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

AI's $344B 'language model' bet looks fragile

bloomberg.com

Post reply on HN
Print view View original

41–50 of 121 posts

Re: AI's $344B 'language model' bet looks fragile

#41

Post by numbers_guy »

redwood wrote: post #4

This technology demos incredibly well and you can just see how everyone gets giddy with excitement around using it. I watch my colleagues and executives proud to show what they could do or make endless jokes about it. It reminds me of when people first got their phones and couldn't stop showing everyone how cool they were. This leads to an over rotation in the perceived value.. the value is significant just as the mo…

> This technology demos incredibly well

You just summed up machine learning, not just AI/LLMs. My domain is very far from LLMs, but even in my domain, you can build a really cool demo, that is entirely misleading.

Re: AI's $344B 'language model' bet looks fragile

#42

Post by hobs »

geon wrote: post #28
wffurr wrote: post #7

Earlier quoted context omitted.

>> in anonymous forums there's a lot more people pointing out that they think this is hype whereas when we wear our professional hats many of us join in When your boss is hyping it up and demanding all hands on deck full steam ahead on the Good Ship AI, lots of people join in out of fear, particularly in the currently awful job market which is partly being ruined by AI itself. Some of us just stay quiet, keep our hea…

No developer job has yet been lost to AI. We are in a huge recession, and companies like to have an excuse to fire people that can be spun as a positive.

At my company at least four developer jobs were lost because of AI tooling, but keep spouting off about the entire economy at once.

Re: AI's $344B 'language model' bet looks fragile

#43

Post by CyanLite2 »

Mizza wrote: post #12

What's the path to recouping that money? Even if every major company in the US spends $100,000 a year on subscriptions and every household spends $20/month, it still doesn't seem like enough return on investment when you factor in inference costs and all the other overhead. New medical discoveries, maybe? I saw OpenAI's announcement about gpt-bio and iPSCs which was pretty amazing, but there's a very long gap between…

Major companies will spend 10-100x that if it resulted in real tangible productivity gains for their businesses.

Re: AI's $344B 'language model' bet looks fragile

#44

Post by andsoitis »

jakeinspace wrote: post #5
redwood wrote: post #2

Oracle's mega moves in the market (cpl hundred $B in market cap),due to their claim that OpenAI was doing a multi-year commit to move much of their workloads to their cloud ... likely a heavy revenue play rather than profit... with an aspiration to push as much CapEx deprecation to out years as possible (btw Oracle where is all the capex?) AKA Financial aengineering... just shows how overly leveraged this bubble has…

This Oracle surge and revenue predictions really feels like jumping the shark. I mean, it's Oracle.... I've never felt confident enough to bet against a company, but a short position on Oracle may well be too tempting for me.

You probably saw in the news yesterday that Oracle stock shot up because they are scoring large AI deals: https://www.wsj.com/business/earnings/oracle-stock-orcl-ai-d...

Re: AI's $344B 'language model' bet looks fragile

#45

Post by zippyman55 »

brookst wrote: post #17
zippyman55 wrote: post #14

Earlier quoted context omitted.

There are no error bars, no confidence intervals. Just a one trick pony that pastes tokens together to give you something that may look good to many people. Sure there are many good use cases, but there are still enough non-patchable and indeterminable in size holes in the watering pail to limit its effectiveness.

Are you saying the Internet is a fad?

Conventional AI and statistics has many opportunities for bounding answers. Then, a less subject knowable person can do something with the result. But if a LLM returns: THE INTERNET IS A FAD, someone may run off and post the result on HN, which would be embarrrasing.

Re: AI's $344B 'language model' bet looks fragile

#46

Post by pjc50 »

geoduck14 wrote: post #11
redwood wrote: post #4

This technology demos incredibly well and you can just see how everyone gets giddy with excitement around using it. I watch my colleagues and executives proud to show what they could do or make endless jokes about it. It reminds me of when people first got their phones and couldn't stop showing everyone how cool they were. This leads to an over rotation in the perceived value.. the value is significant just as the mo…

>It reminds me of when people first got their phones and couldn't stop showing everyone how cool they were. Your comparison to smart phones is interesting. Smart phones are definitely transformative. There was a lot of hype, but still transformative. Do you believe that LLMs and AI is also going to be transformative?

Transformative, but not necessarily in a good way: likely to lead to the end of the open internet, along with all sorts of weird social effects from lowering the cost of convincing fakes.

Re: AI's $344B 'language model' bet looks fragile

#47

Post by freejazz »

brookst wrote: post #17
zippyman55 wrote: post #14

Earlier quoted context omitted.

There are no error bars, no confidence intervals. Just a one trick pony that pastes tokens together to give you something that may look good to many people. Sure there are many good use cases, but there are still enough non-patchable and indeterminable in size holes in the watering pail to limit its effectiveness.

Are you saying the Internet is a fad?

What even is the internet anymore? Do you mean the thing people use to access Facebook, Instagram, TikTok and "X"?

Re: AI's $344B 'language model' bet looks fragile

#49

Post by Fricken »

redwood wrote: post #4

This technology demos incredibly well and you can just see how everyone gets giddy with excitement around using it. I watch my colleagues and executives proud to show what they could do or make endless jokes about it. It reminds me of when people first got their phones and couldn't stop showing everyone how cool they were. This leads to an over rotation in the perceived value.. the value is significant just as the mo…

I've been following progress on self-driving cars since the 2004 and 2005 Darpa Challenges. Things were very exciting in the 2011-2015 timeframe. Demos were mindblowing. Engineers were giddy. The auto industry went from laughing to shaking in their boots, and then scrambling to make it appear as though they had a horse in the self-driving car race. Hundreds of billions in investment were flying around. Now here we are in 2025, things aren't dead at all, but also nobody with skin in the game is under any delusions about how hard it is to build and scale a viable robotaxi operation.

So far the Self driving car hype cycle has served as a useful reference for understanding the hype with LLMs. The main difference with LLMs is that there is 10 times as much money flying around.

Re: AI's $344B 'language model' bet looks fragile

#50

Post by Mizza »

brookst wrote: post #25
Mizza wrote: post #12

What's the path to recouping that money? Even if every major company in the US spends $100,000 a year on subscriptions and every household spends $20/month, it still doesn't seem like enough return on investment when you factor in inference costs and all the other overhead. New medical discoveries, maybe? I saw OpenAI's announcement about gpt-bio and iPSCs which was pretty amazing, but there's a very long gap between…

$100k/year is literally nothing. Think of it as maybe $10k/employee, figuring a conservative 10% boost in productivity against a lowball $100k/year fully burdened salary+benefits. For a company with 10,000 employees that’s $100m/year.

Even at $10k/yr/employee, you'd need 30 million people on the 10k/yr plan to hit 300B ARR. I think that's a hell of a big swing. 3 million, recoup over ten years? Maybe, but I still don't think so. And then competition between 4 or 5 vendors, larger customers figuring out it's cheaper to train their own models for one thing that gives them 90% of the productivity gains, etc.

But rather than speculating, I'm generally curious what the companies are saying to their investors about the matter.

Post reply on HN