Lawsuit protection and the pairing of tweet data with internal demographics data? There’s no telling what Twitter knows about its users from being installed on their phones(geocoding, network analyzation, etc.) to analyzing their advertisement engagement to sentiment analysis of word choice and it goes on and on.

As stupid simple as the tweeting/re-tweeting mechanisms are, that doesn’t mean they don’t have serious targeted marketing algorithms and other high end data analysis tools behind the scenes. It would be silly if they didn’t.

I don’t know, I think he was trying to get the price down because why pay more than you have to?

Also, I think a decade’s-worth of notable human discourse in discreet packets of n-characters in an already hierarchical organization paired with deep demographic metadata could be seriously valuable.

Maybe… I’m sure terms of service involve owning tweets and safety from the same sort of lawsuits you’re seeing against Midjourney for training on other people’s intellectual property.

He’s kicking into x.ai. There are good reasons to think that he bought Twitter to use its massive data-set of human language to train an ai - NOT because he gave/gives a shit about Twitter as a profitable company.

If it becomes a paid competitor to OpenAI, it solves its profitability issues without enhancing “Twitter” as a social media site. Anything he said about improving Twitter was likely a lie designed to prolong usage of the platform and enhance the dataset.

Yeah, I was going to say they just left a default alphabetical sort to their global droplist component and called it a day. Probably works fine in most contexts, but this one - not so much.