Elon Musk wants Twitter to reveal the names of employees responsible for calculating the percentage of social media site users who are bot and spam accounts. Some people close to the entrepreneur, who for some weeks have been engaged in a bitter legal battle with the sociel network after announcing that they want to cancel the agreement signed last April to buy 100% of the company for 44 billion euros, told Bloomberg. dollars.
The bot issue is the main reason Musk took a step back. Yesterday, Tesla’s number one sold 7 billion shares of his company to prepare for the possible acquisition of Twitter which, if it is not made mandatory by the Delaware court called to decide on the case, he will be able to buy back.
Twitter sued Musk to force him to finalize the deal. The trial will begin on October 17th. Both sides have issued subpoenas to banks and consultants to gather evidence. In a letter filed Tuesday, Musk’s lawyers asked the judge overseeing the case to force Twitter to hand over employee names so the defense team can question them, the source said.
The character
Who is Kathaleen McCormick, the judge who will decide who is right between Twitter and Musk
by Arcangelo Rociola
The fake account front: Musk’s calculation, and a possible mistake
An Italian data scientist is at the forefront of the fake account issue, Andrea Stroppawho a few days ago shared an analysis on Twitter in which he stated that those who follow Musk in the calculation of fake accounts use Botometer, a social account investigation tool that would demonstrate how fake accounts can be between 10 and 14 percent, two to three times more than the 5 percent reported by the company founded by Jack Dorsey. Stroppa has received public acclaim several times, sometimes Elon Musk’s support for his analyzes.
Andrea Stroppa, the good hacker who could save Musk
by Pier Luca Santoro
Stroppa does not appear to be one of Musk’s consultants. But it clearly supports his thesis: “For years, most researchers have claimed that Twitter’s estimates of spam and fake accounts were inaccurate. Nobody seemed to care. Now Musk has opened Pandora’s box, “he wrote on her profile, claiming that the company gave” vague answers “to Musk and accusing her of” giving a fake dataset “, already” cleansed of malicious accounts “. Musk also supported her reasoning in this case. “Good summary”, wrote the entrepreneur in response to Stroppa, and then added that if Twitter proves the correctness of its calculation on fake accounts, “the agreement will proceed in the original conditions”, otherwise, “nothing will be done” .
Can Botometer Be Wrong? A thesis
But there are those who doubt the goodness of the bot analysis tools used by Stroppa and Musk’s data scientists. In a post on his blog Matteo Flora, chief executive of The Fool and online reputation expert, believes that while Botometer “remains an incredible tool for identifying potentially automated accounts, and while it may be incredibly accurate, the margin for error, even in idyllic conditions, is too high to be used in econometric estimates. And it can lead to serious miscalculations ”.
The analysis
What’s Happening to Elon Musk?
by Pier Luigi Pisa
What can this entail? “The difference between the 5% calculated by Twitter and the 12-15% declared by Stroppa could simply be within the margin of error of the algorithm used, which does not maintain a 100% accuracy score.” So the analysis could be distorted by some flaws in the search tool.
But there’s more, adds Flora: “Bots aren’t fake accounts. Spam / fake account classification itself is very difficult, because you could manage a fully automated account without it being spam (e.g. automated tickers, rss feeds, or even automated services like Whale Alerts). Botometer researchers’ estimates of the bot population vary between 9% and 15%, but it should be clear by now that these figures do not represent spam, only bots, ”which does not necessarily mean fake accounts.
Whatever happens, the Delaware court is not an easy task.