I’ve seen a few bot-tldrs and while I really appreciate the idea, I’m not sure if they achieve that 100%. The summary is often a bit long (sometimes just as long as the article) and as I understand it the one I see most often is grabbing sentences and stitching them together, which is nice since it keeps the actual content, but it can be a bit awkward to read.
Maybe yours is working differently, I think generally speaking it’d be best to have one paragraph summarizing the article (chatGPT-style), or having a tool that creates new headlines based on the content and replaces the actual headline itself.
Even with the perfect summary, there’s still the issue that people have to look at the comments to see it. I‘d imagine most people just scroll past the post itself.
I think it’s the one you see the most, that stitches sentences together. Using AI for this is very unwise, eventually it’s gonna make something up and it’s gonna straight up lie.
In my opinion the bot does a better job than most actual people would do when summarizing. It’s not perfect and it will never be, but I’m quite ok with how it works.
It’s definitely worth having that bot and I think it’s great that you’re putting the work in. It just doesn’t fix all the “misleading headlines” issues if people don’t even look at the comments or don’t read the (not super short) summary.
Well, to be fair, I’m fixing the issue for myself. I thought it might be nice to share with people who’d appreciate it, but first and foremost I’m writing the software that I want to use myself (which is true for my other bots or Lemmy-related software as well).
You are not logged in. However you can subscribe from another Fediverse account, for example Lemmy or Mastodon. To do this, paste the following into the search field of your instance: !technology@beehaw.org
A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.
Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.
That’s partially the reason for making my AutoTL;DR bot.
Doing god’s work
I’ve seen a few bot-tldrs and while I really appreciate the idea, I’m not sure if they achieve that 100%. The summary is often a bit long (sometimes just as long as the article) and as I understand it the one I see most often is grabbing sentences and stitching them together, which is nice since it keeps the actual content, but it can be a bit awkward to read.
Maybe yours is working differently, I think generally speaking it’d be best to have one paragraph summarizing the article (chatGPT-style), or having a tool that creates new headlines based on the content and replaces the actual headline itself.
Even with the perfect summary, there’s still the issue that people have to look at the comments to see it. I‘d imagine most people just scroll past the post itself.
I think it’s the one you see the most, that stitches sentences together. Using AI for this is very unwise, eventually it’s gonna make something up and it’s gonna straight up lie.
In my opinion the bot does a better job than most actual people would do when summarizing. It’s not perfect and it will never be, but I’m quite ok with how it works.
It’s definitely worth having that bot and I think it’s great that you’re putting the work in. It just doesn’t fix all the “misleading headlines” issues if people don’t even look at the comments or don’t read the (not super short) summary.
Well, to be fair, I’m fixing the issue for myself. I thought it might be nice to share with people who’d appreciate it, but first and foremost I’m writing the software that I want to use myself (which is true for my other bots or Lemmy-related software as well).
I see both sides. Thank you for your service; I for one find it helpful.