[ceejayoz]

I see people holding these sorts of theories about Facebook algorithms, too. They tend to be entirely bullshit.

The folks taking their accounts private aren't doing any sort of scientific test, and there are folks who try it who don't see engagement go up. Unless you're somehow doing statistically significant tests with large groups of users, the same content, at the same time, there are an immense number of confounding variables here.

[CodeWriter23]

Here's what he found subsequent to taking his account private. You can judge for yourself it these were major bugs or not. But the absolute vs. proportional block count is clearly an error that would effect engagement.

To me it seems logical, if you find a certain type of error like that, you go searching elsewhere to see where previous coders made the same kind of mistake. (Even when previous coder = myself)

https://twitter.com/elonmusk/status/1624660886572126209?s=46...

[SilverBirch]

The problem is you can’t believe him. Were those real issues? Maybe, but what we know now that anyone who told him his lack of engagement was organic got fired. So all that’s left is to come up with some reason to justify why Elon is less popular than he thinks he should be. Whether it’s true or false doesn’t matter, because he’ll keep firing people until someone tells him what he wants to hear.

[ceejayoz]

Setting aside "he found" for a moment - given reports of people being fired for giving accurate information, "someone told me" seems more likely - I evaluate Musk's claims like "Fanout service for Following feed was getting overloaded" in the context of previous claims like https://twitter.com/elonmusk/status/1604616863673208832 (lie) and https://twitter.com/elonmusk/status/1026872652290379776 (lie).