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

The Generative AI Con

wheresyoured.at

Post reply on HN
Print view View original

451–460 of 503 posts

Re: The Generative AI Con

#451

Post by HDThoreaun »

jcgrillo wrote: post #237
Aeolun wrote: post #206

Earlier quoted context omitted.

> That's not worth billions. It's definitely not worth trillions. That is a problem for the VC’s that bet wrong, not for the world at large. The models exist now and they’ll keep being used, regardless of whether a bunch of rich guys lost a bunch of money.

> they’ll keep being used How? I get that many devs like using them for writing code. Personally I don't, but maybe someday someone will invent a UX for this that I don't despise, and I could be convinced. So what? That's a tiny market. Where in the landscape of b2b and b2c software do LLMs actually find market fit? Do you have even one example? All the ideas I've heard so far are either science fiction (just wait an…

LLMs are incredible at editing my writing. Every email I write is improved by LLMs. My executive summaries are improved by LLMs. It wont be long until every single office worker is using LLMs as an integral part of their daily stack, people just have to try it and theyll see how useful it is for writing.

Microsoft turned itself into a trillion dollar company off the back of enterprise SAAS products and LLMs are among the most useful.

Re: The Generative AI Con

#452

Post by habinero »

Piskvorrr wrote: post #315
stevage wrote: post #154

Earlier quoted context omitted.

Huh, for me it's the opposite. It does the boring bit, writing pedestrian method bodies. Writing import statements. Closing tags. I do the fun bit: having creative ideas and trying them out.

Looks like you haven't used a decent IDE: these things have been standard for decades , locally and with minimal requirements. But wait, now it happens in the Cloud (meh, that's not gonna fly anymore, too last decade)...AND requires massive amounts of power AND cooling, PLUS it's FUBAR about 50/50. For an incremental improvement...not great, not terrible.

I think LLMs are vastly overhyped and mostly useless, but I use Copilot as glorified autocomplete and like it.

It does what the other poster said: it automates the boring parts of "this db model has eight fields that are mostly what you expect" and it autocompletes them mostly accurately.

That's not something an IDE does.

Re: The Generative AI Con

#454

Post by yunwal »

gamblor956 wrote: post #420
bko wrote: post #391

Earlier quoted context omitted.

The author betrays himself early with this line: > a cynical bubble inflated by OpenAI CEO Sam Altman built to sell into an economy run by people that have no concept of labor other than their desperation to exploit or replace it. He brings up the concept of labor and applies a moral judgement about "replacing" and "exploiting" labor. And then he throws the kitchen sink at the technology. People use it sure, but it's…

Very few people get real value from an LLM. They're useful for minor grammatical checks, and idiot-level summaries. Maybe for entry-level coding. Outside of that, they're basically just gimmicks. You can't trust the output at all; literally everything an LLM outputs must be verified before it can be used for anything important, so you end up doing all of the work you thought you would have avoided by using an LLM.

> literally everything an LLM outputs must be verified

Until you prove P=NP I’ll take that as a win

Re: The Generative AI Con

#455

Post by Aeolun »

jcgrillo wrote: post #237
Aeolun wrote: post #206

Earlier quoted context omitted.

> That's not worth billions. It's definitely not worth trillions. That is a problem for the VC’s that bet wrong, not for the world at large. The models exist now and they’ll keep being used, regardless of whether a bunch of rich guys lost a bunch of money.

> they’ll keep being used How? I get that many devs like using them for writing code. Personally I don't, but maybe someday someone will invent a UX for this that I don't despise, and I could be convinced. So what? That's a tiny market. Where in the landscape of b2b and b2c software do LLMs actually find market fit? Do you have even one example? All the ideas I've heard so far are either science fiction (just wait an…

> Do you have even one example?

My company uses them for a fuckton of things that were previously too intractable for static logic to work (because humans are involved).

This is mostly in the realm of augmented customer support (e.g. customer says something, and the support agent immediately gets the summarized answer on their screen)

It’s nothing that can’t be done without, but when the whole problem can be simplified to “write a good prompt” a lot of use cases are suddenly within reach.

It’s a question if they’ll keep it around when they realize it doesn’t always quite work, but at least right now MS is making good money off of it.

Re: The Generative AI Con

#456

Post by habinero »

paulcole wrote: post #340
lukevp wrote: post #12

I’m a little shocked at how much negativity there is around LLMs among developers. It’s a new tool that requires some learning, and it’s sometimes not so great, but if you’ve used an IDE with real coding assistance built in (eg. VS Code in Edit with Copilot mode - NOT Chat mode, using Claude 3.5), it’s honestly not much worse than a junior dev and 100x faster. And if the code is bad you throw it away and try again 10…

> I’m a little shocked at how much negativity there is around LLMs among developers There's a Quentin Tarantino quote where he says there are 2 kinds of film critics. There are those who love movies and there are those that love the movies they love. A lot of developers really seem to love the technology they love. These people are where most of the negativity is coming from. And my guess is that the people who are e…

I use Copilot for autocompleting the boring boilerplate. I like it. I also think LLMs are mostly useless.

It's not the technology, it's the stupid overhype. It really feels like all the HODL bitcoin cultists have finally gotten over their lost apes and found a new technogod.

So many people in these threads are convinced it's about to gain sentience. That's not going to happen. You get the people outing themselves by saying "it does my job better than me!"

If you say something honest and direct like "their output is mediocre and unreliable" or "the RNG text generator is not capable of thinking, you're just Clever Hans-ing yourself" or "if it does your job better than you, that says more about you than about it", you get people clutching their pearls like a Stanford parent whose kid got a D.

arXiv has turned into a paper mill of AI startups uploading marketing hype cosplaying as "research".

Re: The Generative AI Con

#457

Post by SZJX »

trhway wrote: post #101
automatic6131 wrote: post #87

Earlier quoted context omitted.

This is the crux. A cool thing has been invented, with real usages. Unfortunately, it's cost hundreds of billions of dollars and it has absolutely zero hope of making the trillions needed to justify that. Now someone will respond about how it's just a stepping stone, and how the billions are justified by _something completely imaginary, and not invented yet, and maybe not ever_ e.g. agents.

>it's cost hundreds of billions of dollars and it has absolutely zero hope of making the trillions needed to justify that. The BigTech companies have been flush with liquidity and poured those hundreds of billions into the promising tech, and as result we got a wonderful new technology. There is not much need for those trillions in return - just look at liquidity positions of those companies, they are just fine. If t…

They’ll be fine and will survive regardless, but their current astronomical valuations probably won’t be.

Re: The Generative AI Con

#458

Post by yunwal »

mylastattempt wrote: post #130
danroblew wrote: post #92

Earlier quoted context omitted.

Trillions of dollars, dude. They need to make trillions of dollars to satisfy their investors.

If you are correcting my use of millions to trillions: I was refering to the author himself, who writes like this giant AI bubble is pushing him forward to the edge of the cliff as people keep believing in it, and he is desperately trying to get the bubble to shrink or he'll fall off and die. Methaphorically. But why does he act or feel that way? Let the trillions be lost, it's just how hypes, bubbles and the stock m…

> Let the trillions be lost

Haven’t read any news in the past 20 years? This is all going to be funded by the American taxpayer

Re: The Generative AI Con

#459

Post by habinero »

vonneumannstan wrote: post #412
jcgrillo wrote: post #338

Earlier quoted context omitted.

> top tier people Who? How? This is not what I've seen where I work. There's a bunch of hubbub and generalized excitement, and lots of talk about what could be done, or what might be done, but not very much actual doing . I must just be a clueless "mid".

Yeah unironically. Guido van Rossum - "I use it every day. My biggest adjustment with using Copilot was that instead of writing code, my posture shifted to reviewing code." https://www.youtube.com/watch?v=-DVyjdw4t9I Here's Jeff Dean saying 25% of the characters in new PRs at Google are AI Generated. https://www.dwarkeshpatel.com/p/jeff-dean-and-noam-shazeer Andrej Karpathy - "I basically can't imagine going back to…

These kinds of responses are my favorite dark pattern rhetorical device, because you can assert literally _anything_ in this format and almost nobody will refute you, because the cost of refuting bullshit is 100x the cost of producing it.

Anyways, here goes.

1. Guido uses Copilot like I do - as a StackOverflow replacement to write the dumb boilerplate code. A much less flattering quote is "It doesn't save me much thinking, but [it helps] because I'm a poor typist". Also it's literally a minute or two of a three hour podcast.

2. A lot of code is autogenerated lol. Again, it's all the boring boilerplate stuff.

3. The cofounder of OpenAI is a biased source lol

4. He's an AI researcher, of course he runs that stuff.

5. Again, similar to Guido. He's using it for the boilerplate. Nothing wrong with enjoying using it as a toy, as he is here. But he's not doing serious work with it.

There's no virtue in hyping this stuff like a HODL bitcoin cultist.

Re: The Generative AI Con

#460

Post by yunwal »

sambapa wrote: post #407
sebastiennight wrote: post #396

Earlier quoted context omitted.

This is incorrect. If you take the most basic interpretation of an LLM at temperature 0 as predicting the most likely token, and you run it on, say, 1,000 runs of "complete this Spanish sentence with the word for 'X'", then: - maybe ALL humans would fail the test in some way, eg. let's say everybody gets at least 10 of those wrong, and the average person gets 100 of those wrong. - still, as long as most people correc…

But still, the questions in that test are "solved" in the sense of "I can take a dictionary and answers these questions with full certainty". Beyond established knowledge LLMs are monkeys with typewriters, at best.

I’d like to see you ace even a middle-school level Spanish test with just a dictionary (sub Spanish with some other language if you happen to know Spanish).
Post reply on HN