I recap the news from 40 media everyday with GPT-4 coupled with ML and NLP
https://dupple.com/techpresso
Hey guys!
I've spent the last few weeks working on an algorithm that summarizes the tech news from the last 24 hours, gathering the data from a bit more than 40 media today.
I use the generated data to make a newsletter and twitter posts!
Don't hesitate to share feedback!
subscribed! do you have a tool to automate the newsletter curation/formatting, and an integration into twitter?
Regarding the curation, I use a homemade ML algorithm which I use to give me an idea of the most covered news of the day. I still have to handpick every topic in the end, but it clearly helps me a lot.
Regarding Twitter, once the newsletter is sent, I coded an algorithm that transforms the newsletter into a thread. So basically the integration into twitter is automated and doesn't require any extra work from me!
Hope it was clear!
And thanks a lot for joining, let me know if you have any feedback :)
Interesting idea! Personally I feel like there is too much information coming from different sources these days. So good to have ways to reduce the time I spend on analyzing them.
One initial thought from my side - would be good to be able to customize the types of news you receive. Considering you gather the data from 40 sources I imagine you have a variety of options of what you can highlight in the newsletter.
Hey, thanks for your message :)
Regarding the types of news you receive, it's true that the end goal would be that anyone would be able to choose which kind of news they receive by being able to select topics etc.
For now, I've been dissatisfied by any "algorithmic" solution that would choose which tech topics would interest the audience the most, as I cannot know which articles the readers liked the most (I cannot know which summaries were the most read for instance). Thus, I have to handpick every topic by hand and it's not something that can really scale to allow personalized newsfeeds.
Nonetheless, I plan to launch newsletters on other topics (the next one will probably tackle finance and business as I think there's a huge reader base for this).