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

I genuinely don't understand why some people are still bullish about LLMs

twitter.com

Post reply on HN
Print view View original

781–790 of 1001 posts

Re: I genuinely don't understand why some people are still bullish about LLMs

#781

Post by knowaveragejoe »

IncreasePosts wrote: post #340
ANighRaisin wrote: post #155

Earlier quoted context omitted.

This isn't really a problem in tool-assisted LLMs. Use google AI studio with search grounding. Provides correct links and citations every time. Other companies have similar search modes, but you have to enable those settings if you want good results.

Okay, but it's weird there is a "don't lie to me" button.

How would that ever work? The only thing you can do is continue to refine high quality data sets to train on. The rate of hallucination only trends downwards on the high end models as they improve in various ways.

Re: I genuinely don't understand why some people are still bullish about LLMs

#782

Post by quonn »

whamlastxmas wrote: post #746
WhyOhWhyQ wrote: post #646

Earlier quoted context omitted.

The technology is not just less than superintelligence, for many applications it is less than prior forms of intelligence like traditional search and Stack Exchange, which were easily accessible 3 years ago and are in the process of being displaced by LLMs. I find that outcome unimpressive. And this Tweeter's complaints do not sound like a demand for superintelligence. They sound like a demand for something far more…

I’m sorry but the experience of coding with an LLM is about ten billion times better than googling and stack overflowing every single problem I come across. I’ve stack overflowed maybe like two things in the past half year and I’m so glad to not have to routinely use what is now a very broken search engine and web ecosystem.

It‘s broken now. It was fine 5 years ago.

Re: I genuinely don't understand why some people are still bullish about LLMs

#783

Post by casey2 »

01100011 wrote: post #123
meowface wrote: post #110

Earlier quoted context omitted.

A lot of software developers have an initial bad experience and assume it's terrible and give up on it. I feel bad for people who haven't yet experienced how useful these models are for programming. Some also just prefer manually entering everything. Those people I will never understand.

how much time do I need to devote to see anything but garbage? For reference, I program systems code in C/C++ in a large, proprietary codebase. My experiences with OpenAI(a year ago or more), and more recently, Cursor, Grok-v3 and Deepseek-r1, were all failures. The later two started out OK and got worse over time. What I haven't done is asked "AI" to whip up a more standard application. I have some ideas(an ncurses…

If your mind has trouble expanding outside the domain of "use this well known tool to do something that has already been done" then no amount of improvements will free you from your belief that chatbots are glorified autocomplete.

Re: I genuinely don't understand why some people are still bullish about LLMs

#784

Post by nancyminusone »

brookst wrote: post #684
Thlom wrote: post #669

Earlier quoted context omitted.

A municipality in Norway used LLM to create a report about the school structure in the municipality (how many schools are there, how many should there be, where should they be, how big should they be, pros and cons of different size schools and classes etc etc). Turns out the LLM invented scientific papers to use as references and the whole report is complete and utter garbage based on hallucinations.

And that says… what? The entire LLM technology is worthless for all applications, from all implementations? A company I worked for spent millions on a customer service solution that never worked. I wouldn’t say that contracted software is useless.

No, it says that people dislike liars. If you are known for making up things constantly, you might have a harder time gaining trust, even if you're right this time.

Re: I genuinely don't understand why some people are still bullish about LLMs

#785

Post by rramadass »

I think the author is being overly pessimistic with this. The positives of an LLM agent outweigh the negatives when used with a Human-in-the-loop.

For people interested in understanding the possibilities of LLM for use in a specific domain see The AI Revolution in Medicine: GPT-4 and Beyond by Peter Lee (Microsoft Research VP), Isaac Kohane (Harvard Biomedical Informatics MD) et al. It is an easy read showing the authors systematic experiments with using the OpenAI models via the ChatGPT interface for the medical/healthcare domain.

For a current-status follow-up to the above book, here is Peter Lee's podcast series The AI Revolution in Medicine, Revisited - https://www.microsoft.com/en-us/research/story/the-ai-revolu...

Instead of reading trivial blogs/tweets etc. which are useless, read the above to get a much better idea of an LLM's actual capabilities.

Re: I genuinely don't understand why some people are still bullish about LLMs

#786

Post by dcchambers »

croes wrote: post #237
dcchambers wrote: post #219

Earlier quoted context omitted.

> And people just sit around, unimpressed, and complain that ... what ... it isn't a perfect superintelligence that understands everything perfectly? IMO there are two distinct reasons for this: 1. You've got the Sam Altman's of the world claiming that LLMs are or nearly are AGI and that ASI is right around the corner. It's obvious this isn't true even if LLMs are still incredibly powerful and useful. But Sam doing t…

> LLMs are an existential threat to basically every knowledge worker job on the planet. Given the typical problems of LLMs they are not. You still need them to check the results. It’s like FSD, impressive when it works, bad if not, scary because you never known beforehand when it’s failing

I used to be firmly in this camp, but I'm not any more.

Even if they fail 1% of the time, the cost savings are too great. Businesses will take the risk.

Re: I genuinely don't understand why some people are still bullish about LLMs

#787

Post by rcdwealth »

snitty wrote: post #286
gilbetron wrote: post #109

I get so confused on this. I play around, test, and mess with LLMs all the time and they are miraculous. Just amazing, doing things we dreamed about for decades. I mean, I can ask for obscure things with subtle nuance where I misspell words and mess up my question and it figures it out. It talks to me like a person. It generates really cool images. It helps me write code. And just tons of other stuff that astounds me…

Folks really over index when an LLM is very good for their use case. And most of the folks here are coders, at which they're already good and getting better. For some tasks they're still next to useless, and people who do those tasks understandably don't get the hype. Tell a lab biologist or chemist to use an LLM to help them with their work and they'll get very little useful out of it. Ask an attorney to use it and…

[dead]

Re: I genuinely don't understand why some people are still bullish about LLMs

#788

Post by myko »

dartharva wrote: post #710
jchw wrote: post #153

Earlier quoted context omitted.

Look man, and I'm saying this not to you but to everyone who is in this boat; you've got to understand that after a while, the novelty wears off. We get it. It's miraculous that some gigabytes of matrices can possibly interpret and generate text, images, and sound. It's fascinating, it really is. Sometimes, it's borderline terrifying. But, if you spend too much time fawning over how impressive these things are, you m…

> Well, are they useful? ... Yeah, of course LLMs are useful, but we need to remain somewhat grounded in reality. How useful are LLMs? They are useful enough that they can passably replace (much more expensive) humans in a lot of noncritical jobs, thus being a tangible tool for securing enterprise bottom lines.

Which jobs? I haven't seen LLMs successfully replace more expensive humans in noncritical roles

Re: I genuinely don't understand why some people are still bullish about LLMs

#789

Post by vonneumannstan »

WhyOhWhyQ wrote: post #646
gilbetron wrote: post #109

I get so confused on this. I play around, test, and mess with LLMs all the time and they are miraculous. Just amazing, doing things we dreamed about for decades. I mean, I can ask for obscure things with subtle nuance where I misspell words and mess up my question and it figures it out. It talks to me like a person. It generates really cool images. It helps me write code. And just tons of other stuff that astounds me…

The technology is not just less than superintelligence, for many applications it is less than prior forms of intelligence like traditional search and Stack Exchange, which were easily accessible 3 years ago and are in the process of being displaced by LLMs. I find that outcome unimpressive. And this Tweeter's complaints do not sound like a demand for superintelligence. They sound like a demand for something far more…

[flagged]

Re: I genuinely don't understand why some people are still bullish about LLMs

#790

Post by ubermonkey »

It's a fair point.

An LLM can do some pretty interesting things, but the actual applicability is narrow. It seems to me that you have to know a fair amount about what you're asking it to do.

For example, last week I dusted off my very rusty coding skills to whip up a quick and dirty Python utility to automate something I'd done by hand a few too many times.

My first draft of the script worked, but was ugly and lacked any trace of good programming practices; it was basically a dumb batch file, but in Python. Because it worked part of me didn't care.

I knew what I should have done -- decompose it into a few generic functions; drive it from an intelligent data structure; etc -- but I don't code all the time anymore, and I never coded much in Python, so I lack the grasp of Python syntax and conventions to refactor it well ON MY OWN. Stumbling through with online references was intellectually interesting, but I also have a whole job to do and lack the time to devote to that. And as I said, it worked as it was.

But I couldn't let it go, and then had the idea "hey, what if I ask ChatGPT to refactor this for me?" It was very short (Here's where the story got interesting. YES, the first pass of its refactor was better, but in order to get it to where I wanted it, I had to coach the LLM. It took a couple passes through before it had made the changes I wanted while still retaining all the logic I had in it, and I had to explicitly tell it "hey, wouldn't it be better to use a data structure here?" or "you lost this feature; please re-add it" and whatnot.

In the end, I got the script refactored the way I wanted it, but in order to get there I had to understand exactly what I wanted in the first place. A person trying to do the same thing without that understanding wouldn't magically get a well-built Python script.

Post reply on HN