[anigbrowl]

I extracted a summary using a large language model.

Please do not. While that is an interesting educational tool that I like to use too, republishing it is just information pollution because it will create unproductive feedback loops. We always have abstracts to fulfill this need.

[dilippkumar]

Hard disagree. GP comment was extremely useful to me, and I want this usage to continue.

> republishing it is just information pollution because it will create unproductive feedback loops.

Internet forums have never been grounded in epistemic rigor. If feedback loops have to be avoided, that responsibility does not fall on everyday users of the internet.

Find better sources to train models.

[anigbrowl]

This is what the abstract is for, or you can submit material to an LLM on your own, or read any of many explainer articles by people with varying degrees of scientific knowledge and professional writing ability.

[bsder]

> Hard disagree. GP comment was extremely useful to me, and I want this usage to continue.

And why do you believe that comment to be correct in any way shape or form?

Does it actually reflect the article? Did it get something subtly wrong? Or maybe it's completely wrong?

You'll have to pull up the article to answer those questions anyway, so why add this noise?

[ttpphd]

Wow, good thing that I literally linked to the primary sources AND an additional secondary source for background.

Y'all need to see context.

[bsder]

The fact that you used a GPT summary means that your comment was a net negative to the discussion as I can't even trust that the articles you cited actually are the same articles as quoted in the original paper.

Congratulations on providing an information free post that is in every way inferior to the original article and manages to waste the time of anyone who reads it.

[ttpphd]

You are just being silly and argumentative now.

[AnIdiotOnTheNet]

> Hard disagree. GP comment was extremely useful to me, and I want this usage to continue.

If I told you everything it said was nonsensical and not at all what is reflected by the papers, would you still think it was useful?

I'm not a physicist of any stripe, so I actually have no idea about the accuracy, but I'm not willing to believe a bullshit generator just because it sounds convincing.

[ttpphd]

My only reaction is extreme amusement. The idea that there is some sacred collection of words we must keep pure is funny.

[ThrowawayR2]

dang has indicated that machine generated content is not welcome on HN: "HN has never allowed bots or generated responses. If we have to, we'll add that explicitly to https://news.ycombinator.com/newsguidelines.html, but I'd say it already follows from the rules that are in there." The full post is at https://news.ycombinator.com/item?id=33950747

[ttpphd]

Not breaking any of those rules, but it could be argued that y'all are breaking the one about tangential annoyances.

My comment was good. It provided primary source material and supplementary background material, then provided a clearly and explicitly marked summary of the content.

[anigbrowl]

If that was your takeaway then I did a bad job of communicating, because that was not the idea at all.