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

AI Is Too Big to Fail

sibylline.dev

Post reply on HN
Print view View original

111–120 of 140 posts

Re: AI Is Too Big to Fail

#111

Post by miningape »

travisgriggs wrote: post #106
cudgy wrote: post #103

Earlier quoted context omitted.

401k is voluntary

I guess. My company (and I believe most) highly incentivize it by offering matching funds. The company says, "want to save $1000 of your paycheck? We'll throw in $500". The general wisdom is, you'd be an idiot to leave money like that going unreclaimed.

Where I am there's actually a %-age of your salary/wage that's "returned" to you when you retire. You don't get a choice.

You can choose to contribute to an employer-match scheme (I do) or even fund your own pension account with monthly payments to a pension provider. But these are in addition to the above forced pension.

Re: AI Is Too Big to Fail

#112

Post by array_key_first »

parrellel wrote: post #36
philipwhiuk wrote: post #18

Earlier quoted context omitted.

Pensioners and retirees willingly dumping money into index funds without oversight are the dumb money that enables all of this. So yes, they deserve a haircut.

Or perhaps we shouldn't have killed off the concept of pensions in the United States so we weren't all beholden to godforsaken 401Ks?

To be fair, the pensions also exploded in a lot of cases.

We really just need a more centralized and reliable method of holding money for retirement. Which we have, but we need to expand - SS.

Of course, lots of people think SS will explode, too. But it hasn't! Yet!

Re: AI Is Too Big to Fail

#113

Post by devin »

PerfectElement wrote: post #51
Fernicia wrote: post #14

Earlier quoted context omitted.

When did the HackerNews comment section turn into this? Low quality, aggressive, fervently anti-establishment.

I agree. HN basically became Reddit.

This comment of yours contributes nothing and is exactly the level one would expect of Reddit. If you want better, start by doing better yourself.

Re: AI Is Too Big to Fail

#114

Post by notepad0x90 »

zzleeper wrote: post #98
notepad0x90 wrote: post #58

I think a lot of HNers are too close to the topic for the right perspective. LLMs are as much of a feature of an existing technology as computers are features of electricity. ML and neural networks have been around for a long time and the transition to LLM usage for us feels like "yup, another feature, just a really good auto-complete". For most users of LLMs, this is a magical feature that answers any questions and…

That kinda misses the point for me. I use LLMs a lot and have become an invaluable part of my work (coding, collecting data, proof-reading text) and even personal time (amazing Japanese menu translator!). But I still think they can "fail" in the sense that current investments need an absurd amount of profits in order to have a positive NPV. Data centers depreciate fast, so if you just did $500bn in capex, you better…

So long as there is a reasonable earning potential, or actually a reasonable expectation that the market will expect a decent earning potential the capital will be there.

Unlike crypto (which has yet to fail as a category, just dips), the elephants taking up the tens of billions in investments aren't purely speculative pyramid schemes.

Re: AI Is Too Big to Fail

#115

Post by stevefan1999 »

I'd say the problem is that the fiat currency keeps inflating and thus decreasing in value day-by-day, while AI does not bring that much productivity gain compared to last few decades.

Internet boom does substantially decrease communication cost and bring more people together, making much more liquidity. AI, meanwhile, is not only not flawless (and sometimes fatal, see some horror stories of LLM here in HN doing an oopsie deleting production database and do nothing about recovery), but also a costy beast that only corporations or rich individuals could own, in other words, most people can "download" the "means of production", but not totally use and own it.

And the biggest problem for AI, still, that compared to Internet, is that AI does not make you that much gain in productivity. Back when we are in the office we use pen and paper, now we use Word or LibreOffice depending on where you are, and PC is therefore a must. Now I don't think AI is going to be a must, and they are still shoveling NPU that would end up being a waste in silicon space, if not vaporware.

Re: AI Is Too Big to Fail

#117

Post by estimator7292 »

thewebguyd wrote: post #5
jqpabc123 wrote: post #3

If it wasn't for AI investments, it's likely the United States would be in a recession right now. So the biggest achievement of this amazing new technology is to artificially prop up Wall Street? For those who can read the tea leaves --- it's absolutely a bubble --- a financial bubble. The real reason why China is ahead in many areas is because their leadership is a little more technologically savvy and progressive t…

If this is a pre-war economy, the US is screwed not because of AI but because we started a stupid trade war that's going to cut us off of our supply of dysprosium which is critical for defense manufacturing.

We're doubly screwed because we dismantled most of our industrial base and everyone is scrambling to produce basic goods domestically

Re: AI Is Too Big to Fail

#118

Post by deaux »

candiddevmike wrote: post #24

IMO, the backdrop to all of this is global warming. We're seeing so many warning signs and growth stagnating that governments need something, ANYTHING, to prop up their debt fueled economies and keep the music playing. In the face of impending GDP/population contraction due to global warming and the upcoming breadbasket collapses, GenAI isn't going to do shit for us. All of the money and effort being spent on it is m…

You're having kids while admitting they don't have a glimmer of hope? Ouch.

Re: AI Is Too Big to Fail

#119

Post by Zababa »

candiddevmike wrote: post #24

IMO, the backdrop to all of this is global warming. We're seeing so many warning signs and growth stagnating that governments need something, ANYTHING, to prop up their debt fueled economies and keep the music playing. In the face of impending GDP/population contraction due to global warming and the upcoming breadbasket collapses, GenAI isn't going to do shit for us. All of the money and effort being spent on it is m…

I don't think global warming is the cause of population contraction.

Re: AI Is Too Big to Fail

#120

Post by rwyinuse »

candiddevmike wrote: post #24

IMO, the backdrop to all of this is global warming. We're seeing so many warning signs and growth stagnating that governments need something, ANYTHING, to prop up their debt fueled economies and keep the music playing. In the face of impending GDP/population contraction due to global warming and the upcoming breadbasket collapses, GenAI isn't going to do shit for us. All of the money and effort being spent on it is m…

I think major breadbasket collapses can be avoided through geoengineering. Thanks to volcanoes and emissions from marine traffic, we know for a fact that climate change can be slowed down by spraying enough stuff reflecting sunlight high in the atmosphere.

Of course some conservative US states are enacting laws against geoengineering, but thankfully they can't prevent rest of the world from doing it. I hope we'll eventually reach some kind of a solution where Trump-voting radical Christians can have the biblical apocalypse they want locally, while rest of the world continues to improve.

Post reply on HN