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

The L in "LLM" Stands for Lying

acko.net

Post reply on HN
Print view View original

351–360 of 500 posts

Re: The L in "LLM" Stands for Lying

#351

Post by Marha01 »

simonask wrote: post #236
mikkupikku wrote: post #112

Earlier quoted context omitted.

I think that's true, but something even more subtle is going on. The quality of the LLM output depends on how it was prompted in a way more profound than I think most people realize. If you prompt the LLM using jargon and lingo that indicate you are already well experienced with the domain space, the LLM will rollplay an experienced developer. If you prompt it like you're a clueless PHB who's never coded, the LLM wil…

This is magical thinking. LLMs are physically incapable of generating something “well thought out”, because they are physically incapable of thinking.

It is magical thinking to claim that LLMs are definitely physically incapable of thinking. You don't know that. No one knows that, since such large neural networks are opaque blackboxes that resist interpretation and we don't really know how they function internally.

You are just repeating that because you read that before somewhere else. Like a stochastic parrot. Quite ironic. ;)

Re: The L in "LLM" Stands for Lying

#352

Post by tovej »

vntok wrote: post #114
Jooror wrote: post #74

Earlier quoted context omitted.

I’m curious about how you landed “git gud; prompt better” and not “maybe the domain I work in is a better fit for LLM code”. Or, to be a bit less generous, consider the possibility that the code you’re generating is boilerplate, marshaling, and/or API calls. A facade of perceived complexity over something that’s as complex as a filter-map or two.

The parent is specifically talking about producing boilerplate code -a domain in which LLM excell at- and not having had any success at that. It's therefore not a leap of logic to assume they haven't put (enough) effort into getting better at prompting first, which is perfectly fine per se but leans towards a skill issue and not an immutable property of gen AI. The uncomfortable fact remains that one cannot really ex…

That is not at all what I said, please read my post more carefully before speculating.

I am talking about using LLMs in general, not for boiler plate specifically.

My point about boilerplate is that I have tools that solve this for me already, and do it in a more predictable way.

Re: The L in "LLM" Stands for Lying

#353

Post by miningape »

TimTheTinker wrote: post #205
larodi wrote: post #184

Earlier quoted context omitted.

> No one cares about how the code is written. I would overstate: No one even cares how architecture is done. Unless you are the one fixing it or maintaining it. Sorry, no one. We all know Apple did some great stuff with their code, but we care more about the awful work done on the UI, right? I mean - the UI seems to not be breaking in these new OSs which is amazing feature... for a game perhaps, and most likely the c…

This is overstating it. Computers are amazing machines, and modern operating systems are also amazing. But even they cannot completely mask the downstream effects of poor quality code. You say you don't care, but I bet you do when you're dealing with a problem caused by poor code quality or bad choices made by the developer.

Yep willing to bet that the majority of people saying "users don't care how well the code is written" will crash out when some software they're using is slow and buggy, even more extremely if it glitches and deletes their work.

Just like how most people don't care how well a bridge is designed... until it collapses.

Re: The L in "LLM" Stands for Lying

#354

Post by banannaise »

tanjtanjtanj wrote: post #232
nikitau wrote: post #164

Earlier quoted context omitted.

Roguelike/lites are is of the most popular genres of indie games nowadays. One of it's main characteristics is randomization and procedural generation.

While there are many Roguelikes with procedural generation, I think the most popular ones do not. Slay the Spire, Risk of Rain 2, Hades 1/2, BoE etc are all handmade stages with a random order with randomized player powers rather than procedurally generated.

I've seen a couple roguelike developers report that they played around with procedural generation, but it was difficult to prevent it from creating dungeons that were bad, unfun, or just straight-up killscreens. Turns out it's often easier to simply hand-draw good maps than to get the machine to generate okay-to-good ones.

Procedural generation is good when variety matters more than quality, which is a relatively rare occurrence.

Re: The L in "LLM" Stands for Lying

#355

Post by AnotherGoodName »

raincole wrote: post #34

> Video games stand out as one market where consumers have pushed back effectively No, it's simply untrue. Players only object against AI art assets. And only when they're painfully obvious. No one cares about how the code is written. If you actually read the words used in Steam AI survey you'll know Steam has completely caved in for AI-gen code as well. It's specifically worded like this: > content such as artwork,…

Games with ai art assets are some of the most popular right now in any case. Arc raiders being a great example where some of the voice assets are AI generated.

Be careful of reading any viewpoint on the internet. Apparently no one used facebook or instagram and everyone boycotts anything with ai in it.

In reality i think you’d be foolish not to make use of the tools available. Arc Raiders did the right thing by completely ignoring those sorts of comments. There may be a market for 100% organic video games but there’s also a market for mainstream ‘uses every ai tool available’ type of games.

Re: The L in "LLM" Stands for Lying

#356

Post by kevstev »

salad-tycoon wrote: post #181
hermannj314 wrote: post #138

Earlier quoted context omitted.

only code anyone will be touching in a museum in 800 years will be the good code. I hope they don't talk about what great craftsmen we all were because someone saw an original Fabrice Bellard at the Louvre. Survivor bias plays a role in glorifying the past.

Agreed, I think the good gained by wool mills is greater in that little Timmy is less likely to lose a leg to frostbite than the bad loss of my scarf not passing through a ring. Long term though, I’ve always wondered if the Amish turn out to be the only survivors.

I am kind of lost here on this whole scarf through a ring thing as well. This is just a function of the thickness of the scarf? My wife went through a scarf phase about a decade ago, and I am pretty sure a Pucci scarf could easily fit through a typical sized ring meant to go on a finger?

Its entirely possible that old manufacturing methods produced things that are different, but I would be entirely surprised if they are entirely better overall. If the defining metric for scarves is how well they fit through rings, I am sure they would all be made so you could fit 3 through a ring if people were willing to pony up for that. If you look at a lot of old clothes, they are generally a lot heavier, but I am not sure I would really want to wear them, they look quite uncomfortable. I also think its wonderful that today you can get a set of clothes for a few hours of minimum wage work while in the past this was a major investment. You can also choose to pay thousands for a shirt if you wish, but from 10 feet away its going to be hard to tell the difference.

Re: The L in "LLM" Stands for Lying

#358

Post by tylervigen »

Retr0id wrote: post #346

> Classic procedural generation is noteworthy here as a precedent, which gamers were already familiar with, because by and large it has failed to deliver. With the notable exception of Minecraft terrain generation, which I think most would say was successful in what it set out to achieve.

There are tons of examples of this. Heck, even Tetris has procedural generation. I think this argument was a mistake.

Re: The L in "LLM" Stands for Lying

#359

Post by ducttapecrown »

tptacek wrote: post #338
wolvesechoes wrote: post #155

I am bit tired of such discussions. I don't care if LLMs are good at coding or bad at it (in my experience the answer is "it depends"). I don't care how good are they at anything else. What matters in the end is that this tech is not to empower a common person (although it could). It is not here to make our lives better, more worthwhile, more satisfying (it could do these as well). It is there to reduce our agency, t…

This argument can be used, and has been used, about every innovation in automation since the dawn of the industrial revolution.

It is not the technology that sucks ever more money out of the populace, it's the people at the top!

Re: The L in "LLM" Stands for Lying

#360

Post by qsera »

pixl97 wrote: post #327
qsera wrote: post #192

Earlier quoted context omitted.

>Me labeling "authentic" it or not should not affect it's artistic value. The problem with automated imitation generators is that they can produce thousands of painting that imitate Van Gogh, but does not have the same soul. It is the same reason why these things cannot create genuinely funny jokes. They cannot assess the funnyness of the themselves. They cannot feel, and cannot do the filtering based on emotion. It…

>but does not have the same soul. Define soul, how about a legal/scientific description that accurately covers all bases? The funny jokes thing is funny too, if someone told you a joke and you thought it was funny, then they told you it was from an LLM, would it stop being funny.

That is why I said it is hard to spot an souless painting.

>if someone told you a joke and you thought it was funny, then they told you it was from an LLM, would it stop being funny.

No. It wouldn't. But they can't generate funny jokes.

I would be really happy if I am wrong, and it is possible to laugh all day by reading endless jokes from an LLM...

Post reply on HN