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

Ask HN: What if AI was used to verify users reading articles before discussion?

news.ycombinator.com

Post reply on HN
Print view View original

11–20 of 38 posts

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#11

Post by brudgers »

krapp wrote: post #7
jstx1 wrote: post #4

- The contents of the article are less important than you're implying, most people are here for the discussion. Yes, sometimes not reading the article is a problem but many other times it isn't and this can be clarified in the course of the discussion. Very often the article itself is just a starting point to discuss a broader topic. - People won't answer the questions like you suggest, they'll just leave. In a hypot…

When people don't read the article, they tend to do one of two things - comment on the title (which is often misleading) or use the title as a springboard for politically-driven polemic. Both drive down the quality of discussion considerably. The fewer people who read the article there are, the less substantive the discussion becomes, less insightful, less novel, more tedious and repetitive. The best comments are, in…

Because HN allows comments to wander...or jump on an airplane to a different continent...reading the article is not a prerequisite for good comments and discussion. The articles are mainly an excuse to type into little boxes.

The main thing that produces good discussion is good comments. The double converse is also the case. And a recognized problem: people read the article to find something negative to say...cherry picked quotations taken uncharitably is a proven method with the bonus of suggesting that the commentor actually read the article (but it is easier if something early in the article is a call to angry typing in outrage).

To put it another way, HN is able to have interesting conversations about lists, diagrams, and landing pages. Thin low information blog posts that are only worth reading for their headlines can spark deep informative comment pages.

To put it another way (again), Ask HN is a ready example of how little catalyst the community needs at times. The article is a spark and if it lands on the right intellectual tinder you get a great inferno.

[For polemic reading the article matters not. Polemitians bring their own tinder to the forest and find fuel in the comments no matter how good.]

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#13

Post by sircastor »

This reminds me of a handful of courses I had in school where we had to read an article, take a quiz and then participate in a “discussion”.

To me, the “you didn’t read the article” or “the article directly addresses this” responses are most effective dismissal of such a post, and a bit of a public embarrassment for the poster.

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#14

Post by paulcole »

> I recently read a frustrating thread on HN where it was obvious several people were making arguments against points that were already addressed in the article.

HN wants the engagement and doesn’t give a hoot if people read the articles before commenting.

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#15

Post by muzani »

eimrine wrote: post #3

Gpt-3 can not make questions from text. If would be better to give instant answers if it can read and realize the material, because sometimes I really do not need to read all the article.

There's a discussion on how to do it here: https://community.openai.com/t/ask-and-answer-questions-from...

This was with half an hour of work. I'm sure that with proper R&D, it can do better.

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#16

Post by muzani »

I don't think this would solve it. AI can make mistakes too and quite frequently. As we've learned from CAPTCHAs, they're very annoying and kill engagement.

I think the moderation and "the post answers this" works well enough. If there are lots of straw men, it could be a problem with the content or a clickbaity title.

In many cases, I don't care much about the article at all. Sometimes it's just "new update from Firefox" and people just want to talk about Firefox updates.

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#19

Post by tomjen3 »

krapp wrote: post #7
jstx1 wrote: post #4

- The contents of the article are less important than you're implying, most people are here for the discussion. Yes, sometimes not reading the article is a problem but many other times it isn't and this can be clarified in the course of the discussion. Very often the article itself is just a starting point to discuss a broader topic. - People won't answer the questions like you suggest, they'll just leave. In a hypot…

When people don't read the article, they tend to do one of two things - comment on the title (which is often misleading) or use the title as a springboard for politically-driven polemic. Both drive down the quality of discussion considerably. The fewer people who read the article there are, the less substantive the discussion becomes, less insightful, less novel, more tedious and repetitive. The best comments are, in…

Old angry man rant incoming.

>comment on the title (which is often misleading)

I grew up with the understanding that in a newspaper you should be able to read the title and get to know what happen and then be able to progressively read more if you wanted more details.

When the title is not accurate enough, or is click-bait, maybe we should start there with the AI there?

Re: Ask HN: What if AI was used to verify users reading articles before discussion?

#20

Post by jstx1 »

tomjen3 wrote: post #19
krapp wrote: post #7

Earlier quoted context omitted.

When people don't read the article, they tend to do one of two things - comment on the title (which is often misleading) or use the title as a springboard for politically-driven polemic. Both drive down the quality of discussion considerably. The fewer people who read the article there are, the less substantive the discussion becomes, less insightful, less novel, more tedious and repetitive. The best comments are, in…

Old angry man rant incoming. >comment on the title (which is often misleading) I grew up with the understanding that in a newspaper you should be able to read the title and get to know what happen and then be able to progressively read more if you wanted more details. When the title is not accurate enough, or is click-bait, maybe we should start there with the AI there?

Newspapers use (used?) an inverted pyramid structure because the bottom of the article can often be cut in editing so everything fits correctly on the page. As a newspaper journalist, this is out of your control so you need to make sure that your text makes sense even if one or more paragraphs are cut from the bottom.

I suppose it helps with attention and assimilating information too.

Post reply on HN