So they are part of my daily RSS feed.

I try to review 250 piece of content every single day.

(reviewing includes deleting boring stuff)

I have some filters/priority setup (around China, my favorite technical stack, favorite creators, etc)

So I have a fairly high signal/noise ratio

I don’t try to retain anything, if it’s really interesting I will take some notes, or it will inspire me a new piece of content.

In most cases I just skim through it and I’m just glad to know that it « exists » it will later help me build up connection subconsciously.

But I just do this because I’m really curious, it’s an excellent way to kill time during commute or whatnot

Feedly + Chrome + linkclump + tab groups + autocopy

Balabolka text to speech that watches clip board + appends copied text into one document

TS2 text to speech on android.

I have a ~100 feeds and ~20 news letters.

I have AI filters / deduplicate setup on feedly.

I'll go through priority list and save later anyhthing i'm interested in.

Couple times a week, I'll load up the read later list, use linkclump (hold Z + drag over list of links) to open everything in new tab. Clear all read later. Sort tabs by domain with tab groups.

Go through tabs by topic, drag and over article/parts of article, which gets autocopied to balabolka into a few different text files based on topic. Autocopy has option to save url and access time for later reference. I'll keep the tabs with interesting looking graphs for later reference as well. For frequently used sources, I'll have readermode on, or adguard rules to strip as much superfluous web elements as I can. Balabolka also supports regex to further filter out more things, i.e. sometimes I'll just copy comments from an entire thread and it will only read user name + comment.

Feedly Teams has option to make news letters according to tag, but seems like it's only available for enterprise which is dumb. Shouldn't be hard just to collage everything into a news letter, but here we are. Process usually takes 15m, couple times a week.

Usually works out to 10-15 hours of "audio" per week, which I listen at 2-3X speed on a text to speech app called T2S on android, so about an hour a day at the gym. There's a couple voices that makes listening at higher speed clear, "microsoft zira mobile" on desopt (need reg to unlock), and Android/Google "United States - 1" on mobile. Yes it's still pretty robotic but you acclimatize. If I'm really interested in saving something i'll take a screen shot on my phone of the text and then search for in the text file and find the url.

Have a script that backups the text files, basically keep an archive of everything I've consumed that I'll occasionally search through if I need to find source.

I find audio a good format, especially for keeping up to date. Generally multiple sources will talk about same shit, and you might hear same news/analysis a few times and skip, but repetition helps retain info.

Also useful for podcasts/lectures with transcripts (usually youtube). Throw that in an AI summarizer and copy text to balabolka.

I'm hoping all this will be dramatically simplified in the next few years. Would be nice if there's a RSS +LLM that can filter articles according to topic, summarize and throw into a mp3 by date. Same with podcasts. Maybe better natural language TTS. Maybe have bookmark option like in pocket cast.

E: prior to this there was also good service that makes twitter lists into digests, which was also very useful, but hard to do now. Have a very hacky system to scrape everyday which adds another 5m everyday for 5 lists with ~150 people I follow. Usually 200-300 tweets a day that takes ~30m to listen.