Would be awesome if social media clients somehow integrated with chatgpt and automatically summarized linked articles so it’s easy to see whether an article is worth reading.
I imagine it would save so many human hours.
Would be awesome if social media clients somehow integrated with chatgpt and automatically summarized linked articles so it’s easy to see whether an article is worth reading.
I imagine it would save so many human hours.
I’m also decently sure that soon enough it would mean them summarizing themselves
💯 probably already true to a huge extent. All the more reasons for this functionality.
It will happen soon. But then you have to trust that AI to be unbiased in how it summarises the info...
I don’t care if it is. Low level summary would be enough to decide if I want to see the full details.
Depends on the article. If it's just technical details it doesn't matter. If it's news about something controversial it matters a lot.
You could easily summarize the main points without opinions, just state what the article talks about. It’s really not difficult. Sites like that already exist.
And if you doubted something, just read the source. I’m not going to base an opinion on a summary; just want to see if it’s worth clicking.
You could. I'm just saying, if this relies on OpenAI APIs, I simply don't trust the source of the training data.
I'm not against the idea as a concept. Just highlighting the importance of open LLMs over closed corporate controlled ones.
Didn't think that'd be a controversial take on Nostr of all places.
I’m not sure that it really matters though. How could it be any more “dangerous” than headlines (which we have now)?
It could take an article and totally leave out certain information to make it one sided. It could be programmed to flag things as "fake news."
It is like trusting a single editor to curate every single media outlet.
Then you’d quickly notice a pattern and not use it 🤷♂️
You'd think so but reality tells us that the majority of the population is fairly oblivious to bias that conforms to their beliefs. And centralised editiorial control of all news without open source just seems like a bad thing to me.
It’s not centralized control of all media though. We’re talking about two-sentence summaries.
If widely deployed to "normie" social media it would rapidly become the primary method by which the majority of the population consumes their news.
Exactly. No need to complicate things just a basic summary without an opinion.
“Author describes … and supports his opinion with …” example.
Exactly
And we can only be responsible for ourselves. I don’t think it’s right for us to police things for other people. I guess that’s part of why I’m here.
You can tune OpenAI to your own preferences already.
You can, but ChatGPT as a specific product is trained to be biased, it's been shown many times. OpenAI isn't unbiased so nor are their products.
Such things should be based on actual open source LLMs like the LLaMA models we can tune directly and audit openly.
Biased outside the domain of US politics / current affairs?
a tl;dr bot. sounds cool. but yeah then you have to trust the bot. i like getting summaries from different humans who shared the article because they have their own version. add a bot’s version into that mix.
Everyone is biased. I think a bot could reduce bias by focusing on first order statements, leaving out anything that sounds like an assessment of the content.
It could do. Depends who trains it.
it would sure be better than what happens now…most people read the headline and move on. and we know headlines are heavily manipulated for inducing human behavior. a bot summary that aims to pull out objective content of news articles, ignoring superfluous opinion, that could be 🔥 👍.
Ya id want it opinion-free.
iframe 2.0 😀
That'd be really good. But what I don't get is how we waste so much time on lots of daily crap but when it comes to reading, we want to save time.