Hacker News Reader

Comment by therein | Hacker News Reader

Is it just me or is anyone else worried what will happen when the industry realizes LLMs are just not going to be the path to AGI? We are building nuclear power plants, massive datacenters that make our old datacenters look like toys.

We are investing literal hundreds of billions into something that is looking more and more likely to flop than succeed.

What scares me the most is we are being steered into a sunk cost fallacy. Industry will continue to claim it is just around the corner, more and more infrastructure will be built, even underground water is being rationed in certain places because AI datacenters apparently deserve priority.

Are we being forced into a situation where we are invested too much in this to come face to face with it doesn't work and it makes everything worse?

What is this capacity being built for? It no longer makes any sense.

View on HN

Even if AI progress all stops and we are stuck with these models, it would at least take a few years to make use them to the full potential. What gets lost in this AGI debate is how impressive are these models without even reaching it. And the places where it has been applied are just tip of the iceberg.

It's just like dot-com bubble, everyone was pumped that Internet is going to take over the world. And even though, the bubble popped, the Internet did eventually take over the world.

LLMs are not going to take over the world.

LLMs are not remotely comparable to the Internet in terms of their effect on ordinary people (unless you want to talk about their negative effects...).

People keep trotting out this comparison, and as time goes on it makes even less sense, as we come to see even more clearly just how false the promises of the AI bros are.

If progress stops with these models, we will be left with some interesting curiosities that can help with certain tasks, but are in no way worth the amount of resources it costs to run them en masse.

I'm more worried what will happen when people in denial will be faced with AI replacing their job. I used to hire outsource devs of _relatively_ poor quality (think just below avg. Wipro), but who completed descent amount of projects, for frontend. I will never do that again, unless I have to work with audio or video which modern LLMs can't test on their own.

Probably valid sunk cost fallacy, but it makes me wonder what will happen to the applications and systems being built on top of LLMs? If we face limitations or setbacks, will these innovations survive, or could we see a backlash against all thinking machines, reminiscent of Isaac Asimov's cautionary tales?

so i don't see the path to AGI either, certainly not with LLMs. but there is some very useful stuff to be done with deep learning, and LLMs are a pretty amazin advance for search, translation, communication and user interfaces. but it's not the next industrial revolution.

i don't see the sunk-cost fallacy angle, just the sunk costs. the capital allocators will absolutely shut off the spigot when they see that it isn't going to yield. yeah, there could be some dark data centers. not the end of the world, just another ai winter at worst - maybe a dot-com crash... whatever.

the world is way bigger than the techo chamber

Well, we can for sure put the nuke plants to good use. Maybe for desalination to replace the water supplies.

> We are building nuclear power plants

If only

We are likely in for the biggest market crash in history. Even if LLMs are super useful and will be with us going forward, that doesn't mean they will be profitable enough to justify the spending.

Open source and last years models will be good enough for 99% of people. Very few are going to pay billions to use the absolute best model there is.

> We are likely in for the biggest market crash in history

I don't think there's the kind of systemic risk that you had in a say 2008 is there?, but I do think there is likely to be a "correction" to put it lightly.

Datacenter build-outs are being financed by mega caps with fortress balance sheets and free cash flows that are coming in well above expectations. They are still able to do things like route hundreds of billions in buybacks.

And regardless of being great investments or not, all of those companies have a burning desire for accelerated depreciation to lower their effective tax rate, which data center spend offers.

The more bubbly names will likely come down to earth eventually, but the growth stock sell-off we say in '22 due to the termination of the zero interest rate environment will probably dwarf it in scale. That was a true DotCom 2.0 bubble, with vaporware EV companies with nothing more than a render worth 10 billion, fake meat worth 10 billion, web chat worth 100 billion, 10 billion dollar treadmill companies... Space tourism, LIDAR... So many of those names have literally gone down 90 to 99%. It's odd to me that we don't look at that as what it was - a true dotcom bubble 2.0. The AI related stuff looks tame in comparison to what we saw just a few years ago.

nah, you're being hyperbolic. run the actual numbers. just the US GDP is 28T - AI investment in 2024 was like 250B. it will barely leave a dent

> We are likely in for the biggest market crash in history.

This seems unlikely, because if and when the bottom falls out, it seems implausible that it will be the sort of systemic shock that the financial crisis was, much less the Great Depression. Lots of people would lose lots of money, but you wouldn't expect the same degree of contagion.

I honestly wonder how much they even believe their own hype. Altman is a world class circus barker for AI. I seriously doubt his sincerity about anything. Obviously Zuck is putting his money where his mouth is but idk how a data center the size of Manhattan is a means to any useful end.

I'm just imagining in 2030 when there is an absolutely massive secondary market for used GPUs.

I suspect (but obviously can't confirm, although this is a thought I've had before) that they are now much more of a "normal" company than they were before. I doubt the workforce feels like a god-tier supergroup either - after all there are many places you can work as an AI researcher now, and no doubt many of them left to make a startup or take that hundred million dollar zuck paycheck.

Which leaves Altman frankly looking increasingly awkward with his revolutionary proclamations. His company just isn't as cool as it used to be anymore. Jobs era Apple was arguably a cooler company.

> Is it just me or is anyone else worried what will happen when the industry realizes LLMs are just not going to be the path to AGI?

Yes, some people are concerned, see for example a recent Hank Greene YouTube video:

https://www.youtube.com/watch?v=VZMFp-mEWoM

I'm probably more concerned than he is in that a large market correction would be bad, but even scarier are the growing (in number and power) techo-religious folks who really think we are on the cusp of creating an electronic god and are willing to trash what's left of the planet in an attempt to birth that electronic god that they think will magically save us.

  Is it just me or is anyone else worried what will happen when the industry realizes LLMs are just not going to be the path to AGI? We are building nuclear power plants, massive datacenters that make our old datacenters look like toys.

Nothing will happen. LLMs even at GPT5 intelligence but scaled up with significantly with higher context size, faster inference speed, and lower cost would change the world.

In what way. So far almost all the change has been in spamming social media and customer support.

There are many ideas (society enhancing ones) that I have that I can't do due to context size being too low, inference being too slow, or price being too high.

For example, I want to use an LLM system to track every promise a politician makes ever and see if he or she actually did it. No one is going to give me $1 billion to do this. But I think it would enhance society.

I don't need an AGI to do this idea. I think current LLMs are already more than good enough. But LLM prices are still way to high to do this cost effectively. When inference is as cheap relatively as serving up a web page, LLMs will be ubiquitous.

Out of pure curiousity: do your really think that would make any significant difference for voter behavior? My understanding is that one of the biggest misunderstandings of the last decade+, was that center and leftist politicians assumed it would keep people from voting against their interests if you point out the lies of the relevant politicians and how their policies actually go against their voters interests. I mean that was the whole point of the fact-checking stuff gaining traction in mainstream-media during that time: just to be mostly abolished again because people are just not interested in truth and facts as much as we like to assume. Not not at all. But not as much as we tend to think.

Please don't get me wrong, I am not trying to be sarcastic here. I would love to see a perspective – just any perspective – of how to get out of the current political situation. Not just in the US but in many other countries the same playbook is followed by authoritarians with just as much success as in the US. So if you have material or some reasoning at hand why more information for the population would make a difference on the voting behaviour I would be super-interested. Thanks in advance!

I don't know if it will make any difference. But this sort of large scale data categorization is possible only with an LLM. Previously, you probably needed dozens of people working full time to do this. Now you just have a few GPUs do it.

I have a lot more ideas that are gated by low context size, inference speed, and price per token.

The bottomline is that we don't need AGI to change the world. Just more an cheaper compute.

I have been toying with idea to make candidate chooser on what members of parliament actually voted for. That data set is pretty limited and readily available. Just get the vote record and identify key bills. No LLM needed.

> Just get the vote record and identify key bills. No LLM needed.

How do you compare their vote record with the things they've said publicly?

> For example, I want to use an LLM system to track every promise a politician makes ever and see if he or she actually did it.

You're describing a list. Why do you need GPU farms to create a list?

How do you populate that list and verify the promises on the list?

You can't verify anything with a LLM.

> We are building nuclear power plants

Who is? There'll be about 10 nuclear reactor construction starts this year, largely either replacing end-of-life reactors, or in China (which had a well-established nuclear build-out prior to AI stuff). Beyond media hype, there's little reason to think that the AI bubble is actually leading to nuclear reactors being built anywhere.

> Are we being forced into a situation where we are invested too much in this to come face to face with it doesn't work and it makes everything worse?

Yes, but don't worry, they're getting close to the government so they'll get bailed out.

> What is this capacity being built for? It no longer makes any sense.

Give more power to Silicon Valley billionaires.

Reply on news.ycombinator.com