I want to talk about uptake of Fediverse sharing on Threads so far. I'm gonna share some information from my personal network. I'm very involved in Fediverse Threads, so these are probably high.
I used a technique I read about from @quillmatiq . the steps are as follows:
1. request threads data download.
2. extract data of accounts i follow
3. use `toot` command line tool to follow from masto
4. count how many follows fail
i did steps 2-4 with bash, sed, toot and grep. i might try to automate it in the future.
out of 865 people i follow on threads, 668 were not followable from masto. That is 78%; so about 22% have sharing enabled.
people i follow might not be representative. a weird thing in directed networks, like this one, is that we mostly follow people who have more followers than we do. and people who have lots of followers probably value being followed, so the probably are more likely to turn on Fediverse sharing.
im going to try the same experiment with my followers and see what the stats are for them.
OK! Of 2034 followers extracted, 621 had fediverse sharing enabled. That's 30%. a surprising result. my guess is that i probably follow a lot of people not interested in the Fediverse, but interest in the Fediverse is one of the main reasons to follow me.
Anyway, if take the smaller number and say about 22%, and extrapolate to all 200M Threads users, that means we have something like 44M Threads users on the Fediverse *right now*. HFS!
Earlier you correctly warned that your sample set likely is not representative of Threads users as a whole, though. So, extrapolation to the general is likely not valid.
As a counter example, @jwz found:
> Of the 655 accounts I used to follow on Instagram, […] exactly ZERO can be followed from Mastodon dot social. If there's something they have to turn on to make it happen, zero of them have done so.
https://www.jwz.org/blog/2024/08/rumors-of-threads-dot-net-federation-have-been-greatly-exaggerated/
So, extrapolate from that, and it's *zero*.
@bignose @evanprodromou @jwz absolutely. Another option is that @jwz wrote his testing code wrong and got zero results. Jamie, if I send you a bash script to run on a file with one username per line, would you try it again? Or do you feel pretty confident you got it right the first time?
@bignose @evanprodromou @jwz just read his post. He did it manually. I'd still be happy to help confirm.
@bignose @evanprodromou @jwz I'll just scrape his following list from IG, actually. Easier to double check!
@evan @bignose @evanprodromou My technique was brute-force simple. Get the list of every Instagram account I followed. Convert those to [@]name[@]threads.net. 41% were not 404. Of all the rest, none succeeded at federating when I pasted them into the search field on mastodon.social. (Though I do follow one or two other threads accounts, such as POTUS, so it's not that they don't work at all on this instance.)
@jwz @bignose @evanprodromou excellent. Sounds pretty reasonable. I've had problems in Mastodon search when an account hasn't been followed on that server before, but I'm sure you have too so it's probably not that.
@jwz @bignose @evanprodromou So, I scraped your following list on IG and found only two hits, @therealelvira and @whscullin , but they just enabled Fediverse sharing today. Even some of the people I expected to have it turned on, like @bradfitz@inuh.net and @textfiles didn't have it. By contrast, 22% of my followeds and 30% of my following on Threads have it turned on. So, very different uptake rate! Thanks for doing the analysis.
@evan Well at least I can follow Elvira now :-)
@jwz Elvira was already on Mastodon: @TheRealElvira :-)