[2023-09-26]

I want to vent a bit about some concepts which only recently become possible and if they’d work for us.

As billions of internet users, would we get products that work for our needs a bit more than today?

I mean individually, there are probably 20+ software products with Bn+ MAU, each if them obviously doing a good job on average - would individuality take us a bit more forward?

Few points right away on personalisation things that I’d want to see there in 1-2+ yrs:

• Local-first
• Product experience completely revolves around a human’s request (calling it “offensive privacy in my short demo”)
• As it’d be a new product, data needs are majorly different too

On first few points, not a big fantasy - have you seen a D1 releases and ideas around, that’s like available around the corner?

Individual DB per user (scale up to millions databases), feature flags on different scale, distribution of opt-ins, opt-outs, deletion of data from models - a lot of things just enabled by infra. (link for the tweet).

Of course, it’d be a bit until adoption, but does personalisation pays off? For largest tech - yes, on average idk.

Maybe the most promising use-case of above lies in the shadows - each person could get a time representation.

If it can enable cheaper multi-faceted recommendations for large systems - well, profits on ads / streams and anything on e-comm could be gigantic. (not creating a billions of world’s GDP serving ads on just purchased things for example will be nice, or simpler - not changing your Spotify/Youtube recommendation streams by few random searches)