zlacker

What a great way to see the bias play out in real time.

>>msla+(OP)
I noticed some bias because I only analyze sources in English. It should be easy to make ChatGPT also read news in other languages, maybe that will help. Is that what you meant?

>>msla+(OP)
A sentiment analysis plot of all the sources could be insightful both for users as well as a way to tune the weight of sources in the future. Very cool

>>spacep+Xi
Could you please expand your thought? Not sure I fully understand the idea.

>>yakhin+27
It also seems to be very US-centric and seems to put a strong focus on reporting on the economy. I personally found nothing particularly relevant to my existence featured on the site.

These two biases are probably indicative of a host of other, less obvious biases. To be clear: all media is biased, because it is created by biased humans and it has been well demonstrated that algorithms replicate the biases of their creators.

A good newspaper/website knows their audience and delivers reporting relevant to it in a language they appreciate and cognizant of its societal impact.

>>tfourb+Lq
The recent one is economic focused, but if you look back it does focus on other stuff like Starship launch, climate, health, geopolitics, energy etc.

>>yakhin+27
> It should be easy to make ChatGPT also read news in other languages

This should indeed be easy, and given how good it is at translation, very interesting - taking too news sources and national broadcast services; rate as you do now - filter the top, summarize in English, store and dedup on similarity vector - it would be great. For example i don't read any Chinese or African news sources - but it would be great to have them in the mix.

Might even index on similarity first then ask for a summary on all different reports on the same story?