Ya im a big fan. The way they break down topics is great. But I was still left scrolling and scrolling.
I am using LLMs for what I am building and i'm envisioning a system that understands what you read yesterday so its able to give you more concise and direct information.
I also enjoy the completeness of a set amount of stories.
I think there should be a middle ground where you dont have to completely remove yourself from the news, just the endless feeds and opinions. It does seem helpful to know what is going on with the world. Whether thats socially or just understanding other events. I fully agree there is a limit and most of what is out there is junk but not all of it.
I have been actively trying to build something to get away from the endless feeds of news.
Essentially a modern day newspaper. So you can read what is important and then be done for the day.
Thats exactly what I am doing here just with a massive list of RSS feeds. Changing it to a list of personal RSS feeds would be a pretty cool idea. It would allow your own personal digest for the day.
Last year artifact inspired me to build out my own AI news app. Even though they decided to shut down I still think there is potential here for more niche products not looking to grow to massive scale.
My news app, Quill, focuses more on completion than endless scrolling. Similar to a traditional newspaper. ~10 stories a day and you can be done. Similar to yahoo news digest from years back.
Hi this is a small side project I've built using the latest LLMS. It gathers 10,000+ articles every morning and creates a digest of the most talked about stories.
So far it's done a good job at keeping me on top of the news without having to visit different sites or doom scroll Apple News.
It's an ever improving project to keep up with the AI advances. Even in the past 6 months there has been a 3x reduction in cost and a huge improvement in reasoning. In the future this could be an even more powerful newsroom of sorts.
I've been building a daily news digest with intent of learning something new and creating something cool.
Although AI is the heart of the product, I have noticed that I have been using it as an "anything tool". Though the articles talks about more broad use cases like hardware and robotics, I've found its an anything tool for most of the boring use cases.
Using python for the first time in a while, ask GPT. Need a tweak on the app store copy, ask GPT. What are the top news sources, ask GPT. Data formatting & manipulation, save 5 minutes ask GPT.
I've built Quill to help simplify the news. I found myself doom scrolling through news apps and wanted something different.
At the moment Quill is built on the same foundation that Yahoo News Digest was. Summaries, Quotes, Images, and more.
Every day it combines thousands of articles and creates 6-10 collections of the most important news stories. I've found it helps me keep up to date without spending hours reading the news.
I think that its a helpful addition to anyone who regularly uses apps like Apple News or similar.
Thanks for sharing newsminimalist. I've unknowingly built nearly the same thing but more visual with additions like maps and images. Similar to Yahoo News Digest from years back.
I like the "relevance score", it would be interesting to implement something similar to that.
I've been able to cut out the news for the most part with a small side project. It combines hundreds of sources and thousands of articles every day to make a concise digest.
In about 5 minutes I can get caught up without doom scrolling.
I built a news digest for myself after I couldn't find a good yahoo news digest alternative. It's a once a day feed of 6-10 stories. Simple and quick enough to keep myself informed.
This was an issue when I first started. With minimal sources a lot of time the top collections were low quality SEO articles.
After adding a sufficient amount of sources I've noticed a decent reduction in the echo chamber. Although by ranking importance by the most talked about topic, it is going to have some sort of echo chamber.
Adding left and right leaning publishers for instance has helped. Although one might say something is good and something is bad, the embeddings pick it up as the same topic.
I had nearly the same idea as you around surfacing what is "important" vs just a large list of RSS articles.
The main differences compared to what you are thinking are two things. One for the `Significance of the event` I've used the number of publishers talking about that event. So more publishers == more important. Two, I've done this in a daily fashion instead of a weekly report.
I can also confirm that the LM has the capability to do at least a days worth (2500+) of articles. I would doubt its capability to produce an entire article but it does a great job at a small summary.
Quill is a modern newspaper delivering 6-10 stories a day. Our main goal is to let you quickly catch up on the news and get on with your day. We user over 70+ sources to collect over 2500 stories daily and using AI (Embeddings) we combine all of those stories into Collections. A summary is then generated (LLM) for every article in the collection and the best is chosen by a human.
The old Yahoo News Digest app from years back was incredible and I never could find a replacement. I decided to build my own.
It uses Embeddings to gather thousands of articles and compare against each other. Then uses a relational graph to combine those into collections and uses an LLM to create a succinct summary, quote, map, and other info about the topic.
It lets me the news for the day within minutes instead of endlessly scrolling.