Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How the AI Bubble Will Pop

derekthompson.org

Post reply on HN
Print view View original

311–320 of 531 posts

Re: How the AI Bubble Will Pop

#311

Post by fragmede »

camillomiller wrote: post #178
donatj wrote: post #123

I feel kind of like a Luddite sometimes but I don't understand why EVERYONE is rushing to use AI? I use a couple different agents to help me code, and ChatGPT has largely replaced Google in my everyday use, but I genuinely don't understand the value proposition of every other companies offerings. I really feel like we're in the same "Get it out first, figure out what it is good for later" bubble we had like 7 years a…

There is none, zero value. What is the value of Sora 2, if even its creators feel like they have to pack it into a social media app with AI-slop reels? How is that not a testament to how suprisingly andvanced and useless at the same time the technology is?

It's in an app made by its creator so they can get juicy user data. If it was just export to TikTok, OpenAI wouldn't know what's popular, just what people have made.

Re: How the AI Bubble Will Pop

#312

Post by fragmede »

shakadak wrote: post #304
tbrownaw wrote: post #185

Earlier quoted context omitted.

Well there was that example a while back of some store's product recommendation algo inferring that someone was pregnant before any of the involved humans knew.

The ~woman~ teenager knew she was pregnant, Target's algorithm noticed her change in behavior and spilled the beans to her father.

Back in 2012, mind you.

Re: How the AI Bubble Will Pop

#313

Post by ajross »

incomingpain wrote: post #96

>Others insist that it is an obvious economic bubble. The definition is that the assets are valuated above an intrinsic value. The first graph is Amazon, Meta, Google, Microsoft, and Oracle. Lets check their PE ratio. Amazon (AMZN) ~ 33.6 Meta (META) ~ 27.5 Google (GOOGL) ~ 25.7 Microsoft (MSFT) ~ 37.9 Oracle (ORCL) ~ 65 These are highish pe ratios, but certainly very far from bubble numbers. OpenAI and others are al…

> Objectively there is no bubble. Economic bubble territory is 100-200+ PE ratios.

Not sure I buy that analysis. That was certainly true in 2001. The dot com boom produced huge valuations in brand new companies (like the first three ones in your list!) that were still finding their revenue models. They really weren't making much money yet, but the market expected them to. And... the market was actually correct, for the most part. Those three companies made it big, indeed.

The analysis was not true in 2008, where the bubble was held in real estate and not corporate stock. The companies holding the bag were established banks, presumptively regulated (in practice not, obviously) with P/E numbers in very conventional ranges. And they imploded anyway.

Now seems sort of in the middle. The nature of AI CapEx is that you just can't do it if you aren't already huge. The bubble is concentrated in this handful of existing giants, who can dilute the price effect via their already extremely large and diversified revenue sources.

But a $4T bubble (or whatever) is still a huge, economy-breaking bubble even if you spread it around $12T of market cap.

Re: How the AI Bubble Will Pop

#314

Post by walleeee »

uladzislau wrote: post #145

Calling this an “AI bubble” reads like pure sour grapes from folks who missed the adoption curve. Real teams are already banking gains - code velocity up, ticket resolution times down, and marketing lift from AI-assisted creative while capex always precedes revenue in platform shifts (see cloud 2010, smartphones 2007). The “costs don’t match cash flow” trope ignores lagging enterprise procurement cycles and the rapid…

> Habit formation is the moat

Well, at least you're honest about it.

Re: How the AI Bubble Will Pop

#315

Post by _heimdall »

throw0101d wrote: post #249
_heimdall wrote: post #38

Earlier quoted context omitted.

My understanding is that the 40 hour work week (and similar) was talked about for centuries by workers groups but only became a thing once governments during WWI found that longer days didn't necessarily increase output proportionally. For a 4 day week to really happen st scale, I'd expect we similarly need the government to decide to roll it out rather than workers groups pushing it from the bottom up.

> My understanding is that the 40 hour work week (and similar) was talked about for […] See perhaps: * https://en.wikipedia.org/wiki/Eight-hour_day_movement Generally it only really started being talked about when "workers" became a thing, specifically with the Industrial Revolution. Before that a good portion of work was either agricultural or domestic, so talk of 'shifts' didn't really make much sense.

Oh sure, a standard shift doesn't make much sense unless you're an employee. My point was specifically about the 40 hour standard we use now though. We didn't get a 40-hour week because workers demanded it, we got it because wartime governments decided that was the "right" balance of labor and output.

Re: How the AI Bubble Will Pop

#316

Post by blonder »

boxed wrote: post #207
skybrian wrote: post #194

Earlier quoted context omitted.

How do we know this?

Statistically this is obvious. Most people use the free tier. Their total losses are enormous and their revenue is not great.

Sam has claimed that they are profitable on inference. Maybe he is lying but I don't think speaking so absolutely about them losing money on that is something you can throw around so matter of fact. They lose money because they dump an enormous amount of money on R&D.

Re: How the AI Bubble Will Pop

#317

Post by rimunroe »

danaris wrote: post #303
1123581321 wrote: post #135

Earlier quoted context omitted.

It isn’t a question asked by people generally suspicious of productivity claims. It’s only asked by LLM skeptics, about LLMs.

As a counterexample to your assertion, I've seen it a lot on both sides of the RTO discourse.

This is another example of the phenomenon they’re describing, not a counterexample.

Re: How the AI Bubble Will Pop

#318

Post by cyberpunk »

sseagull wrote: post #299
delichon wrote: post #70

Earlier quoted context omitted.

Someone is paying. OpenAI revenue was $4.3 billion in the first half of this year.

OpenAI "spent" more on sales/marketing and equity compensation than that: "Other significant costs included $2 billion spent on sales and marketing, nearly doubling what OpenAI spent on sales and marketing in all of 2024. Though not a cash expense, OpenAI also spent nearly $2.5 billion on stock-based equity compensation in the first six months of 2025" ("spent" because the equity is not cash-based) From https://archi…

How the fuck does anyone spend 2 billion dollars on sales and marketing. I’ve seen the odd ad for openai but thag number seems completely bananas.

Re: How the AI Bubble Will Pop

#319

Post by dns_snek »

maxglute wrote: post #274
donatj wrote: post #123

I feel kind of like a Luddite sometimes but I don't understand why EVERYONE is rushing to use AI? I use a couple different agents to help me code, and ChatGPT has largely replaced Google in my everyday use, but I genuinely don't understand the value proposition of every other companies offerings. I really feel like we're in the same "Get it out first, figure out what it is good for later" bubble we had like 7 years a…

See kids hooked on LLMs. I think most of them will grow up paying for a sub. Like not $15/m streaming sub, $50-100/m cellphone tier sub. Well until local kills that business model.

Local models won't kill anything because they'll be obsolete as soon as these companies stop releasing them. They'll be forgotten within 6-12 months.

Re: How the AI Bubble Will Pop

#320

Post by thisisit »

xiphias2 wrote: post #147
donatj wrote: post #123

I feel kind of like a Luddite sometimes but I don't understand why EVERYONE is rushing to use AI? I use a couple different agents to help me code, and ChatGPT has largely replaced Google in my everyday use, but I genuinely don't understand the value proposition of every other companies offerings. I really feel like we're in the same "Get it out first, figure out what it is good for later" bubble we had like 7 years a…

Just you switching away from Google is already justifying 1T infrastructure spend. Just think about how much more effective advertisements are going to be when LLMs start to tell you what to buy based on which advertiser gave the most money to the company.

> Just think about how much more effective advertisements are going to be when LLMs start to tell you what to buy based on which advertiser gave the most money to the company.

This has been the selling point of ML based recommendation systems as well. This story from 2012: https://www.forbes.com/sites/kashmirhill/2012/02/16/how-targ...

But can we really say that advertisements are more effective today?

From what little I know about SEO it seems nowadays high intent keywords are more important than ever. LLMs might not do any better than Google because without the intent to purchase pushing ads are just going to rack up impression costs.

Post reply on HN