Wouldn't it be much easier to simply blacklist and put out some bland statement about "not supporting 'hate speech'" or something? It doesn't matter if it's true or not, it just matters that they paint the picture of them opposing something "bad". Lying just seems a lot easier than going to all of that effort.
It doesn't require more work than they are already presumably implementing for spam detection (unless I way too much overestimate big tech's abilities), it's a little touch to an already-being-implemented system with a bonus of being able to block competitors to some extent, without direct liability.
It's dead obvious that they do not want it. The reasons why is beyond me.
Someone whose job it is made a few mistakes. you must have made a shit ton of mistake too. Claiming otherwise is stupid.
What you and i can pinpoint on twitter spam is independent of "everything is evil". Difference.
Twitter is just too powerful with all those engineers and access to ML to have this just as a mistake.
Whoever is responsible for that flagging is either:
- Incredibly incompetent and not knowing what they were doing (which I honestly don't want to believe)
- Knew exactly what they were doing and it's intentional with a benefit of doubt with plausible deniability.
Unintentionally blocking a whole domain "by mistake" falls into the first point above.
Knowing track record of social media companies hiring good engineers and being evil, also knowing Mastodon being a competitor to Twitter, I don't think I need to tell much more.
Design the system in a way that it'd pick up competitor links as spam (without directly blacklisting them explicitly) using fine-tuned training data.
In case of media attention, tell it's an error and ML algorithms are blackboxes, and take action to "whitelist" those domains only when needed.
Yeah, we buy it, Twitter.