๐Ÿ’ค 71387

Today I read an interesting article online but then saw that it says at the top "Summarize this article with" and then lists a bunch of AI models. I would personally never be able to post my articles on a site that has such a feature. It sickens me.

3 weeks ago ยท ๐Ÿ‘ knusper

Actions

๐Ÿ‘‹ Join Station

6 Replies

๐ŸŒ™ xoagray

Yeah, those thngs are pretty much everywhere now. I've gotten to where I for the most part just ignore them. Half the time they seem to pull in stuff that isn't even related to what I'm looking for anyway. ยท 2 weeks ago

๐Ÿ‘ฝ lars_the_bear

Personally, I have no objection to LLMs being used this way. It seems like something we might actually be able to trust an LLM to do adequately. In academic publishing, though, I would have thought publishers still expected authors to provide an abstract.

I think it would be interesting to carry out a study, comparing the AI's summary with the authors', to see which people found most helpful. Perhaps this has already been done. ยท 3 weeks ago

๐ŸŒฒ half_elf_monk

Summarizers have been around for a while now, well before the bloom of "LLM" tech. That it appears on an academic journal is awful, but perhaps is cause to reflect... inhowfar do people treat "peer reviewed" as a lazy crutch, an appeal to guild authority rather than a mark of excellence. ยท 3 weeks ago

๐Ÿช knusper

Peer review should catch this, but Academia often is a self-perpetuating mediocracy ยท 3 weeks ago

๐Ÿ’ค 71387

@knusper I just remembered that I've watched a video some time ago of someone programming an algorithm based on a paper where it was extremely hard to understand the actual algorithm because the whole article was filled with a bunch of unconventional phrases and unneeded "slop", so it felt like they are trying to release a paper but "gatekeep" that information from people not being used to reading such papers. Maybe they should implement something to clear this up first, I doubt anyone would be mad about it. ยท 3 weeks ago

๐Ÿช knusper

Sadly, even reputed peer reviewed scientific journals start implementing such a feature. ยท 3 weeks ago

Proxied content from gemini://station.martinrue.com/71387/89f18b0a392745dcbcbd3f14d7a83cb1 (external content)

Gemini request details:

Original URL
gemini://station.martinrue.com/71387/89f18b0a392745dcbcbd3f14d7a83cb1
Status code
Success
Meta
text/gemini
Proxied by
kineto

Be advised that no attempt was made to verify the remote SSL certificate.