Lightnews — Scholar-powered news

spacecowboy @spacecowboy17.bsky.social · 8m

2. The items and the dividers are ordered by the user. I'm thinking of how to attach the notion of "priority" to them for the purposes of making recommendations to the user. Should we have a "rating" attached to each item and/or divider. Or should we use the order in the collection as the priority?

spacecowboy @spacecowboy17.bsky.social · 13m

Thanks! Now I know: www.sofahq.com

Please let me know if the demo UI could be improved with some features from that app.

Sofa: Downtime Organizer - Sofa

Create lists of apps, books, movies, tv shows, podcasts, music albums, and video games to enjoy later.

www.sofahq.com

1

spacecowboy @spacecowboy17.bsky.social · 18m

Yes, the initial step is the single-player experience where you can create collections.

The real goal is a collaborative discovery.

spacecowboy @spacecowboy17.bsky.social · 22m

Share our thoughts on the lexicon draft: rankthat.net/about

Maybe we could build a lexicon that is useful across apps.

RankThat - Collections

rankthat.net

1 3

spacecowboy @spacecowboy17.bsky.social · 26m

I asked Claude to fake the PDS calls with localStorage. So the UI works with it as if it's an async PDS API. Hopefully, it should be easy to switch to the real PDS backed implementation.

1 3

spacecowboy @spacecowboy17.bsky.social · 29m

Open questions:

1. How to represent seen and disliked items? Store them as a divider that has "seen items"/"disliked items" type? Or should they not be stored in PDS at all and be private on the app server? Or maybe give that choice to the user? Or encrypt them in PDS so only the user can read?

1 1

spacecowboy @spacecowboy17.bsky.social · 34m

Once I'm happy with the UI (which I think is pretty close) I'm planning:
1. Create the lexicon
2. Replace faked PDS calls on the client with the real PDS calls
3. Build a recommendation "app server" that would have similar logic to For You

1 1

spacecowboy @spacecowboy17.bsky.social · 37m

I'm starting with a client only UI prototype and I'm faking the PDS calls with localStorage: bsky.app/profile/spac...

spacecowboy @spacecowboy17.bsky.social · 44m

I want to make an atproto app for creating collections of anything - create book shelves, rate movies, collect science papers, fictional character tier list, etc.

Before committing to a lexicon I want to understand what a good user experience would look like and then build the corresponding lexicon

4 8

spacecowboy @spacecowboy17.bsky.social · 38m

I changed that to two levels: collections that hold items and dividers (like a book divider) as a flat list. You can drag either items or dividers, you can delete dividers and add new ones.

Give a try to the demo and let me know what you think. I'll pass it to Claude :)

1 2

spacecowboy @spacecowboy17.bsky.social · 40m

It’s been super useful to do the prototype. At first I wanted to use a 3 level hierarchy: collection -> shelf -> item. But that turned out to be too heavyweight. As a user you need to think about the shelf structure and then dragging items within and between shelves was awkward.

1 4

spacecowboy @spacecowboy17.bsky.social · 41m

So I’ve spent some time with Claude Code to build a demo website: rankthat.net

This is just a UI prototype - all data is stored in local storage, no PDS integration. The recommendations are hard-coded. More details and the current lexicon draft here rankthat.net/about

RankThat - Collections

rankthat.net

1 2

spacecowboy @spacecowboy17.bsky.social · 44m

I want to make an atproto app for creating collections of anything - create book shelves, rate movies, collect science papers, fictional character tier list, etc.

Before committing to a lexicon I want to understand what a good user experience would look like and then build the corresponding lexicon

2 1 18

spacecowboy @spacecowboy17.bsky.social · 52m

github.com/bluesky-soci...

I'm doing a prototype UI for an atproto app and I fed ^ to Claude.

Draft Lexicon Style Guide (Lexinomicon) · bluesky-social atproto · Discussion #4245

More folks are starting to design Lexicon schemas from scratch, which is great! This is particularly driven by ecosystem projects like Slices and microcosm which make it easier to work with AT netw...

github.com

2

spacecowboy @spacecowboy17.bsky.social · 4h

Let me know what works and what doesn't in For You so we can make it better.

It already demotes generally popular content (often politics) but we can further tweak it.

2

spacecowboy @spacecowboy17.bsky.social · 9h

For You stores all likes over the last ~60 days and then generates the personalized feed on user request.

1 1

spacecowboy @spacecowboy17.bsky.social · 9h

Can you elaborate? Are you talking about the Bluesky app UI?

I was thinking creating a "Customized For You" - a version of "For You" that people could configure the parameters like how much to penalize popular content (popularity_penalty=0.2).

Are you thinking along those lines?

1

spacecowboy @spacecowboy17.bsky.social · 10h

You can peek at anyone's For You feed here: linklonk.com/bluesky?user...

I find it interesting to see how different the perspectives/interests can be.

Note, this playground mode only uses your public likes - not your private "show less"/"show more".

Bluesky "For You" feed playground

linklonk.com

1 1

spacecowboy @spacecowboy17.bsky.social · 10h

Does "show less like this" not work well for you?

Regarding trusting dislikes of others I am thinking of this as part of a solution for sorting the reply section: bsky.app/profile/spac...

spacecowboy @spacecowboy17.bsky.social · 3d

Then when sorting the replies we sum the "like trust scores" of people who liked that reply.

There is another set of "trust scores" - how much you trust the dislikes of other people.

1 1

spacecowboy @spacecowboy17.bsky.social · 10h

Yes, it only affects your feed. Here is how it works:

spacecowboy @spacecowboy17.bsky.social · Sep 3

When you press "show less like this" we make likes of people who liked that post have less importance for you.

Specifically, we disregard one like that you have in common with them.

As a result you would see less content from them.

1 1

spacecowboy @spacecowboy17.bsky.social · 1d

Glad you liked it! Here is the link:

1

spacecowboy @spacecowboy17.bsky.social · 1d

Thanks! I guess you need to rebuild it if the follow graph changes? Is it an incremental update or from scratch?

Also can you please share the QPS for Following vs Discover? I want to get a sense of the worst case scenario for serving For You.

5

spacecowboy @spacecowboy17.bsky.social · 1d

Is it literally fetching post ids of all your follows from a SQL db? Or does it have custom logic?

1 6

spacecowboy @spacecowboy17.bsky.social · 1d

Whoever you are paying - it is working!

teq @quilling.dev · 8d

Im being extorted Im paying £70 a month to show up in the For You feed this is criminal

1 6

spacecowboy @spacecowboy17.bsky.social · 1d

Try For You - it finds people who liked the same posts as you, and shows you what else they've liked recently.

Right now it is mostly BB27 content for you.

spacecowboy @spacecowboy17.bsky.social · 1d

One thing I want to explore in the future is to give more weight to "show less".

1 2