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

ChatGPT Can't Kill Anything Worth Preserving

biblioracle.substack.com

Post reply on HN
Print view View original

61–70 of 363 posts

Re: ChatGPT Can't Kill Anything Worth Preserving

#61

Post by gtirloni »

101008 wrote: post #22

I agree with the article but maybe you are trying to do the wrong thing with ChatGPT. I am not a native English speaker (as many of you already noticed due to my grammar mistakes), so ChatGPT have been very useful to rewrite my texts, fix grammar and maybe "beautify" them a bit. It's has been an invaluable tool in that sense. I feel much more confident now sending emails and writing important messages knowing that th…

The second version is only marginally better and ignores your own style. How's that invaluable?

Re: ChatGPT Can't Kill Anything Worth Preserving

#62

Post by Semaphor »

heavyset_go wrote: post #41
Mechanical9 wrote: post #35

Earlier quoted context omitted.

Huh, that hasn't been my experience, at least not yet. Are you talking about newer movies and shows?

Mostly videos I stream online. YouTube is a good example, but I've noticed an uptick in bad subtitles on other streaming platforms that stream content with higher production values.

English dubbed English subs are completely different, the subs (presumably) being translations of the original language instead of the dubs. It's very jarring and I can only imagine this getting better with AI.

Re: ChatGPT Can't Kill Anything Worth Preserving

#63

Post by ryanklee »

ARandomerDude wrote: post #18
ryanklee wrote: post #12

Earlier quoted context omitted.

Hmmm, I suppose I'm interested in what you mean by bullshit. Just because AI makes something easier to do, doesn't mean the work is bullshit. Lots of research, composition, and communication work becomes much easier to do with an AI assist, but that doesn't mean the work is bullshit.

> I'm coming around to the idea that in general the more bullshit something is the easier it is for AI to disrupt. > in general

Sorry, but I don't see how this is a clarifying response.

Research, composition, and communication are a big chunk of what ChatGPT is good at assisting. I don't know enough to claim that it /is/ the general case, but it's very likely not a minority.

And if that's the case, and these activities are not /essentially/ bullshit, then I'm still wondering how AI disruption can be claimed to be a reliable signal of bullshit.

Re: ChatGPT Can't Kill Anything Worth Preserving

#64

Post by dv_dt »

_petronius wrote: post #14

I'm less worried about stuff like ChatGPT killing things off, but more about just making everything noticeably a bit worse. To take an analogy: bad voice recognition software abounds everywhere, not because it is better than what it replaced in terms of UX, but because it works just enough and allows massive cost-savings on hiring people to do customer service jobs. A world where most marketing copy is written by med…

ChatGPT is basically automated reversion to a slightly worse mean for applicable areas. The algos as sophisticated as they are produce a credible remix of the training input.

Re: ChatGPT Can't Kill Anything Worth Preserving

#66

Post by visarga »

NotYourLawyer wrote: post #17

It’s gonna kill the internet as a place to find useful information. Google is already almost unusable for researching any topic that’s even adjacent to anything that can be monetized. Widespread, good-enough, AI-generated, SEO’d “articles” will push it over the edge.

I don't think so. The internet is already full of shit sites, you can automate spam very easily in any quantity.

But chatGPT would actually write useful texts, most of them more useful than the average blog post. Over a few years they might be just as good as the best human articles.

Normal people could write a draft version and have chatGPT reword it, that would control the contents of the article. But I expect spammers would just clone other articles, they don't have the time.

One thing AI could do is to scale up validation. Search every topic and note the answers - do they concur? do they disagree? what is the distribution. Maybe there is no answer?

This would be a reference for AI to stop hallucinating and making factoid mistakes. The model should know when it doesn't know, or when it is stepping on a landmine by forgetting to mention something.

In the end I think the internet is going to be full of spam, AI generated or not. And we will need to use more AI to extract the signal from the noise. This time it should be local AI under user control.

Re: ChatGPT Can't Kill Anything Worth Preserving

#67

Post by 082349872349872 »

alsodumb wrote: post #39
debacle wrote: post #34

I see two big problems with GPT: 1. It presents information really well that is completely false. I don't know exactly how it works under the hood, but sometimes it spits out paragraphs that read like the fever dreams of a flat earther. 2. It will cause our communication skills to further atrophy. Maybe in an innocuous way (no one memorizes epic poems anymore), but with GPT I think it will make people less likely to…

Regarding 1, it's well known that LLMs are pretty good at hallucinating.

Regarding 2, it's well known that wetware online are already pretty bad at passing a persuasiveness turing test.

(spellcheck means we no longer have to worry about getting distracted by orthography; if gpt means we no longer have to worry about getting distracted by grammar, I should hope we could then move on to helping people express actual ideas)

Re: ChatGPT Can't Kill Anything Worth Preserving

#68

Post by KevinGlass »

Mechanical9 wrote: post #35
heavyset_go wrote: post #20

Earlier quoted context omitted.

Subtitles got noticably worse now that their generation is often farmed out to AI, especially in translations. Drives me nuts as someone who likes using subtitles.

Huh, that hasn't been my experience, at least not yet. Are you talking about newer movies and shows?

Sometime last year, someone published a guide on auto-generating subtitles using an emacs plugin and WhisperAI. A lot of anime release groups use it so they can get an English subtitled version of a show released a couple hours after it airs in Japan.

It's nice to have access sooner, especially for less popular shows that might never get any kind of release outside of Japan, but it's a lot worse then most fan subs and sometimes misses entire sections of dialog.

Re: ChatGPT Can't Kill Anything Worth Preserving

#69

Post by dgb23 »

_petronius wrote: post #14

I'm less worried about stuff like ChatGPT killing things off, but more about just making everything noticeably a bit worse. To take an analogy: bad voice recognition software abounds everywhere, not because it is better than what it replaced in terms of UX, but because it works just enough and allows massive cost-savings on hiring people to do customer service jobs. A world where most marketing copy is written by med…

When everything will suck a bit more in the future, will it then recursively suck even more over time given that AI feeds itself content?
Post reply on HN