I wonder if you can re-create subreddits with custom algorithms.

Another approach would actually be to focus on single community rigorously testing the content being delivered by members of this community and tweaking the algorithm until everyone is satisfied. This might guarantee a higher quality of the content being delivered but will be a bit specialised.

The most basic thing I can think of is classifying the posts using an LLM or the such into pre-defined categories (anime, movies, etc.) and then allowing users to choose a set of these pre-defined categories.

Ideally the algorithm would even be tweak-able by individuals and not a one-size-fits-all solution. However, I am not sure how I can even go about this.

asking the important questions

@jeffry.blue this is so cool

@decentralgabe.bsky.social how does it differ from Web3? Is it just a subset of apps/concepts that people understood under the Web3?

Did Web3 or even Web4 ever truly arrive for us to move on to Web5?

Are u sure Web5 is already a thing? 😅

I am starting to like you Paul. Respect for not attacking back

https://dance.co/ Dance Bikes subscriptions are cool. I pay ~50 per month and they take care of everything.

notifications too

Managed to get a first version of the profile tab working

the goal is to build my own recommendation algorithm that I can tweak and customize as much as I want.

It is getting there

As always will open-source the whole thing when I am done

the problem is that no matter what the outcome is it is very unlikely Google will be able to have a monopoly over whatever replaces search. They majority of their money is made out of ads displayed within search results. Major problem.

As correctly stated in the document there is no reason to believe Google will be able to ship the best LLM/product. Search as we know it is over. Google’s best strategy is to make as much money while it lasts while trying to make money in new ways.

why?