Hacker News Reader

Which is nonsense, since the numbers in question are not verifiable.

How can you identify a "spam account"? It's not possible to definitively determine the intent of someone opening or using a new account.

So the numbers are arguable either way. Musk is using this fact to try to wriggle out of a disastrous impulse buy.

View on HN

It’s not that subjective anyone using twitter still sees spam accounts promoting a website, business, product, or scam on every post. Twitter just said they remove 1M spam accounts each day. Their mDAU is claimed to be around 200M. This means they are deleting around half a percent of their users every single day. That seems like a lot to be identifying to me, but doesn't really determine if they are deleting enough or not but does show me that being higher than 5% is very possible when you have to delete that many accounts every 10 days.

Why do you think Twitter list accounts it knows to be bots as mDAU? Those are very different categories.

I’m not saying they are. I’m saying that they are banning enough bots to make up 5 percent every 10 days. With that amount of bots being banned, the possibility of that amount also being missed from being banned is substantial.

Reply on news.ycombinator.com