:13
these? How many of those paying subscribers, those 5 million business subscribers have paid OpenAI $1 once?
We do not know and the media does not ask these questions and we actually can't find out because they're not public. What I fear is that things are much
19:29
worse at OpenAI and Anthropic than we currently know. If I'm wrong and they're exactly what they're saying, they're still in trouble because they're still losing billions of dollars.
It's bad either way when we come back. So, okay, bubbles
19:47
pop. What would that look like in this case?
Running a business comes with a lot of whatifs, but luckily there's a simple answer to them. Shopify.
It's the
20:03
commerce platform behind millions of businesses, including Thrive Cosmetics and Momaf Fuku. And it'll help you with everything you need from website design and marketing to boosting sales and expanding operations.
Shopify can get the job done and make your dream a reality. Turn those whatifs into.
20:21
Sign up for your $1 per month trial at shopify.com/special offer. This episode is brought to you by Indeed.
When your computer breaks, you don't wait for it to magically start working again. You fix the problem.
So, why wait to hire the people your company
20:37
desperately needs? Use Indeed sponsored jobs to hire top talent fast.
And even better, you only pay for results. There's no need to wait.
Speed up your hiring with a $75 sponsored job credit at inddeed.com/mpodcast.
20:53
Terms and conditions apply. Lowe's knows you've got a job to do and we help get it done.
With the Milo's Pro Rewards program, eligible members save more with volume discounts on qualifying orders through a quote of $2,000 or
21:10
more. Join for free today.
Lowe's. We help you save.
Offer can't be combined with any other discount contract and/or special pricing. Exclusions, more terms, and restrictions apply.
Details at lowe's.com/terms. Subject to change.
If you were my age, almost 50, you
21:28
remember when the dot bubble burst. Price to earnings ratios on tech stocks had soared to crazy levels.
Pets.com was a household name. Companies were spending gobs of money.
But then, crucially, they started running out of
21:43
it. The government then won its case against Microsoft, and the Fed began raising interest rates.
All of a sudden, tech didn't look so hot anymore. Ed argues that something similar is happening now and he says Open AI is the
21:59
canary in the bubble coal mine. Open AAI raised oh god they raised 8 point they've raised $18.3 billion this year.
They're meant to receive another $20 billion by the end of the year. But
22:14
to do that they need to convert their and this is a very simplified version. They need to convert from a nonprofit to a for-profit entity.
Long story short, OpenAI began as a nonprofit. Then it started a for-profit arm.
In May, the company transitioned to
22:32
a public benefit corporate structure. Now, they need to convert by the end of the year.
Otherwise, Soft Bank, who is who is the one handling their $40 billion round, which is not fully raised, just to be clear, cuts their round in half if they don't convert by
22:48
the end of December 31st, 2025. Standing in the way of that is Microsoft.
Microsoft in their investment has access to all of OpenAI's IP, all of their research, and has exclusive rights to sell their models and gets a 20% uh revenue cut from everything OpenAI does.
23:05
If OpenAI cannot convert by the end of the year, they don't get a bunch of that money. Convert to a for-profit indeed.
And if they don't by the end of the year, Soft Bank likely cuts the round in half. Even if they do get that money somehow by raising from other venture capitalists, if they cannot
23:20
convert to a for-profit, which may not be possible, this is very important, they will die. They cannot survive as a nonprofit because effectively they would become a Ponzi scheme.
They would be a scheme where they burn money, people
23:36
invest into them and the only way to to get liquidity, which is the term for when you sell your stock in a company, would be to sell it to someone else. And quite frankly, why would an investor invest in a company which really has no liquidity event in front of it?
Open AAI
23:53
if they cannot convert will die. And open AI is effectively the AI industry.
So that brings me I guess to sort of three points that I want to put to you in in in varying orders. So the first
24:09
one is what I would call the Uber hypothesis which is okay so this thing has been subsidized by investors. It's not quite a onetoone because Uber was also subsidized by a very very um free interest rate environment, right?
But
24:26
the idea that we need a little time, we'll work the kinks out and we'll come up with a profitable business. Mhm.
I know you think that's bull. Tell me why.
Well, very important place to start. Uber in its entire lifetime, so over 10
24:43
years, over a decade, burned about 2530 billion total. That it and a lot of that was in marketing.
You remember Groupon? Groupon did the same thing.
They burned all of this money to market their service and eventually had no way of turning things profitable. Uber, same
24:59
deal. Uber, however, had a business that made sense.
There was a way to make this profitable. It still burns a bunch of cash because the cost of customer acquisition is incredibly high.
The cost of running the service was not. The actual internal costs of running Uber
25:14
are expensive because it's an atscale company, but the infrastructure costs were not. The big mistake that people make with the Uber comparison is just saying Uber burned 25 to 30 billion so it can't be that bad for these companies.
No, no, no. It's actually so much worse.
People are missing
25:30
something. And that's how because the compute cost so much because of the infrastructure to provide the compute.
No one sees and no one wants to admit that Microsoft and Amazon have pro and Google as well between them probably put $250 billion just into the
25:45
infrastructure to run OpenAI and Anthropic. And those numbers are never included in the evaluation of these companies.
On top of that, OpenAI when they closed this round has raised like 60 70 billion. And on top of that, Anthropics what 3040
26:02
billion at this point. So instead of they've burned like 10 times what Uber burned in 10 years in three and I actually think my capex estimate is somewhat low.
On top of that, the internal fuel of AI running a GPU on an
26:20
hourly basis may not be profitable. Semi analysis, which is a big expensive $500 uh newsletter, you you look at them and they say, "Yeah, for hyperscalers it's about a buck 776 an hour to run the average GPU." To turn the green on that,
26:36
you need to be running them at 80% utilization for 3 years. And that's to just break even.
Here's the crazy thing. This would be if like Uber cars all took $50,000 of fuel a month.
It is that
26:52
ridiculous. It is that different.
And there is no knob to turn to turn these profitable. I want to talk about the hey but AI works in some areas in medicine
27:07
in research in the fact that you know my oncologist could make this incredible model of my tumor before they cut it out of me like that stuff is amazing and and so that's where I wonder that's where I have some pause
27:22
sure but that that thing you just described doesn't exist there are machine learning learning models that are doing incredible things in the sciences. Those are not large language bottles.
And people like Sam Orman love to conflate them. And so you get these situations where and I'm I'm not
27:38
critiquing you. I'm just saying, yeah, if they could do that, that would be great.
That isn't what they're doing. The way that they're being you way that AI is being used within healthcare, for example, is as a scribe thing, like a bridge, very big company for taking notes from doctors.
And yes, the
27:54
administrative burden is a huge thing and it's good to reduce that. I'm not getting that wrong.
But again, we're talking about something that is going that hyperscalers are going to spend $400 billion dollars this year building out. And the actual use cases are nowhere near that kind of thing.
The
28:11
actual use cases are nowhere near the myth, but the myth is what's used to sell it. And so people will say, well, they're getting exponentially better.
And the reason they say that is because these things are improving on benchmarks that are specifically built for large
28:27
language models. These things are not getting better actual use cases that would change the world.
The actual use cases of generative AI are fairly similar to where they were 6 to 12 months ago. In fact, the only thing that's really changed is the launch of reasoning models at the end of last
28:43
year. and those have increased the prices of the costs of LLM significantly while not really creating any new use cases.
Now let's talk about where politics comes into the picture, right? And not just politics, but as we
29:01
have seen in the US government starting to take cuts of, uh, tech company deals, talking about Nvidia, sales to China, um, and whatever
29:16
the heck it's doing at Intel. There is a question to me as to whether the Trump administration in particular is going to attempt to think of these
29:32
companies as too big to fail. If OpenAI disappeared tomorrow, nothing would change.
Legitimately, I don't think anything would change. Like, yeah, there would be news outlets would have to find something else to lie about, but I mean, I'm sure they could find something.
These companies are not essential. Their
29:49
outcomes are not essential. And indeed, their involvement in government doesn't make them essential at all.
If these things fail, there isn't really an obvious thing that changes. It isn't something that would affect the market other than the fact the myth will die.
30:05
And even that's going to happen either way or I guess people won't want to buy as many GPUs. To be clear, the only possible too big to failish thing they could do would be to permanently subsidize every AI company and every single purchaser of GPUs and effectively
30:22
nationalize Nvidia. This isn't going to happen.
You can't bail this out because in the too big to fail example of the past, something would fail and you could plug that with money top. So you'd have a fund that would fund that and then there would be an end point to that funding.
There isn't such a thing for
30:39
this. There isn't anything to bail out.
Microsoft, Google, Amazon, Meta, they're going to be fine. Their data centers could all melt tomorrow.
Not all of them, but the GPU ones. Eh, they'll still make a bunch of money from Microsoft 365 or advertisements on Facebook or what have you.
30:55
There is no way you can plug the gap. If OpenAI fails, yeah, there'll be people who lose money, but that's not something where you can just give a bunch of money.
These services are not essential and they're not they're these businesses are not large parts of our economy. From
31:10
what I can tell, the combined revenue of every single AI company, including the revenues from companies like Microsoft and Amazon and Google, is like 35 to40 billion, which is about how much uh smartwatches made last year. It's not It's very loud.
It's kind of like the
31:26
MIT study. Lots of adoption, not much disruption.
So, why are people listening now? I think because the time has run out, nothing has changed, nothing has happened, and the costs have gone up and we now have, thanks to people like Paul
31:42
Kadroski and Christopher Mims at the Wall Street Journal, real insight into the rotten economics beneath. And on top of that, I think people have just been using these things for long enough to go, all right, when's the thing I was meant to be exciting by happening?
So
31:58
let's say this bubble does burst. What does that mean for a casual user of these products or even a nonuser of the products?
Because generally companies don't
32:14
explode without some kind of follow-on consequence. So pretty much I believe that chatgptt.com will forward to copilot.
I think Microsoft will just absorb open AAI into its body like just in it goes and I think that
32:33
the version of chat GPT you see today will be dramatically different in the sense that it will be so heavily rate limited that you'll be able to use it a couple times a day. I think subscriptions to these products will shoot up to 50, 100, 250 a month just to
32:49
start if not more because the economics underlying them are so bad. So so so very bad.
I cannot I don't think people truly know yet because we don't have the underlying accounting how bad this will be. But everything I'm seeing suggests
33:05
that this is the single most aggressive burn rate on software ever made. I think a lot of these products will simply cease to be.
Will you take a victory lap? I don't know.
I think when Open AI dies, I might pop some champagne, smoke a big cigar, but I don't think it's going to
33:22
be that clean. I think there could be a chance that whatever my editor, Matt Hughes, uh, he had the idea that OpenAI will exist as a PO box that sues people.
It's like a patent troll. But I think that the collapse of one of those companies will be a moment I don't know
33:38
to celebrate because real people are going to lose money. There are going to be actual economic consequences here.
Nevertheless, the celebration I have will be to say to people, you should have listened earlier. The harm could have been reduced and perhaps there are people that I will name that were principally
33:54
responsible for spreading the myth. And those people must be held accountable.
I think there are members of the media who will be held accountable. But the big thing I will want to say is that the amount of the many many many people I hear from that say I feel like I'm being gaslit.
I think those people will feel vindicated for believing their eyes for
34:12
believing what they felt when they used these products was real versus everyone telling them otherwise. [Music] Ed Zitron, thank you so much for coming on.
34:27
Thank you for having me. It's been wonderful.
Ed Zitron writes the newsletter, Where's Your Ed at? and hosts the podcast Better Offline.
All right, that is it for our show today. What Next TV is produced by Evan Campbell and Patrick Fort.
Our show
34:43
is edited by Paige Osborne. Mia Loel is the executive producer here at Slate, and TBD is part of the larger What Next family.
And if you're looking for even more Slate podcasts to listen to, head on over to our Plus feed, which is exclusively for Slate Plus subscribers.
34:58
We have a new episode of The Discourse out today where Slate writer Natish Pawa and I discuss the outcome of the landmark Google Antirust case, The Search One, and the government's remedy. All right, we'll be back on Sunday with another episode.
I'm Lizio Liry. Thanks
35:14
for listening. [Music] This episode is brought to you by Sakf Avenue.
Saks Fth Avenue makes it easy to get creative with your personal style
35:31
and find the best arrivals for fall. Whether you're trying out the latest fashion trends or curating a closet that stands the test of time, Sax has you covered.
Saxs makes shopping feel personal, their trusted instore stylists are there to guide and inspire you,
35:47
helping you find pieces that fit every aspect of your lifestyle. Prefer online shopping?
Saxs.com offers a curated selection tailored to your taste. From new arrivals by your favorite designers like Prada jackets and Gucci loafers to stylish pieces for your everyday needs.
36:05
Whether it's date night, a vacation, a special occasion, or just lounging in style, Sax helps you find the perfect look for every occasion. If you want shopping to be fun and easy, then head to Saks Fifth Avenue for inspiring ways to elevate your personal style everyday.
36:23
Listen. That's the sound of the fully electric Audi Q6 Ron.
The sound of captivating electric performance, dynamic drive, and the quiet confidence of ultra smooth handling.
36:39
The elevated interior reminds you this is more than an EV. This is electric performance redefined.
The fully electric Audi Q6 Ron.