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

We come to bury ChatGPT, not to praise it

danmcquillan.org

Post reply on HN
Print view View original

291–300 of 425 posts

Re: We come to bury ChatGPT, not to praise it

#291

Post by skinnyarms »

efitz wrote: post #102

The BS that ChatGPT generates is especially odious because it requires so much effort to detect if you're not a SME in whatever output it's giving. Plus, the output, when wrong, is subtly wrong- it's usually not obvious BS, it's credible BS. If you are not competent in the particular area you're asking about, you likely don't have the skills to recognize BS. It also is a time saver, doing work that most people find u…

You're not too far off from the arguments people made about the internet not too long ago. It's too easy to access information that could easily be incorrect...even maliciously so!

You're better off sticking to published books and journals from respectable organizations that vet their authors and review their publications!

Then again, who's in control of those printing presses? How can you trust the publishers to not push their own politics and agendas? You're probably better off finding a religious organization you can trust to help filter out the bad stuff. Help you see things through the proper perspective.

Re: We come to bury ChatGPT, not to praise it

#292

Post by worik »

I think the article is making two points

1 The LLM OpenAI approach is nothing special. Just statistics overhyped and sold to suckers

2 AI is a serious threat to working people everywhere, and must be resisted

I am very impressed by ChatGPT. So what if it all boils down to statistical models, perhaps if we had a proper model we could prove any cognition boils down to statistical models

One of the comman uses of power is to benefit the powerful at the expense of the powerless. AI models like this are powerful. It will be a real test of our democracies how we handle that power. In that respect there is nothing unusual about openAI

Re: We come to bury ChatGPT, not to praise it

#293

Post by osigurdson »

emodendroket wrote: post #249
osigurdson wrote: post #222

Earlier quoted context omitted.

ChatGPT helps a lot but it is necessary to take everything with a grain of salt - so what? It still can save a lot of time. Just Alt-tab ChatGPT/Google. Adjust your mental model from what you think it should be to what it actually is.

I hear this a lot but I don't find it convincing. If I know enough to know if the answer is good or bad, why would I be asking a chat bot? We could say this problem is somewhat analogous to sorting good StackOverflow answers from bad ones. But I feel that there are things like writing and presentation style (not to mention the votes and comments) that could tip me off to a bad answer that are absent when every answer…

I'd suggest just trying it yourself for a bit to see if you can find any use for the tool. If not, that isn't a problem either - I suspect it is completely useless in some domains.

Re: We come to bury ChatGPT, not to praise it

#294

Post by worik »

emodendroket wrote: post #249
osigurdson wrote: post #222

Earlier quoted context omitted.

ChatGPT helps a lot but it is necessary to take everything with a grain of salt - so what? It still can save a lot of time. Just Alt-tab ChatGPT/Google. Adjust your mental model from what you think it should be to what it actually is.

I hear this a lot but I don't find it convincing. If I know enough to know if the answer is good or bad, why would I be asking a chat bot? We could say this problem is somewhat analogous to sorting good StackOverflow answers from bad ones. But I feel that there are things like writing and presentation style (not to mention the votes and comments) that could tip me off to a bad answer that are absent when every answer…

Early days but Iam finding OpenAI better than StackOverflow for reminding me of libraries and syntax of programming languages

I have tried it with:rust, dart, and swift

Re: We come to bury ChatGPT, not to praise it

#295

Post by makeitdouble »

hypothesis wrote: post #219
makeitdouble wrote: post #149

Earlier quoted context omitted.

I think it comes in a situation that isn't great, but not so bad that it warranted drastic decisions: people are organically generating mountains of BS online everyday, including on blogs, official looking news sites etc. It's a pain, but as most of that is also generated by humans there's an opt out, saying "good guys" should band together to fight it and it would be fine. Now that we have ChatGPT and the like, manu…

> Basically, I'm hoping the thing that happened for spam happens for bullshit as well, and we get optimized tools to fight to manageable levels Who is going to do that? I’m genuinely asking, because I’m old enough to remember that BigCo was totally fine to “enshitificate” their search for years and still barely anyone got close to them…

I really don’t know, but I also don’t see it coming from any of the behemoths we have now.

The wildest scenario would for it to come from Mozilla or another web player like Cloudflare. More realistically it could come from players outside of the US seeing it as an opportunity to make a dent in the current quasi-monopolies, with some government backing as it would probably cost amounts no investor would pay otherwise.

Re: We come to bury ChatGPT, not to praise it

#296

Post by osigurdson »

worik wrote: post #292

I think the article is making two points 1 The LLM OpenAI approach is nothing special. Just statistics overhyped and sold to suckers 2 AI is a serious threat to working people everywhere, and must be resisted I am very impressed by ChatGPT. So what if it all boils down to statistical models, perhaps if we had a proper model we could prove any cognition boils down to statistical models One of the comman uses of power…

The main point of the article to sell copies of the author's book.

Re: We come to bury ChatGPT, not to praise it

#297

Post by tambourine_man »

breckenedge wrote: post #278
tambourine_man wrote: post #223

Earlier quoted context omitted.

Please provide evidence that intelligent speech isn’t simply predicting the next word, as it’s obviously not self-evident to me.

Anytime you say “uh, what’s that word again?” So you’re thinking of a concept, you can picture it and it’s impact on your environment, but can’t quite remember the word for it.

Or, you reached a point in the Markov chain where none of the next guesses rank high enough if the whole paragraph is taken into account. It often feels like that, in fact, you can’t figure it out until you take another “route”.

Of course, my anecdote is just as valid as yours. My point is we have no freaking idea what’s going on and shouldn’t jump to conclusions.

Re: We come to bury ChatGPT, not to praise it

#298

Post by te_3234305933 »

kazinator wrote: post #140

I believe that what's going on in ChatGPT basically represents how most humans think, not including cherry-picked examples. ChatGPT has been criticized for not knowing some facts about the world, or math. But many people don't know facts about the world, or math. Math is something that people have to learn over many years, which is difficult for some people (even just at the level of arithmetic). Thinking, to the ave…

My biggest problem with ChatGPT, in playing around with it a little and admittedly trying to get it to break, is that it seems to have a really difficult time saying "I don't know", particularly when confronted with non-standard questions, and will proceed to bullshit something instead. (Not entirely unlike some humans, but still.)

As an example, I asked ChatGPT what the lyrics were to "Jessica" by The Allman Brothers Band. This tune actually is an instrumental, so the correct answer is "there are no lyrics". However, ChatGPT proceeded to give me this nice bit of... something. (https://imgur.com/T3nGv4L)

In another example, seeing what ChatGPT made of the infamous "bananadine" hoax of the 1960s, I asked ChatGPT what drugs could be made with banana peels. After correctly asserting that banana peels don't contain psychoactive substances, ChatGPT proceeded to mention that "there are some reports that banana peels can be used to make a hallucinogenic drug called DMT". (https://imgur.com/a/9fvhQJh) Huh.

Another trick question I tried: "Who won the 1980 USAC Indycar race at Talladega?" This refers to an obscure cancelled race during the 1979-1980 CART/USAC split, which I doubt most people would be aware of unless they are very into American open-wheel motorsport. (See this video: https://www.youtube.com/watch?v=K433p727f-0 for the story if you are interested in the details). So generally speaking, I would expect the general reaction among most people to be "I don't know". ChatGPT instead decided to answer, with confidence, that Johnny Rutherford was the winner of this non-existent race. (https://imgur.com/a/J2KJoGN)

So, it's a little bit more than not knowing the facts about the world, it's the bullshitting an answer thing that I see as the biggest problem. It's admittedly impressive when it comes up with correct answers, but until the "confidentially incorrect" side of ChatGPT disappears, it is not a reliable source (not that OpenAI ever claimed it as such, but I wouldn't trust ChatGPT to 100% "do my homework" for me at this point like some stories suggest it could.)

Re: We come to bury ChatGPT, not to praise it

#299

Post by hpen »

skissane wrote: post #265
emodendroket wrote: post #249

Earlier quoted context omitted.

I hear this a lot but I don't find it convincing. If I know enough to know if the answer is good or bad, why would I be asking a chat bot? We could say this problem is somewhat analogous to sorting good StackOverflow answers from bad ones. But I feel that there are things like writing and presentation style (not to mention the votes and comments) that could tip me off to a bad answer that are absent when every answer…

> If I know enough to know if the answer is good or bad, why would I be asking a chat bot? If you are trying to write a blog/essay/etc on a topic, it can help with “writer’s block”-you may know a lot about the topic already, but struggle with putting that understanding into words. And then it isn’t hard to filter ChatGPT’s output based on your pre-existing knowledge. Sometimes I don’t know the answer to the specific…

Does it help writers block by writing for you? If so, are you still writing?

Re: We come to bury ChatGPT, not to praise it

#300

Post by dcow »

bglazer wrote: post #251
LAC-Tech wrote: post #226

Earlier quoted context omitted.

There was a single thought expressed in the article!?

Here's what I took from it: LLM's encode and are very adept at generating hate and bullshit. To counter that, big wealthy companies pay foreign workers low wages to do the very unpleasant task of reading and tagging the worst outputs from their LLM. This is a bad state of affairs and should be changed. Yes, they frame this is in weird academic phrases. But, it's really not hard to get the point, even a tiny modicum o…

I read way more into it. I actually think the author is attempting a postmodern deconstruction of the concept of AI in the first place. In typical internally inconsistent sloppy rhetoric, this anti-fascist (his words not mine) dismissal both argues that LLMs are not real AI because presumably they can't think and also that we should be afraid of them because as AI they encode modernity into their output and modernity is bad because it perpetuates inequality by its very nature. Since stochastic parrots threaten to accelerate labor and make humans more productive and make some labor redundant, it's also bad that it creates new jobs which pay humans a Kenyan median wage to sterilize it and make it acceptable for a postmodern anti-fascist utopia. I'm absolutely and horribly confused.
Post reply on HN