Short form content feeds are like unloading a dumptruck full of random items into your driveway. Is it actually better if all that stuff you didnt ask for is real information that needs to be organized and pieced together with what you already know without any of the associated context that helps you do that? Or is it better if you know 99% of it is trash and you dont need to remember any of it?

I think a tool like this is great for people who want to use short form content intentionally, and personally that only happens when I am bored and in need of a new topic to research. I think of all short form content like marketing/ads, just showcasing something i might be interested to dig into on my own. It's how i used StumbleUpon website back in the day.

But I have noticed I am rarely using short form content with intention. its because i want to check what my friends have posted, and then with the extra downtime i scroll a little bit, and sometimes get stuck

These attempts to make educational short form content still suffer from the same drawbacks, so I wonder how effective they truly are.

[0] https://cyberpsychology.eu/article/view/33099

[1] https://www.tandfonline.com/doi/10.1080/09658211.2025.252107...

I think it would be nice if you could do a non simple English version but nevertheless happy with what you've put together, and I've added a shortcut to my phone. Please don't let the negativity stop you from continuing to work on it.

1. https://www.wikitok.io/

The United States Virgin Islands are a group of islands in the Caribbean Sea. They are currently owned and under the authority of the United States Government. They used to be owned by Denmark (and called Danish West Indies). They were sold to the U.S. on January 17, 1917, because of fear that the Germans would capture them and use them as a submarine base in World War I.

https://simple.wikipedia.org/wiki/United_States_Virgin_Islan...

I easily have over 100 tabs of wikipedia open at any one time, reading about the most random stuff ever. I'm the guy who will unironically look up the food I'm eating on wikipedia while I'm eating it.

No need to try to make it "doomscrollable" when it's already got me by the balls.

- https://sixdegreesofwikipedia.com/

- https://www.wikitok.io/

I've been swiping a lot for the last 10 minutes and I'm not sure how much it's learning. I have some feedback.

- I have never liked or clicked a biography but it keeps suggesting vast amounts of those

- It does not seem to update the score based on clicking vs liking vs doing both. I would assume clicking is a solid form of engagement that should be taken into consideration

- It would be interesting to see some stats. I have no idea how many articles i've scrolled through or the actual time spent on liked vs disliked article previews. If you can add such insight it would be interesting

- A negative feedback mechanism would be interesting as well. There is no way to signal whether I'm just neutral towards something (and swipe through) or actively negative about it (which is a form of engagement the doomscroll would actually use to show me such content once in a while)

- since this website has already shown me multiple pages about things I'm learning about thanks through it, it might benefit from a "share" button (another engagement signal) as HN folks are likely to want to share on HN things they've just learned

- Would you be willing to make the experiment open source?

But then I look at the comments, and it really looks like some people want this.

Now I'm depressed.

1. https://apps.apple.com/us/app/egghead-scroll-learn/id6630364...

It was timewasting or avoidant behavior for sure, and often described negatively. But at least you were learning something.

Silicon Valley then spent the next few decades trying to understand that behavior so they could isolate it, strip all positive value out of it, and make it highly profitable.

It's kind of like the opposite to my Wikipedia project Redactle.net which takes a lot of effort.

OP, since you're encountering load issues I would suggest narrowing your corpus to Wikipedia vital level 3 and caching all the content since it's only 1000 articles.

Should be: Please only continue if you're not at work.

DuckDB loaded in the browser via WebAssembly and Parquet files in S3.

Is it just not implemented, or would it be too much effort to undo the effect on your weights from liking the article?

Update, it does not run properly in Firefox on iOS. After it has loaded to 100%, the site refreshes.

Yeah, it got really sticky real fast. From the random (?) selection it starts off where I couldn't recognize anything but popular TV shows, it immediately over-indexed on that content and I had to fight for my life to see anything else in the feed that I would recognize and consider a good algorithmic pick for my interests.

Which is brilliant, because Instagram has the same issue for me - absolute metric tons of garbage and whenever there is a gem in that landfill of a feed that I interact with positively, it's nothing but more of that on my feed for weeks until I grow sick of that given thing. In conclusion, Instagram could have used this 30 line algorithm and I'd have the same exact experience when using it.

Algorithmic feeds are obviously problematic for turning several generations into lobotomized zombies, but they are also just not very good at nuance, so it is not even a case of something that's bad for you but it just feels so good. It's just something that's bad, but is able to penetrate the very low defenses in human psychology for resisting addiction and short-term gratification and there is no incentive to improve them for the sake of the user as long as they work for the advertisers.

I gave up after about a minute.

And a plug for my own (fiendishly difficult) Wikipedia-based game:

https://studiowgx.willgrant.org/wikipedia-golf/