Recent AI model progress feels mostly like bullshit • Hacker News Board

Re: Recent AI model progress feels mostly like bullshit

#131

Post by leereeves » Sun, Apr 06, 2025, 9:19 PM UTC

pclmulqdq wrote: post #99

blitzar wrote: post #56

Earlier quoted context omitted.

Does the as yet unwritten prequel of Idiocracy tell the tale of when we started asking Ai chat bots for facts and this was the point of no return for humanity?

It turns out there's huge demand for un-monetized web search.

I like that it's unmonetized, of course, but that's not why I use AI. I use AI because it's better at search. When I can't remember the right keywords to find something, or when the keywords aren't unique, I frequently find that web search doesn't return what I need and AI does.

It's impressive how often AI returns the right answer to vague questions. (not always though)

Re: Recent AI model progress feels mostly like bullshit

#132

Post by pclmulqdq » Sun, Apr 06, 2025, 9:20 PM UTC

leereeves wrote: post #131

pclmulqdq wrote: post #99

Earlier quoted context omitted.

It turns out there's huge demand for un-monetized web search.

I like that it's unmonetized, of course, but that's not why I use AI. I use AI because it's better at search. When I can't remember the right keywords to find something, or when the keywords aren't unique, I frequently find that web search doesn't return what I need and AI does. It's impressive how often AI returns the right answer to vague questions. (not always though)

Google used to return the right answer to vague questions until it decided to return the most lucrative answer to vague questions instead.

Re: Recent AI model progress feels mostly like bullshit

#133

Post by spudlyo » Sun, Apr 06, 2025, 9:20 PM UTC

pclmulqdq wrote: post #99

blitzar wrote: post #56

Earlier quoted context omitted.

Does the as yet unwritten prequel of Idiocracy tell the tale of when we started asking Ai chat bots for facts and this was the point of no return for humanity?

It turns out there's huge demand for un-monetized web search.

Soon sadly, there will be a huge demand for un-monetized LLMs. Enshitification is coming.

Re: Recent AI model progress feels mostly like bullshit

#134

Post by aspenmayer » Sun, Apr 06, 2025, 9:21 PM UTC

stavros wrote: post #101

more_corn wrote: post #87

Earlier quoted context omitted.

Ironically though an LLM powered search engine (some word about being perplexed) is becoming way better than the undisputed king of traditional search engines (something oogle)

That's because they put an LLM over a traditional search engine.

Google Labs has AI Mode now, apparently.

https://labs.google.com/search/experiment/22

Re: Recent AI model progress feels mostly like bullshit

#135

Post by MegaButts » Sun, Apr 06, 2025, 9:21 PM UTC

fancyfredbot wrote: post #129

MegaButts wrote: post #122

Earlier quoted context omitted.

I vehemently disagree. If I ask a question with an objective answer, and it simply makes something up and is very confident the answer is correct, what the fuck has it understood other than how to piss me off? It clearly doesn't understand that the question has a correct answer, or that it does not know the answer. It also clearly does not understand that I hate bullshit, no matter how many dozens of times I prompt i…

It didn't understand you but the response was plausible enough to require fact checking. Although that isn't literally indistinguishable from 'understanding' (because your fact checking easily discerned that) it suggests that at a surface level it did appear to understand your question and knew what a plausible answer might look like. This is not necessarily useful but it's quite impressive.

There are times it just generates complete nonsense that has nothing to do with what I said, but it's certainly not most of the time. I do not know how often, but I'd say it's definitely under 10% and almost certainly under 5% that the above happens.

Sure, LLMs are incredibly impressive from a technical standpoint. But they're so fucking stupid I hate using them.

> This is not necessarily useful but it's quite impressive.

I think we mostly agree on this. Cheers.

Re: Recent AI model progress feels mostly like bullshit

#136

Post by maccard » Sun, Apr 06, 2025, 9:21 PM UTC

ohgr wrote: post #107

maccard wrote: post #28

Earlier quoted context omitted.

> where’s the business model? For who? Nvidia sell GPUs, OpenAI and co sell proprietary models and API access, and the startups resell GPT and Claude with custom prompts. Each one is hoping that the layer above has a breakthrough that makes their current spend viable. If they do, then you don’t want to be left behind, because _everything_ changes. It probably won’t, but it might. That’s the business model

You missed the end of the supply chain. Paying users. Who magically disappear below market sustaining levels of sales when asked to pay.

I never said it was sustainable, and even if it was, OP asked for a business model. Customers don’t need a business model, they’re customers.

The same is true for any non essential good or service.

Re: Recent AI model progress feels mostly like bullshit

#137

Post by stavros » Sun, Apr 06, 2025, 9:22 PM UTC

aspenmayer wrote: post #134

stavros wrote: post #101

Earlier quoted context omitted.

That's because they put an LLM over a traditional search engine.

Google Labs has AI Mode now, apparently. https://labs.google.com/search/experiment/22

Hm, that's not available to me, what is it? If its an LLM over Google, didn't they release that a few months ago already?

Re: Recent AI model progress feels mostly like bullshit

#138

Post by maccard » Sun, Apr 06, 2025, 9:23 PM UTC

grey-area wrote: post #130

maccard wrote: post #37

Earlier quoted context omitted.

Which one? Nvidia are doing pretty ok selling GPU's, and OpenAI and Anthropic are doing ok selling their models. They're not _viable_ business models, but they could be.

NVDA will crash when the AI bubble implodes, and none of those Generative AI companies are actually making money, nor will they. They have already hit limiting returns in LLM improvements after staggering investments and it is clear are nowhere near general intelligence.

All of this can be true, and has nothing to do with them having a business model.

> NVDA will crash when the AI bubble implodes, > making money, nor will they > They have already hit limiting returns in LLM improvements after staggering investments > and it is clear are nowhere near general intelligence.

These are all assumptions and opinions, and have nothing to do with whether or not they have a business model. You mightn't like their business model, but they do have one.

Re: Recent AI model progress feels mostly like bullshit

#139

Post by fancyfredbot » Sun, Apr 06, 2025, 9:23 PM UTC

bobsmooth wrote: post #90

MegaButts wrote: post #68

Earlier quoted context omitted.

> LLMs aren't good at being search engines, they're good at understanding things. LLMs are literally fundamentally incapable of understanding things. They are stochastic parrots and you've been fooled.

What do you call someone that mentions "stochastic parrots" every time LLMs are mentioned?

That makes me think, has anyone ever heard of an actual parrot which wasn't stochastic?

I'm fairly sure I've never seen a deterministic parrot which makes me think the term is tautological.

Re: Recent AI model progress feels mostly like bullshit

#140

Post by kayo_20211030 » Sun, Apr 06, 2025, 9:29 PM UTC

aerhardt wrote: post #34

My mom told me yesterday that Paul Newman had massive problems with alcohol. I was somewhat skeptical, so this morning I asked ChatGPT a very simple question: "Is Paul Newman known for having had problems with alcohol?" All of the models up to o3-mini-high told me he had no known problems. Here's o3-mini-high's response: "Paul Newman is not widely known for having had problems with alcohol. While he portrayed charact…

This may have hit the nail on the head about the weaknesses of LLM's.

They're going to regurgitate something not so much based on facts, but based on things that are accessible as perceived facts. Those might be right, but they might be wrong also; and no one can tell without doing the hard work of checking original sources. Many of what are considered accepted facts, and also accessible to LLM harvesting, are at best derived facts, often mediated by motivated individuals, and published to accessible sources by "people with an interest".

The weightings used by any AI should be based on the facts, and not the compounded volume of derived, "mediated", or "directed" facts - simply, because they're not really facts; they're reports.

It all seems like dumber, lazier search engine stuff. Honestly, what do I know about Paul Newman? But, Joanne Woodward and others who knew and worked with him should be weighted as being, at least, slightly more credible that others; no matter how many text patterns "catch the match" flow.