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

971–980 of 1001 posts

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

#971

Post by mollerhoj »

janalsncm wrote: post #926
mollerhoj wrote: post #925

Sabine is an obvious grifter. Crazy to see her on the frontpage of HN

It’s weird that LLM bulls not not directly engage with the criticism, which would characterize as a fairly standard criticism of hallucination. Hallucinations are a problem. I don’t know who Sabine is but I don’t have to. Same with the top reply from Teortaxes containing zero relevant information, which Twitter in its infinite wisdom has decided is the “most relevant” reply. (The second “most relevant” reply is some…

It's weird that you'd guess (wrong) that I'm an "LLM bull", then start generalizing the actions of this supposed group. Hallucinations are a problem, no news value there. Sabine seems not to be able to get value out of LLMs, and concludes that they aren't useful. The logic is not exactly impressive.

It's a tool. It can be useful, it doesn't always work. Some people claim it's better than it is, some people claim it's worse. This isn't exactly rocket science.

Regardless of the tweet in question, Sabine is a grifter. Her novel takes on academia being some kind of conspiracy of people milking the system, and of physicists not being interested in making new discoveries are nonsensical and only serves to increase her own profile. Look at this video of her trying to convince the world she received an email that apparently proves all her points correct. My BS detector tells me she wrote that email herself, but you be the judge: https://www.youtube.com/watch?v=shFUDPqVmTg

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

#972

Post by simoncion »

retrac wrote: post #866
simoncion wrote: post #535

Earlier quoted context omitted.

> ...there was no such thing as live machine transcription ten years ago. What? Then what the hell do you call Dragon NaturallySpeaking and other similar software in that niche?

Thru the 90s and 00s and well into the 10s I generally dismissed speech recognition as useless to me, personally. I have a minor speech impediment because of the hearing loss. They never worked for me very well. I don't speak like a standard American - I have a regional accent and I have a speech impediment. Modern speech recognition doesn't seem to have a problem with that anymore. IBM's ViaVoice from 1997 in partic…

So, you started by saying

> ...there was no such thing as live machine transcription ten years ago.

Now you're saying that live machine transcription existed thirty years ago, but it has gotten substantially better in the intervening decades.

I agree with your amended claim.

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

#974

Post by hbogert »

joquarky wrote: post #957
glimshe wrote: post #91

Earlier quoted context omitted.

I'm tired of people bashing LLMs. AI is so useful in my daily work that I can't understand where these people are coming from. Well, whatever... As you said, examples where I wouldn't expect LLMs to be good at from people who dismiss the scenarios where LLMs are great at. I don't want to convince anyone, to be honest - I just want to say they are incredibly useful for me and a huge time saver. If people don't want to…

I'm starting to resign myself to just enjoying the benefits and let those who can't evolve fall behind. I'm growing weary of trying to help people use these tools properly.

I'm pretty weary of all the people telling me I'm "holding it wrong"

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

#975

Post by glimshe »

hbogert wrote: post #974
joquarky wrote: post #957

Earlier quoted context omitted.

I'm starting to resign myself to just enjoying the benefits and let those who can't evolve fall behind. I'm growing weary of trying to help people use these tools properly.

I'm pretty weary of all the people telling me I'm "holding it wrong"

I'll give you a simple and silly example which could give you additional ideas. LLMs can be great for checking whether people can understand something.

One day I came up with a joke and wondered whether people would "get it". I told the joke to ChatGPT and asked it to explain it back to me. ChatGPT did a great job and nailed what's supposedly funny about the joke. I used it in an email so I have no idea whether anyone found it funny, but at least I know it wasn't too obscure. If an AI can understand a joke, there's a good chance people will understand it too.

This might not be super useful but demonstrates that LLMs aren't only about generating text for copy-and-paste or retrieving information. It's "someone" you can bounce ideas with, ask opinions and that's how I use it most frequently.

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

#976

Post by alxjrvs »

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…

Because the marketers oversold it. That is why you are seeing a pushback. I also outright rejected them because 1) they were sold and marketed as end all be all replacements for human thought, and 2) they promised to replace only the parts of my job that I enjoy. Billboards were up in San Francisco telling my "bosses" that I was soon to be replaced, and the loudest and earliest voices told me that the craft I love is dead. Imagine Nascar drivers excitedly discussing how cool it was they wouldn't have to turn left anymore - made me wonder why everyone else was here.

It was, more or less, the same narrative arc as Bitcoin, and was (is) headed for a crash.

That said, I've spent a few weeks with augment, and it is revelatory, certainly. All the marketing - aimed at a suite I have no interest in - managed to convince me it was something it wasn't. It isn't a replacement, any more than a power drill is a replacement for a carpenter.

What it is, is very helpful. "The world's most fully functioning scaffolding script", an upgrade from copilot's "the world's most fully functioning tab-completer". I appreciate it usefulness as a force multiplier, but I am already finding corners and places where I'd just prefer to do it myself. And this is before we get into the craft of it all - I am not excited by the pitch "worse code, faster", but the utility is undeniable in this capitalistic hell planet, and I'm not a huge fan of writing SQL queries anyway, so here we are!

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

#977

Post by simonw »

xigoi wrote: post #906
KoolKat23 wrote: post #749

Earlier quoted context omitted.

This is more a lack of understanding of it's limitations, it'd be different if they asked for it to write a python script to collate the data.

If the LLM is intelligent, why can’t it figure out that writing a script would be the best way to solve the problem?

ChatGPT has been able to do exactly that (using its Code Interpreter tool) for two years now. Gemini and Claude have similar features.

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

#978

Post by simonw »

ulech wrote: post #969

> "By my personal estimate currently GPT 4o DeepResearch is the best one." uhm, I dismiss this statement here? if you call 4o the best, that means you haven't genuinely explored other models before making such claims...

GPT 4o Deep Research is currently the only available interface to the full o3 (as opposed to o3 mini).

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

#979

Post by whamlastxmas »

blactuary wrote: post #808
whamlastxmas wrote: post #746

Earlier quoted context omitted.

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.

The search ecosystem is broken now because google is focused on LLMs

It’s broken bc google has spent 20+ years promoting garbage content in a self-serving way. No one was able to compete unless they played by googles rules, and so all we have left is blog spam and regular spam

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

#980

Post by whamlastxmas »

player1234 wrote: post #798
whamlastxmas wrote: post #746

Earlier quoted context omitted.

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.

How did you measure and compare googling/stack overflow to coding with an LLM? How did you get to the very impressive number ten billion times better?! Can you share your methodology? How have you defined better?

I take calipers to my boss’s forehead veins and see how pissed he is routinely throughout the day
Post reply on HN