Stephanie Jones
W.A good turn in theater deserves another. Elon Musk has done just that.
The world’s richest man has complained over the past few weeks that Twitter is underestimating the amount of automated spam bots on his site and will not give him access to the data needed to conduct an independent assessment, which he previously needed. The acquisition of his company will be completed. According to Twitter, bot accounts are less than 5%. This number is consistent with many years of SEC filings. Musk was able to access this image from the beginning. His sudden obsession with it seems to be a work in progress. This is one way to make an excuse to reconsider its $ 44 billion offer for Twitter in the midst of a sharp fall in stock market prices.
On Monday, Musk’s lawyer sent a trivial letter on Twitter saying it was better to override the data. Or. Or? Otherwise, Musk and his attorneys would consider their refusal a violation of the consolidation agreement and rescind the agreement, a dubious argument that probably won’t stand up in court – but one they could use to make things unpleasant for Twitter.
On Wednesday, we learned about Twitter’s own mask-type move. This business That’s it Spoken Preparing to give him access to his so-called “Firehose” API, a stream of every tweet sent. (Each. Tweet.) This is about 500 million microblog posts every day. A Twitter spokesperson did not comment on exactly how he would share the trove with Musk, saying only: We believe this agreement is best for all shareholders. We want to close the transaction and implement the consolidation agreement at agreed prices and terms. “
Let’s be very clear: allowing Mask so much access was a mistake for Twitter. It’s a ridiculous move to make him understand. Do you need information? Here we have all the data. We’re not hiding anything. Enjoy! “It simply came to our notice then. You don’t want access to everything, “said Goran Murrick, a computer scientist at the USC’s Institute for Information Sciences who has worked with a similar Twitter API. In fact, Mask probably only needs to make the “Dichaus” API Twitter available to some researchers, which is 10% of all tweets. The difference between the results obtained from Decahose and Firehouse is “equal to the difference between a poll and a census,” Murrick said. Smaller, more easy-to-use sample sizes are good for poles. “And you can always get multiple polls and get pretty accurate results,” Murrick said, referring to what would create a more comprehensive census.
Of course, it’s funny to see Twitter pulling on the mask — to say its bluff, obviously, to try. However, the outcome of the lawsuit between Billionaire Mask and his company may not be significant. It will not be easy for Musk to analyze Twitter bot activity. To do this, a team of researchers would need to create software that could review tweets. Most of all, Mask could use his access to the Firehouse API to bring a guess of bot activity on Twitter, making it almost inevitable that his image would not match Twitter’s.
To begin with, the definition of what constitutes a bot account can easily differ from Twitter. Musk can define bot but he likes. Even among the top researchers in the field, there is no universally accepted definition. “If you keep two people at home and ask about the definition of anything, they will have different opinions – and especially on Twitter about the definition of what a bot is,” Murrick said. “So maybe if someone tweets more than 1,000 times a day, he’s a bot, right?” But maybe someone else will say, it should be if they tweet more than 50 times. The best tool for bot detection, the Botometer, which comes from the Indiana University team, offers only a fair measure of bot activity, giving a probability score that an account is a bot, never a guarantee. Mask has plenty of flexibility to use API Firehouses to draw the decisions needed to achieve its goals. This, in turn, could persuade Twitter to pay less.
Moreover, it is unclear whether even an honest-to-goodness effort by Musk could completely replicate the Twitter process, doubling its internal speculation. Specifically, Twitter says that less than 5% of its “monetized daily active users” account bots, an image of Twitter’s own creation. A Twitter app reports monthly active users if they log in at least once every 30 days. So it is uncertain whether even a well-intentioned mask can use the Firehouse API to count the same number of monetized daily active users.
A second question: What if Mask discovered a problem with Twitter’s Firehose API? Facebook bought CrowdTangle from Brandon Silverman. He mentioned on Twitter that Crowdtangle could make the situation worse.
Really, it doesn’t matter if the extra problem is real or fake. Theater is basically about make-believe.
