Hacker News Reader

Comment by vzaliva | Hacker News Reader

This is a good example where a summary produced by much-hated AI (GPT 4o) is quite useful (to people who do not want to read all details in he article):

"In his study, James Pae examined the hypothesis: "The closer to the train station, the worse the kebab." Focusing on Paris, he analysed kebab shops' proximity to train and metro stations alongside their Google review ratings. His findings revealed a negligible correlation between a kebab shop's distance from stations and its review ratings, suggesting the hypothesis lacks substantial evidence. Pae acknowledges potential influencing factors such as tourism and review biases and expresses intentions to revisit the study for further analysis."

View on HN

I don't think a lot of people have a problem with AI summaries (a lot of people are using AI for exactly that). I think the 'hate' mostly comes from the fact that people tend to copy/paste whatever AI says without adding anything to the conversation.

It's the same as me running a query on Google and copy/pasting a list of 10 results. It doesn't really add anything to a conversation - anyone can go to Google and look something up.

For what it's worth, I do hate people pasting their AI summaries to the comments. Not only are they adding nothing, they are actively detracting from the conversation; they have just pasted a wall of text without fact-checking it. And in fact, this "summary" misrepresents the article; it completely ignores the humor and presents it as a serious scientific endeavour.

But judging from the rest of the comments, it seems like most people barely managed to finish reading the title, so perhaps there's no need to worry about them reading this AI slop...

I read the article and fact-checked the summary before posting. The original article is quite long, so the summary may be useful for those who are intrigued by the title and just want to know the outcome.

I don’t see any need to mention the "humor" aspect here. Many seemingly laughable hypotheses have turned out to be true when rigorously tested. The author did a good job investigating this one.

Asking a question to a coworker and having him copy paste from chatgpt has to be the worst thing about being a programmer in 2025.

If I'm asking you something it's because I want to discuss it, not because I couldn't be bothered to google it myself. Do people really not realize that?

Reply on news.ycombinator.com