Judge Orders Twitter to Provide More Spam Account Data to Elon Musk's Lawyers (cbsnews.com) 85
From the Washington Post earlier this week:
On Tuesday, The Washington Post reported that Twitter's former head of security, Peiter Zatko, had filed a whistleblower complaint with federal regulators, including the Securities and Exchange Commission, accusing Twitter of "Lying about Bots to Elon Musk...."
"Twitter executives have little or no personal incentive to accurately 'detect' or measure the prevalence of spam bots," the complaint alleges, adding "deliberate ignorance was the norm" among its executive team.
The same article notes that three people familiar with Twitter's spam-detection, processes said Twitter's "internal bot prevalence numbers" were almost always less than 5%. (And the article reminds readers that Musk himself had waived his right to perform "due diligence" prior to striking the deal.)
But here's that Tuesday article's most prescient sentence. "The judge has rejected Musk's requests for information from more than 20 company leaders — including Zatko — but the whistleblower claims could open the door for them to make further requests, legal experts said."
Sure enough, Friday night CBS News reported that the judge "ordered both Twitter and Tesla CEO Elon Musk to turn over more information to opposing lawyers..." Chancellor Kathaleen St. Jude McCormick on Thursday ordered Twitter to provide Musk's attorneys more data regarding the company's estimates that less than 5% of the accounts on its platform are fake.
The judge also rejected Musk's attempts to shield details about analyses he used in his attempt to terminate the deal. That work was done by data scientists who examined live-feed information from Twitter about public user accounts to test the company's daily-user counts....
The judge rejected more comprehensive data requests from Musk's attorneys as "absurdly broad," noting that a literal reading of the request would require Twitter to produce "trillions upon trillions of data points" reflecting all data collected on roughly 200 million accounts over three years. But McCormick did order Twitter to produce information on 9,000 accounts that were reviewed in connection with company's fourth-quarter audit, a data subset that has been described as a "historical snapshot."
McCormick also ordered Twitter to turn over documents regarding other metrics, regardless of whether they expressly address "monetizable daily active users," or mDAU. Musk's attorneys have suggested that a comparison of Twitter's mDAU with other metrics, such as "User Active Minutes," could support their theory that the company has fraudulently misled investors and securities regulators about the scope of activity on its platform.
"Twitter executives have little or no personal incentive to accurately 'detect' or measure the prevalence of spam bots," the complaint alleges, adding "deliberate ignorance was the norm" among its executive team.
The same article notes that three people familiar with Twitter's spam-detection, processes said Twitter's "internal bot prevalence numbers" were almost always less than 5%. (And the article reminds readers that Musk himself had waived his right to perform "due diligence" prior to striking the deal.)
But here's that Tuesday article's most prescient sentence. "The judge has rejected Musk's requests for information from more than 20 company leaders — including Zatko — but the whistleblower claims could open the door for them to make further requests, legal experts said."
Sure enough, Friday night CBS News reported that the judge "ordered both Twitter and Tesla CEO Elon Musk to turn over more information to opposing lawyers..." Chancellor Kathaleen St. Jude McCormick on Thursday ordered Twitter to provide Musk's attorneys more data regarding the company's estimates that less than 5% of the accounts on its platform are fake.
The judge also rejected Musk's attempts to shield details about analyses he used in his attempt to terminate the deal. That work was done by data scientists who examined live-feed information from Twitter about public user accounts to test the company's daily-user counts....
The judge rejected more comprehensive data requests from Musk's attorneys as "absurdly broad," noting that a literal reading of the request would require Twitter to produce "trillions upon trillions of data points" reflecting all data collected on roughly 200 million accounts over three years. But McCormick did order Twitter to produce information on 9,000 accounts that were reviewed in connection with company's fourth-quarter audit, a data subset that has been described as a "historical snapshot."
McCormick also ordered Twitter to turn over documents regarding other metrics, regardless of whether they expressly address "monetizable daily active users," or mDAU. Musk's attorneys have suggested that a comparison of Twitter's mDAU with other metrics, such as "User Active Minutes," could support their theory that the company has fraudulently misled investors and securities regulators about the scope of activity on its platform.
You can buy twitter followers... (Score:2, Interesting)
Re: (Score:3, Interesting)
Re: You can buy twitter followers... (Score:3)
Re: You can buy twitter followers... (Score:1)
Re: You can buy twitter followers... (Score:1)
As is all social media....except for youtube. Ok not kidding, yes I am, really.....
Re: (Score:2)
Re:You can buy twitter followers... (Score:5, Insightful)
None of that matters at all unless Twitter has documentation which acknowledges that, because the argument hinges on Twitter knowingly pretending they have less bots on their service than they do. If there are just more bots than they know about, that's Musky's problem, because he skipped due diligence. THAT was the time to determine whether Twitter had a clue about how many bots were on the service. Now he has to prove that they knew they had more than they were saying, because he thought he was playing four dimensional chess and that skipping due diligence would be a master stroke of something other than his penis. He clearly thought that avoiding due diligence would mean that he could utilize the argument that Twitter was wrong and therefore he shouldn't have to pay, but finding out whether Twitter was competent to report their own numbers is literally part of the due diligence that he skipped.
They call it due diligence because it is due. Elon didn't do the due and now he's screwed, unless he can find a smoking gun. And the judge is not letting him drag net Twitter's documentation to find it, he's got to pick a spot if he wants to fish. If he'd done due diligence, he might have found something to poke at, but again, he didn't. Let's hear again about how Elon is the most competent genius around, I could use a chuckle.
Re: (Score:1, Interesting)
Now he has to prove that they knew they had more than they were saying, because he thought he was playing four dimensional chess and that skipping due diligence would be a master stroke of something other than his penis. He clearly thought that avoiding due diligence would mean that he could utilize the argument that Twitter was wrong and therefore he shouldn't have to pay, but finding out whether Twitter was competent to report their own numbers is literally part of the due diligence that he skipped.
Elon Musk skipped due diligence because he knew that he didn't need it. He knew that he could do exactly what he is doing right now.
That's one of the benefits of being rich. He can literally do anything he wants, and if there is a problem he can just keep hiring more lawyers until the problem goes away. And if the lawyers can't make the problem go away he will offer to "settle" for what amounts to pocket change.
Elon Musk is the real life version of the old comic book character Richie Rich. A 12 ye
Re: (Score:2)
What's he's doing now is PR, and PR doesn't mean much in court. It's the same that Trump is doing, getting the public on your side when you know you've got a poor legal leg to stand on. Why rich people do the PR instead of getting decent legal advice upfront is beyond me.
Re: (Score:3)
Elon Musk skipped due diligence because he knew that he didn't need it. He knew that he could do exactly what he is doing right now.
Wouldn't that be fraud? I'm not a lawyer, but it has to be some sort of tort. You're supposed to enter into contracts in good faith.
Re: (Score:3)
Re: (Score:2)
Re: (Score:2)
Twitter's house of cards will soon collapse.
This is such a dumb thing to say, why would it? Everybody knows advertising is a scam, but everyone still does it, because they are unwilling to compete on quality. It's hard to justify when most people are happy to buy shit.
Re: (Score:2)
Re: (Score:2)
Doesn't matter. Musk made a decision to buy, and agreed to, without doing due diligence. He's gotta work harder than "I think there are spambots" to get out of it. Buyer beware.
Re: (Score:3, Insightful)
Is this the first opportunity we've had to kick over the rocks of "social media" and see how fake it all really is? I think so, and we have Elon to thank for it.
Re: (Score:2)
Is this the first opportunity we've had to kick over the rocks of "social media" and see how fake it all really is? I think so, and we have Elon to thank for it.
Smart people have always known this, and Musk has not made stupid people any more smart.
Re: (Score:3)
I think neither - he knew (more or less) going in, but didn't care since it doesn't matter, and backed out simply because the stock market went down so the company isn't worth what he offered any more.
Knowing the numbers on spam and bots doesn't change anything really. Everbody who uses the site knows how it is, or isn't, including both users and advertisers whose bottom line is and always was conve
Re: Cooking the books (Score:5, Insightful)
Yeah, it does matter. If the site's actual number of eyeballs is lower than what it claims it is:
1) The company is overvalued, mainly because:
2) Advertisers are paying too much
3) When they find out, the revenue shrinks
4) When the revenue shrinks, the share price tanks
5) Shareholders will demand an explanation, and they have the right to one
Musk or no Musk, they've got some splainin' to do to their current shareholders if their actual eyeballs are less than what they're leading the public to believe. Your argument is no different from saying that people know Enron is a big energy company and that shareholders should just take their word for it that the money they're making is both real and sustainable in the long term. The only difference in this case is that it's not accounting fraud, rather it's something more resembling click fraud.
There's a real possibility that they want to avoid having to face that music, so they're depending on the Musk buyout to happen. Who knows at this point.
Re: Cooking the books (Score:1)
Re: (Score:2)
Re:Cooking the books (Score:5, Interesting)
> It looks to me like Twitter was cooking the books Enron-style for years before Elon got involved.
How so? Had Twitter made a secret of the fact that the vast majority of accounts are effectively loss-leaders, the cost of business? Or the fact that they struggle daily against automated spam factories?
> The only question is if Elon knew this going in and planned for this outcome...
No, that's not the question. Musk went in with bravado, and then had a change of heart. Twitter told him how they counted accounts that they might generate revenue from, and Musk discovered that the spam accounts weren't among them. ... but the contracts that he signed weren't conditional on "spam/bot accounts total" but instead on "spam/bot accounts mistakenly counted as possible revenue sources" (mDAU).
And now, he's generating a circus, hoping everyone will forget about what the contract was actually about.
> Regardless, win-win for anyone that thinks it is too dangerous to have Twitter insiders decide for the rest of us what is true..
Why would you think that? Twitter hasn't changed its moderation strategies in response to the contract or the lawsuit, which I presume is what you're trying to say there. While there might be some overlap, Spam != Bots != disinfo/misinfo. Perhaps you are thinking that such people will now know a tiny bit more about Twitter's procedures? IE that while they have automated moderation, they have a human spot-check behind it?
Re: (Score:2, Insightful)
>Why would you think that?
My hope that as the body of evidence that Twitter does not represent consensus or even popular opinion grows we will have more institutions learning to ignore Twitter storms. As far as influence goes, the power of Twitter is way disproportionate to actual representation. So you have a loud minority propped (and often egged on) by bots pre
Re: (Score:2)
I don't know what the fuck you're thinking with this whole 'cooking the books' thing. This is a pretty damn clear case of someone trying to weasel their way out of writing some checks with their ego that they didn't actually think they'd have to cash, and the orders from the judge reflect they seem to agree.
Ultimately, I think the destruction of Twitter is good for humanity, so I'd lov
Re: Cooking the books (Score:2)
I'm pretty sure they never disclosed that multiple employees were actively working for other corporations and nation-states.
Re: (Score:2)
That's derpy as fuck.
Yes, twitter had disclosed that they cooperate with various countries around the world. In this case, India.
Re: (Score:2)
And now, he's generating a circus, hoping everyone will forget about what the contract was actually about.
The media and public may forget things like that, but there's a reason court cases take months and judgements run between the many 10s to the 100s of pages for even simple issue. The legal system does not forget.
Twitter data (Score:5, Informative)
I've worked at a mid sized social media company and seen first hand how that company and our other social media partners (not Twitter but other big names you know) track, handle and create user meta data.
The odds that Twitter has a handle on this are very low. Not necessarily because they're maliciously hiding it. I don't know either way on that. But, because this is an incredibly difficult problem. Looking at any random account you simply can't know for sure in many cases as most people aren't super active and they just resend the same links and memes to their friends the same way a bot would.
Doing the same analysis in bulk with code? Good luck getting an accurate number of humans vs bots if the bots are putting in any reasonable effort to hide and not do dumb things like post hundreds of times a minute.
Re:Twitter data (Score:5, Insightful)
Re: (Score:2)
Musk seems really ignorant when it comes to technology.
This is what blows my fucking mind.
If whatever methodology he comes up with for determining if something is a bot is as bad as his statistical analysis methodology, it's a bad fucking joke.
Re: (Score:2)
Twitter claims in its court filings that the tools Musk is using claims Musk's twitter account is a bot!
https://arstechnica.com/tech-p... [arstechnica.com]
Re: (Score:2)
Judging from hearing him talk, and his personal accomplishments, I wouldn't rate him much more than slightly above average intelligence.
Wish the dude would stick to being a nerdy entrepreneur. He seems pre
Re: (Score:3)
The interesting part here is that if Twitter can identify spam to such a degree that they exactly know what is a spam account or not they would probably earn more money selling that technical solution than allowing spam accounts inflating mDAU revenue.
Especially when you don't want to (Score:2)
It is a non-trivial problem.
Note, however, you need not accurately identify each specific bot account in order to get a reasonable count. Sampling works. It's simple multiplication if you do human and in-depth automated analysis of 10,000 accounts, and although you're wrong about some you're wrong roughly equally in each direction.
That is, if you wrongly label a few bots that are in fact human and vice versa, that doesn't change your count. Your count changes significantly only if you systematically and sig
Re: (Score:2)
It is a non-trivial problem.
Note, however, you need not accurately identify each specific bot account in order to get a reasonable count. Sampling works. It's simple multiplication if you do human and in-depth automated analysis of 10,000 accounts, and although you're wrong about some you're wrong roughly equally in each direction.
That is, if you wrongly label a few bots that are in fact human and vice versa, that doesn't change your count. Your count changes significantly only if you systematically and significantly err in a chosen direction.
However, the fact that it's not stupidly easy comes in handy when Twitter execs don't WANT to tell advertisers "20% of the clicks you think you're getting on Twitter are actually fake".
But how do you make sure that in-depth analysis isn't biased? One number I'd be curious to see is definitely human vs definitely bot vs unsure.
It would be certainly be more informative than the 5 vs 20% numbers that have been thrown around.
Re: (Score:1)
Hey, don't get me wrong, I'm not saying if Twitter is trying or not. I don't use Twitter and have no idea.
I'm just saying -if- they were honestly trying to do proper analysis it is a hard problem.
Given your experience though, no, I'd agree they're not trying if that's how it really works. I'd assume their bot levels are much higher than 5% if they're not trying.
Re: (Score:2)
Proceeds to tell story about Twitter trying.
I work for an organization that has a lot of customers. Nowhere near Twitter's user count, of course, but large enough to know that the proposed mitigation strategies:
Blocking your IP, your username/e-mail pattern, and your hours of use are absurdly fucking ignorant.
Even your custom domain is pushing it.
No, when we try to solve problems like bots hitting our services, we're not looking at the things that identify you as an obviou
Due Diligence (Score:3)
Re: (Score:1)
Re: (Score:2)
The only question is how many past CFOs and CEOs going to go to jail in addition to the current crew.
CFO and CEO going to jail for lying about bots? https://a.pinatafarm.com/620x4... [pinatafarm.com]
The Wolf Of Wallstreet guy defrauded hundreds of millions and got two years in minimum security with Tommy Chong as a cell mate. You really have some anger issues over twitter.
Re: (Score:3)
Re: (Score:2)
Securities fraud is still nominally illegal, so yes jail.
Are we talking about Twitter or Musk? Cuz really that statement could go either way.
Re: (Score:2)
Re: (Score:2)
So no, not jail. Because that's not how that shit happens.
Now Musk, he's treading very close to securities fraud, and if he's not careful, he may end up with a company he didn't want, and a prohibition against running it.
Re: (Score:2)
Re: (Score:1)
Re: (Score:2)
Waiving due diligence means that Musk has to prove malice, because incompetence would have been revealed during the due diligence he didn't do.
It means that unless the guys at Twitter are a lot less competent than I'd bet they are, ol' Musky is screwed. He's got to prove they know they're wrong.
Re: Due Diligence (Score:2)
Re: (Score:1)
Re: Due Diligence (Score:2)
Very highly unlikely. If Twitter was maliciously misrepresenting what they already knew... Elon would be the least of their problems. Elon is a potential owner, there already are heavy institutional owners of Twitter (60%).
If Twitter was doing something this illegal, the SEC will ruin the Executives lives. The current owners will burn it to the ground in additional lawsuits. Their Public Auditor would lose credibility.
You think heavy weights like Vanguard, Morgan, Wells, etc would let this stuff fly?
Re: (Score:2)
Re: Due Diligence (Score:1)
Re: (Score:2)
Proving fabrication is quite a high bar for a problem as complex as identifying which of the 1.3billion twitter accounts are bots. The defense against fabrication becomes easier the more complex the problem.
You really need us to say it? (Score:2)
Fine, it's all bots. Every last account. All the way down.
Isn't it standard discovery? (Score:2)
I am clearly not a lawyer. But doesn't this look like standard discovery?
If you are going to use a claim in court, you need to provide the basics evidence of your claim to the opposing side. And you are making a wild claim, you can not request the opposing side to provide you with an absurd amount of data in hope to find the needle in the haystack.
Re: (Score:2)
Yes, I think so. But ISTM that Twitter is more likely to lose from the discovery than Musk.
I'd hope they could both lose, but that would mean only the lawyers win.
Re: (Score:2)
Discovery could mean that Musk doesn't even get a trial, he gets a summary judgement that yes, he signed a contract requiring specific performance and yes, he waived due diligence and bought it as-is.
He's required to share the data analysis that caused him to back out of the deal. That's highly unlikely to show that he had a valid reason, as per the contract.
All the nonsense about it being bad for twitter requires some vast conspiracy to exist, and twitter to be completely inept at said conspiracy, and for
Re: (Score:2)
They're a corporation that sells advertising; do they actually have any need for the purported conspiracies? They make the amount the of advertising money they claim to make, right? That's all that matters when somebody is buying the company as-is; the money. Nothing else is material.
I think that sums it up very well.
Re: Isn't it standard discovery? (Score:1)
It's NOT so much the money a biz has made in the past BUT the money it will make in the future that drives any deal.
If your future projections were just wrong its fine, but if they were fraudulent then you do end up in jail.
More importantly, if the money they have been making was based on "knowingly" reporting wrong metrics to advertisers, then it means it's not going to continue at same level since now advertiser's know you were misrepresenting.
Whichever way you look at it Twitter as we know it is dead and
Huh? (Score:2)
Why? Why? Why? Why can't I say whatever nonsense I want on a media channel I don't own?!
Re: (Score:1)
Twiter is full of trolls, bots, and spammers! Who knew!?
Elon Musk [twitter.com] knew.
Elon Musk
@elonmusk
If our twitter bid succeeds, we will defeat the spam bots or die trying!
11:53 AM Apr 21, 2022Twitter for iPhone
More realistic reading over at Techdirt! (Score:2)
Musk seems like a very smart guy (Score:2)
... who keeps doing or saying very dumb things, then often doubling down on them. Is it just that he has no impulse control, is it because he craves attention and adulation, or is it something else?
Re: (Score:2)
The dude is obviously smart, but if you were to analyze his behavior, you'd think he was a teenager. That intelligence doesn't do him one bit of good when he's led around by his ego. I'm pretty sure he's the kind of dude who would save the last 20k of the fortune he burned to win an argument just to whack the guy who beat him.
Delay tactics (Score:2)
The whole thing seems to be an attempt to create a justification for a continuance (the judge has set a trial date much sooner than Musk's team wants) by creating a situation where they can claim:
"We cannot possibly sift thru all of the data we received in discovery in time for the currently scheduled trial date, so lets push it back."
so what? (Score:1)
85% (Score:2)
85% are just posting to create awareness and sell their products, brand of whatever.
Take it with a grain of salt (Score:1)
Re: (Score:2)
But I know half a dozen >40 people who buy into this nonsense hook, line and sinker.
No one wants the deal (Score:2)
Musk wants out and Twitter's board doesn't want him to buy and ruin Twitter. So why not just shake hands and everyone walks away? I guess the lawyers need to be fed.