Yes...haha....haha....yes!

10% is likely the floor in this case, as somewhere between 5 and 7% have already been taken down. So higher.

Let me know if you find all the off label Wordle clones. I ended up implementing a rule to exclude all those posts from analysis to keep them from clogging up my reports, heh.

This one has been on the list for a while. There is (was?) a whole cluster of accounts with the same persona all doing the exact same bit in various languages.

It depends on what "verification" represents. Ask multiple people and you will get different answers in response.

Also, the social graph is not part of the core protocol.

@xblock.aendra.dev classifies those sorts of images. And if you read through Aendra'a posts, you'll quickly get a sense of why image classification is exceptionally difficult.

Imo, it isn't just about the compute required- there's also the overhead of dealing with erroneous classifications.

You're absolutely right about this. There are a fairly significant number of people who are engagement farming with tools on their otherwise normal accounts.

And at least one new tool pops up every week.

It's very difficult to analyze the numbers without first separating follows from starter packs from regular follows. I recommend treating them separately for the purposes of analysis.

Note the recent surge in operations.

>$500 ticket
>$19.99 payments

I’m sorry the phrase “two year Coachella loan” might actually kill me.

it is if it's hooked up to factorio

never been myself. and honestly it doesn't feel like i've missed much.

glad there's a livestream.

Someone's gotta do it
bsky.app/profile/case...

Literally just saw my log file and thought that was odd. You should be off the list within the next several minutes.

jerry no

restaurant profit margins are thin enough, but any restaurants that heavily feature coconut milk, fish sauce, gochujang, teff, berebere, cumin, sichuan pepper, avocado, black vinegar, chili crisp oil, cardamom, tamarind, or soy sauce are in for a world of pain