Elon Musk, unsatisfied with the ongoing court case over his attempt to break a $44 billion merger contract, has challenged Twitter CEO Parag Agrawal to a public debate.
"I hereby challenge @paraga to a public debate about the Twitter bot percentage," Musk wrote in a tweet on Saturday. "Let him prove to the public that Twitter has <5% fake or spam daily users!"
Of course, a Musk/Agrawal debate is unlikely to happen, and Musk's proposed debate would not be likely to prove any facts about Twitter spam that couldn't be proven at trial. Musk, Agrawal, or both could also choose to testify at the upcoming trial in the Delaware Court of Chancery. CNBC reported, unsurprisingly, that a "source close to the company says a debate is not going to happen outside of a pending trial."
Despite Musk's claimed eagerness to prove his point in a public debate, he tried to have that trial delayed until February 2023. Judge Kathaleen McCormick rejected Musk's request for a delay while granting Twitter's motion to expedite the trial, now scheduled to begin on October 17. "The reality is delay threatens irreparable harm to the sellers," McCormick said in her ruling.
In May, when Agrawal posted a thread explaining Twitter's spam-estimate process, Musk responded with a poop emoji.
Musk may be worried the trial won't focus enough on his claims that Twitter's spam numbers are accurate. "That's not what this case is about," Twitter attorney William Savitt said at the hearing on the trial date, calling Musk's spam complaint a "manufactured issue."
Twitter wrote in a court filing last week that Musk has no right to exit the merger based on the number of spam accounts, saying the agreement contained no references to false or spam accounts. "When Musk offered to buy Twitter, he did not ask for—and Twitter did not make—any representations regarding the number of false or spam accounts," Twitter wrote, adding that "Musk forwent all due diligence—giving Twitter twenty-four hours to accept his take-it-or-leave-it offer before he would present it directly to Twitter's stockholders."
Twitter also pointed out that Musk's analysis purporting to show that at least 10 percent of Twitter's active daily users are spam or fake used a web tool that recently labeled his own account a likely bot.
Musk's attempt to break the merger deal centers on his unproven claim that Twitter's publicly stated spam numbers are incorrect. Specifically, Twitter reports in Securities and Exchange Commission filings that fewer than 5 percent of its monetizable daily active users (mDAU) are spam or fake.
Musk's legal team wrote in a court filing that after he agreed to buy the company, "Musk was flabbergasted to learn just how meager Twitter's process was. Human reviewers randomly sampled 100 accounts per day (less than 0.00005 percent of daily users) and applied unidentified standards to somehow conclude every quarter for nearly three years that fewer than 5 percent of Twitter users were false or spam. That's it. No automation, no AI, no machine learning."
In another tweet on Saturday, Musk wrote, "If Twitter simply provides their method of sampling 100 accounts and how they're confirmed to be real, the deal should proceed on original terms. However, if it turns out that their SEC filings are materially false, then it should not." Musk subsequently started a poll in which 65 percent of over 822,000 respondents answered "Lmaooo no" to the question of whether "less than 5% of Twitter daily users are fake/spam."
While Twitter said its "quarterly estimates are based on daily samples of 100 mDAU," that adds up to 9,000 per quarter. "[A]s a basic statistical matter, the approximately 9,000-account sample Twitter reviews of accounts included in mDAU each quarter is sufficiently sized to extrapolate across the mDAU population," Twitter wrote in last week's court filing. Twitter also said it performs "multiple human reviews (in replicate) of thousands of randomly selected accounts each quarter using both public and private data."
A Musk court filing claimed that Twitter "does not perform even the most basic of human-verification processes—such as contacting the sampled accounts to determine if they are real, including by sending an email, text, or even a push notification on Twitter requiring them to enter a CAPTCHA" and "does not remove suspended accounts (which Twitter otherwise does not count as monetizable) from previous mDAU calculations—even when they are suspended for spam within the same quarter."
Twitter's response said that "the accounts included in Twitter's sample of mDAU have already been subjected to Twitter's automated spam-detection processes, which include processes requiring certain users to respond to phone or text notifications or complete a CAPTCHA. Twitter further avers that, after it determines an account is spam, malicious automation, or fake, Twitter stops counting it in mDAU." Twitter said it "locks millions of accounts each week that cannot pass human-verification challenges, such as CAPTCHAs or phone verifications."
Agrawal wrote in May that Twitter's "actual internal estimates for the last four quarters were all well under 5%... The error margins on our estimates give us confidence in our public statements each quarter."
Apart from the debate over spam estimates, Twitter wants the judge to focus on Musk's contractual obligation to complete the merger. "Musk's Counterclaims, based as they are on distortion, misrepresentation, and outright deception, change nothing. Musk signed and is obligated to consummate the merger agreement," Twitter wrote. Musk is simply trying to escape a merger agreement that he "no longer found attractive once the stock market—and along with it, his massive personal wealth—declined in value," Twitter wrote.