Hacker News Reader

Comment by d4rkp4ttern | Hacker News Reader

I have a daily workflow of scanning r/ML and HN and I subscribe to a few newsletters that I came across. I save bookmarks of tools and repos to raindrop.io and articles to readwise/reader. One good trick is to use the readwise feed email when subscribing to newsletters, so the newsletters go to Readwise instead of your personal email.

My big unsolved problem is Twitter — how do I avoid going on twitter more than a half hour a day, by using some type of twitter based filter/aggregator? Labml daily is a relatively good trend aggregator informed by Twitter. But I still keep discovering interesting things on Twitter not covered by any of the above. And BTW I bookmark twitter threads to Readwise/reader as well.

View on HN

I used to deal with the "avoiding going on Twitter" problem by subscribing to interesting AI Twitter feeds in Feedbin, an RSS aggregator. Unfortunately that doesn't work any more because the "genius" in charge revoked Feedbin's Twitter API access a week or so ago. So now I don't check Twitter at all.

https://feedbin.com/blog/2023/03/30/twitter-access-revoked/

The genius didn't lose anything here. If you never go on twitter you are a feedbin user. You make twitter no money. No one likes a mooch.

You gotta be trolling, nobody is that dumb...

The people posting on Twitter do it to be heard by others. As the audience decreases, the significance of the platform decreases. Thus, people stop posting on said platform and use other avenues to get their voices heard.

As users drop off, advertisers leave, removing a large part of their revenue.

Your claim is so completely out of touch with reality...

The koolaid level is high with you. No one stops posting on twitter because feedbin can't repost. You care about people engaged not someone who doesn't follow or comment or show up in the view column

Who are a few of your best twitter follows for AI?

I think Twitter lists are a good starting point. There are plenty of ML/AI lists around. I start from there and whittle down to good signal/noise ratio, meaning avoid people who post overly frequent fluffy hot-takes. People who retweet good stuff are also worth a follow. Ultimately what I want to get out of twitter is tools, papers, good blog-links.

For specific research-oriented follows I would suggest starting with Ilya Sutskever and see who he is following or retweeting. And for practical stuff start with Simon Willison

Reply on news.ycombinator.com