jack on Nostr: The first Twitter “algorithm” (which I guess is what we’re calling it now) was ...
The first Twitter “algorithm” (which I guess is what we’re calling it now) was named “In Case You Missed It.”
It was a small module of 3-7 tweets since your last session at the top of your timeline which, based on your history, would have the highest probability of you replying, retweeting, or favoriting (our version of “like” at the time).
It was relatively dumb but very effective (for both our customer and the company). A quick way to save time from scrolling through hundreds of tweets and “missing” something.
I was reminded of it last night as I tried to catch up after nostrica. Obviously that lead to whatever the “for you” tab is today, which as far as I can tell is more or less the same as when I left, but all the context labels are removed. And maybe there are some new weights on news topics and media?
Fascinating that most of this work will now likely be offloaded to GPT-*. And concerning that the only “openness” we’ll get out of the decisions it’s making are the prompts.
It was a small module of 3-7 tweets since your last session at the top of your timeline which, based on your history, would have the highest probability of you replying, retweeting, or favoriting (our version of “like” at the time).
It was relatively dumb but very effective (for both our customer and the company). A quick way to save time from scrolling through hundreds of tweets and “missing” something.
I was reminded of it last night as I tried to catch up after nostrica. Obviously that lead to whatever the “for you” tab is today, which as far as I can tell is more or less the same as when I left, but all the context labels are removed. And maybe there are some new weights on news topics and media?
Fascinating that most of this work will now likely be offloaded to GPT-*. And concerning that the only “openness” we’ll get out of the decisions it’s making are the prompts.