Agreed, it's when. They're hoping to stave it off or maybe stretch out the pop into a correction by all hedging together with all these incestuous deals, but you can't hold back the tide. They debuted this tech way too early, promised way too much, and now the market is wary about buying AI products until more noise settles out of the system.

▲

dylan604 8 hours ago | parent | next [-]

> They debuted this tech way too early, promised way too much,

finally, some rational thought into the AI insanity. The entire 'fake it til you make it' aspect of this is ridiculous. sadly, the world we live in means that you can't build a product and hold its release until it works. you have to be first to release even if it's not working as advertised. you can keep brushing off critiques with "it's on the road map". those that are not as tuned in will just think it is working and nothing nefarious is going on. with as long as we've had paid for LLM apps, I'm still amazed at the number of people that do not know that the output is still not 100% accurate. there are also people that use phrases as thinking when referring to getting a response. there's also the misleading terms like "searching the web..." when on this forum we all know it's not a live search.

	▲	burnte 5 hours ago \| parent [-]
		> sadly, the world we live in means that you can't build a product and hold its release until it works. you have to be first to release even if it's not working as advertised. You absolutely can and it's an extremely reliable path to success. The only thing that's changed is the amount of marketing hype thrown out by the fake-it vendors. Staying quiet and debuting a solid product is still a big win. > I'm still amazed at the number of people that do not know that the output is still not 100% accurate. This is the part that "scares" me. People who do not understand the tool thinking they're ACTUALLY INTELLIGENT. Not only are they not intelligent, they're not even ACTUALLY language models because few LLMs are actually trained on only language data and none work on language units (letters, words, sentences), tokens are abstractions from that. They're OUTPUT modelers. And they're absolutely not even close to being let loose unattended on important things. There are already people losing careers over AI crap like lawyers using AI to appeal sanctions because they had AI write a motion. Etc. And I think that was ultimately the biggest unforced error of these AI companies and the ultimate reason for the coming bubble crash. They didn't temper expectations at all, the massive gap between expectation and reality is already costing companies huge amounts of money, and it's only going to get worse. Had they started saying, "these work well, but use them carefully as we increase reliability" they'd be in a much better spot. In the past 2 years I've been involved in several projects trying to leverage AI, and all but one has failed. The most spectacular failure was Microsoft's Dragon Copilot. We piloted it with 100 doctors, after a few months we had a 20% retention rate, and by the end of a year, ONE doctor still liked it. We replaced it with another tool that WORKS, docs love it, and it was 12.6% the cost, literally a sixth the price. MS was EXTREMELY unhappy we canceled after a year, tried to throw discounts at us, but ultimately we had to say "the product does not work nearly as well as the competition."

▲

ares623 8 hours ago | parent | prev | next [-]

It’s going to pop as soon as they get confirmation the govt will bail them out. Until then they’re going to give it their all to keep it growing.

▲

pksebben 8 hours ago | parent [-]

I think they already have that confirmation. When we bailed the banks out in 08 we basically said "If you're big enough that we'd be screwed without you then take whatever risks you like with impunity".

That's a reduction of complexity, of course, but the core of the lesson is there. We have actually kept on with all the practices that led to the housing crash (MBS, predatory lending, Mixing investment and traditional banking).

▲

n8cpdx 8 hours ago | parent | next [-]

> If you're big enough that we'd be screwed without you then take whatever risks you like with impunity".

I know financially it will be bad because number not go up and number need go up.

But do we actually depend on generative/agentic AI at all in meaningful ways? I’m pretty sure all LLMs could be Thanos snapped away and there would be near zero material impact. If the studies are at all reliable all the programmers will be more efficient. Maybe we’d be better off because there wouldn’t be so much AI slop.

It is very far from clear that there is any real value being extracted from this technology.

The government should let it burn.

Edit: I forgot about “country girls make do”. Maybe gen AI is a critical pillar of the economy after all.

▲

thewebguyd 6 hours ago | parent | next [-]

> I’m pretty sure all LLMs could be Thanos snapped away and there would be near zero material impact.

I mostly agree, but I don't think it's the model developers that would get bailed out. OpenAI & Anthropic can fail, and should be let to fail if it comes to that.

Nvidia is the one that would get bailed out. As would Microsoft, if it came to that.

I also think they should be let to fail, but there's no way the US GOV ever allows them to.

	▲	palmotea 4 hours ago \| parent [-]
		> Nvidia is the one that would get bailed out. As would Microsoft, if it came to that. > I also think they should be let to fail, but there's no way the US GOV ever allows them to. There's different ways to fail, though: liquidation, and a reorganization that wipes out the shareholders. OpenAI could be liquidated and all its technology thrown in to the trash, and I wouldn't shed a tear, but Microsoft makes (some) stuff (cough, Windows) that has too much stuff dependent on it to go away. The shareholders can eat it (though I think broad-based index funds should get priority over all other shareholders in a bankruptcy).

▲

pksebben 7 hours ago | parent | prev [-]

I expect the downvotes to come from this as they always seem to do these days, but I know from my personal experience that there is value in these agents.

Not so much for the work I do for my company, but having these agents has been a fairly huge boon in some specific ways personally:

- search replacement (beats google almost all of the time)

- having code-capable agents means my pet projects are getting along a lot more than they used to. I check in with them in moments of free time and give them large projects to tackle that will take a while (I've found that having them do these in Rust works best, because it has the most guardrails)

- it's been infinitely useful to be able to ask questions when I don't know enough to know what terms to search for. I have a number of meatspace projects that I didn't know enough about to ask the right questions, and having LLMs has unblocked those 100% of the time.

Economic value? I won't make an assessment. Value to me (and I'm sure others)? Definitely would miss them if they disappeared tomorrow. I should note that given the state of things (large AI companies with the same shareholder problems as MAANG) I do worry that those use cases will disappear as advertising and other monetizing influences make their way in.

Slop is indeed a huge problem. Perhaps you're right that it's a net negative overall, but I don't think it's accurate to say there's not any value to be had.

	▲	s1mplicissimus 5 hours ago \| parent [-]
		I'm glad you had positive experiences using this specific technology. Personally, I had the exact opposite experience: Wrong, deceitful responses, hallucinations, arbitrary pointless changes to code... It's like that one junior I requested to be removed from the team after they peed in the codebase one too many times. On the slop i have 2 sentiments: Lots of slop = higher demand for my skills to clean it up. But also lots of slop = worse software on probably most things, impacting not just me, but also friends, family and the rest of humanity. At least it's not only a downside :/

▲

exasperaited 8 hours ago | parent | prev | next [-]

Which of these firms is too big to fail though?

It all depends on whether MAGA survives as a single community. One of the few things MAGA understands correctly is that AI is a job-killer.

Trump going all out to rescue OpenAI or Anthropic doesn't feel likely. Who actually needs it, as a dependency? Who can't live without it? Why bail out entities you can afford to let go to the wall (and maybe then corruptly buy out in a fire sale)?

Similarly, can you actually see him agreeing to bail out Microsoft without taking an absurd stake in the business? MAGA won't like it. But MS could be broken up and sold; every single piece of that business has potential buyers.

Nvidia, now that I can see. Because Trump is surrounded by crypto grifters and is dependent on crypto for his wealth. GPUs are at least real solid products and Nvidia still, I think, make the ones the crypto guys want.

Google, you can see, are getting themselves ready to not be bailed out.

	▲	Capricorn2481 5 hours ago \| parent [-]
		> One of the few things MAGA understands correctly is that AI is a job-killer Trump (and by extension MAGA) has the worst job growth of any President in the past 50 years. I don't think that's their brand at all. They put a bunch of concessions to AI companies in the Big Beautiful Bill, and Trump is not running again. He would completely bail them out, and MAGA will believe whatever he says, and congress will follow whatever wind is blowing.

▲

bigbuppo 8 hours ago | parent | prev [-]

If Meta or Google disappared overnight, it would be, at worst, a minor annoyance for most of the world. Despite the fact that both companies are advertising behemoths, marketing departments everywhere would celebrate their end.

▲

palmotea 4 hours ago | parent | next [-]

> If Meta or Google disappared overnight, it would be, at worst, a minor annoyance for most of the world.

I'll grant you Meta, but losing Google in that way would be highly disruptive because so many people have their primary email account on it.

▲

NicoJuicy 7 hours ago | parent | prev | next [-]

Watching this on my Android on a chrome browser.

Hard disagree

▲

lazide 7 hours ago | parent | prev [-]

Considering how much of the world runs on WhatsApp alone, that’s laughably wrong.

▲

hermanzegerman 6 hours ago | parent | next [-]

Then they would just use another Messenger or fall back on RCS/SMS.

The only reason WhatsApp is so popular, is because so many people are on it, but you have all you need (their phone number) to contact them elsewhere anyway

▲

diroussel 6 hours ago | parent | prev [-]

So if WhatsApp had an outage, but you needed to communicate to someone, you wouldn't be able to? Don't you have contacts saved locally, and other message apps available?

▲

lazide 6 hours ago | parent [-]

In most of Asia, Latin America, Africa, and about half of Europe?

You’d be pretty stuck. I guess SMS might work, but it wouldn’t for most businesses (they use the WhatsApp business functionality, there is no SMS thing backing it).

Most people don’t even use text anymore. China has it’s own Apps, but everyone else uses WhatsApp exclusively at this point.

	▲	hermanzegerman 6 hours ago \| parent [-]
		Nobody uses WhatsApp Business in Germany, Austria or Switzerland in a way that you would be stuck without

▲

project2501a 8 hours ago | parent | prev | next [-]

They got enough slush money to make this go on for a couple of years.

I am shocked at the part they know it is a bubble and they are doing nothing to amortize it. Which means they expect the government to step in and save their butts.

... Well, not that shocked.

	▲	bigbuppo 8 hours ago \| parent [-]
		They're floating 40 year bonds for technology with a three year lifecycle. They do not have the actual cash for this.

▲

somewhereoutth 4 hours ago | parent | prev [-]

Yet another [1] AI winter [2]

[1] https://en.wikipedia.org/wiki/AI_winter#The_setbacks_of_1974

[2] https://en.wikipedia.org/wiki/AI_winter#AI_winter_of_the_199...