Hacker News Reader

I’m not sure why we need comments from people proud that they couldn’t be troubled to read the article that is supposedly being discussed, but it seems like a popular genre of reply whenever something tech-critical is posted.

I tried to read the article too and found it both visually unpleasant and intellectually unfocused. GP is making the rather comically clever point that ChatGPT is useful to summarize text that might be difficult for humans to parse, even a text claiming that ChatGPT is a useless model only capable of spewing Islamophobic speech and in need of burying. I mean the article ends on some tangent about reactionary solutionism... If you don't see the satire in that, well, you might be taking life too seriously.

>both visually unpleasant

Glad I wasn't the only one who noticed that. It's like the author picked a font/CSS combo to simulate a book with horrible typesetting.

Does it look like a blurred screenshot of text to anyone else?

Love the book, and LaTeX, but it does give me these vibes https://www.google.com/books/edition/Walk_Through_Combinator....

I tried to read the article too and found it both visually unpleasant and intellectually unfocused.

That's fine but then you're probably not in a position to have curious conversation about it. Which is also fine! But the solution to that is to find something else that's interesting to you, rather than to gunk up the thread for the people who are interested in discussing the piece.

Despite the qualities I ascribed to it, I actually read the piece. That's why I chuckled at the satire of having ChatGPT summarize it. I thought it sufficiently concise and every so slightly poignant commentary, daresay curious. I haven't disparaged anybody else for their own curious conversation, have I? If you'd like to discuss whether, "The structural injustices and supremacist perspectives layered into AI put it firmly on the path of eugenicist solutions to social problems.", well, be my guest.

I didn't say you've disparaged anyone, you're defending the paste of generated content into the thread as if it's a good thing because, I dunno, you found it amusing or it tickled your sneerbone. But pasting generated content into the thread is not a good thing by HN's standards of good things for the reasons I described and the reasons you described are fine reasons but they are bad HN reasons.

I don't care where the content came from. I judged it on its merit not on its origin. If it was a bad summary or more difficult to understand than the article then it wouldn't have worked quite as well. The additional context of it being generated by the [harmful technology in question] made it meta humorous, yes, not least because I found it to be a better representation of the argument that I think the author is attempting to communicate than the form in which it was originally presented. Ultimately, the author's piece sounds more like a stochastic parrot to me than the summary by ChatGPT.

Hence, I find satire an entirely appropriate tactic to spark curious discussion, in this case. It's able to communicate all that remarkably succinctly. And, though we have not discussed whether ChatGPT is harmful because OpenAI paid Kenyan works $2/hr to "clean up" the model, I do believe we have discussed the nature of the argument in the first place, which is whether a LLM can produce anything of value. It seems preposterous to claim that ChatGPT is utterly useless and harmful when a harmless use for it is shoved in your face, no? And since I laughed, I guess I value the output, oddly. Which all quite contradicts the author's assertion that bad evil technology has no value and is destroying humanity.

Summarising the article was useful for me, as user of HN since it was Startup News. It certainly doesn’t violate the HN Guidelines.

While it may be frustrating to see comments from people who haven't read the article being discussed, it's important to recognize that everyone consumes and processes information in different ways. Some people may prefer to skim articles or jump straight to the comments section to get a sense of the wider conversation, rather than reading every detail of the article itself.

Some people may not have the time or the interest to read lengthy articles and may prefer to get the gist of the discussion through comments. It's important to keep an open mind and respect the ways in which others consume and engage with information.

—thus spake chatGPT

It's tech-person-splaining

AI is still not AI, still cannot pass the 'critical thinking' task.

I can prove this without having ever interacted with ChatGPT.

The problem with "A.I." is that it continues to only be able to take the 'almost best' sum of ideology and not invent new ideas. AI cannot innovate. It can only imitate.

What about the people who actually read the original article word for word - aren't they the real victims?

I don’t think so. I don’t agree with everything it says but I found it interesting enough. “Is generating a lot of nonsense and boilerplate a justifiable use of scarce resources?” is actually a rather profound question.

The article does not contribute anything new.

At this point if I see any variation of "bullshit generator" used to downplay the significance of LLMs I will instantly discount the author as a person who has already lost the plot.

Just smile politely and leave them to their devices as the world changes around them.

This is not even a correct summary of the content of the post. I'm genuinely entertained by the sheer amount of shamelessness demonstrated here. Seriously, though, technology at wrong hands endorses bad behaviors.

I think this is where the OP says "QED".

I went and wasted my time reading the post and it turns out the ChatGPT summary was accurate and this comment that it is not accurate was grossly wrong. Wish I hadn't wasted my time.

ChatGPT is not really meant for summarizing things, Kagi Summarizer for example:

"This article discusses the dangers of ChatGPT, a large language model, and how it is used to generate 'bullshit' and propagate existing power structures. It argues that the model is harmful and that its plausibility makes it even more dangerous. The article also highlights the exploitative labour practices that go into maintaining the model, as well as the underlying supremacist perspectives that are embedded in AI. The author suggests that instead of embracing AI, we should focus on centring activities of care and search for alternatives to algorithmic immiseration."

Why is it incorrect? You haven’t provided any supporting arguments.

Here is Kagi’s summary:

This article discusses the dangers of ChatGPT, a large language model, and how it is used to generate 'bullshit' and propagate existing power structures. It argues that the model is harmful and that its plausibility makes it even more dangerous. The article also highlights the exploitative labour practices that go into maintaining the model, as well as the underlying supremacist perspectives that are embedded in AI. The author suggests that instead of embracing AI, we should focus on centring activities of care and search for alternatives to algorithmic immiseration.

Both miss that it’s an advertorial for the author’s book.

From the HN guidelines on comments: “Be kind. Don't be snarky. Converse curiously“

I downvoted you because your comment is both snarky and profoundly incurious.

Seriously, either take the time to engage with whats actually written in the post or don’t bother posting lazy swipes. As has been noted by others, the ChatGPT summary isn’t even really correct.

I’ll admit the piece is written in jargony social science language, but it makes real, interesting points about the social changes that are going to accompany introduction of LLM’s into society.

If authors want people to read their articles they should learn to write.

Let's put snarkiness to one side - the article was shit. You know it, I know it, and everyone else knows it. Poorly written, with the same dumb thoughts - and I use the term loosely - I can get from any spoiled sophmore at any university in the western world.

The only reason it got any traction was a clickbait title about a hot topic. So I feel perfectly OK about shitting all over it. If you don't, flag me. It's the internet, it really doesn't matter.

You still haven’t referenced a single thought actually expressed in the article.

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 of intellectual curiosity would get you there.

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.

Interesting. I think a more coherent form of the argument is this: LLM's aren't capable of reasoning or self-updating. Capitalists are exploiting powerless labor to encode their preferences and goals into LLM's. Because the LLM can't really think and isn't human they can control it. So, we're going to be forced to deal with robots that perfectly embody the worst aspects of the modern economy and society.

It's hard because the author is just generating poorly disguised backlinks to other peoples' thoughts. Follow some of the links and see for yourself e.g. Stop feeding the hype and start resisting (which I also read and lacks any explanation of exactly how one might resist if one was so inclined) which links to the same CogSci2022 talk as the author does on resisting the dehumanization of technology. All three of which cite the resistance slogan Stochastic Parrots from a paper published in the proceedings of the Fairness Accountability and Transparency conference. Sociologists don't like large language models because they have the potential to make humans do less menial labor [which isn't good for a capitalist society because it frees humans up to think]. I mean I have no idea why these people feel so threatened by a large language model, but outrage porn gets clicks.

The bias and motive is also obnoxious. The author is very clearly generating backlinks to other pseudo-academics in some postmodern anti "AI" résistance. It's just regurgitation of other peoples' original commentary.

And what makes you think this is a correct summary

You won my internet for the day.

Thanks LAC-Tech!

It runs against the purpose of HN, though. Take a look at:

https://news.ycombinator.com/item?id=34662167

https://news.ycombinator.com/item?id=34240213

https://news.ycombinator.com/item?id=34178237

The parent isn’t a bot account. The summarisation was added by a human and adds to the discussion.

This is directly addressed by the moderator comments.

I found the article so devoid of insight, just a regurgitation of clichés from the left, that I was left wondering if it was written by ChatGPT.

My impression is that dumb people (for lack of a better term) find ChatGPT useful, while smart people (again, for lack of a better term) don't.

It's kind of like how a jazz musician doesn't like for example Ed Sheeran, but the average person does.