Live data from Hacker News

Twitter says Musk’s spam analysis used tool that called his own account a bot

arstechnica.com

551–560 of 585 posts

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#551

Earlier quoted context omitted.

> If twitter bot number is, say, 30% Can you define with specificity what the “twitter bot number” is? Because Twitter doesn’t claim that 5% or fewer of accounts are bots. They claim that “according to their ongoing reviews and subjective numbers” less than 5% of accounts they qualify as an monetizeable user are “false or spam”. Twitter could be 90% bots and spam, and not be even slightly inaccurate in their statemen…

> They claim that “according to their ongoing reviews and subjective numbers” less than 5% of accounts they qualify as an monetizeable user are “false or spam”. Yes, it's just a short version for the sake of the discussion. Let's just say this number is not 5%. It's 30% when evaluated by an independent third party. Other comments claim that this number can be 90% 80% and it wouldn't matter. Twitter would be relieved…

> Other comments claim that this number can be 90% 80% and it wouldn't matter.

I don't see any other comments saying the number can be 90% or 80% (other than mine, where I was pointing out the difference between before removal and after removal). Can you cite Which comments specifically are you referring to and whihc part of them you're disagreeing with? I'm just not seeing that.

I mean, it's fine for you to disagree with _that_, but it seems like you're disagreeing with a straw-man.

> You'd imagine it is impossible, but it's not

I wouldn't imagine it would be impossible. I think it's unlikely, but it's always possible that Twitter is perpetrating a deliberate fraud in its SEC filings. My priors on that are pretty small (1-5%?), but sure, of course it's possible.

If Twitter is actually perpetrating a deliberate fraud, and Elon is able to get the goods in discovery, then sure he'll likely win the case. That's a really high bar, though, and does not square with your prior argument that if a third-party analysis comes to a 30% bot number, then the SEC will bring charges.

That's simply not sufficient to cause the SEC to bring charges. Rather, you would need the additional evidence you think might come out in discovery.

My point is this: "The true bot number is 30%" and "The true bot number is 30%, and Twitter knew that and deliberately lied in its SEC filings" are two different statements. One is sufficient to bring a charge, and the other is not.

> You can bet there will be at least one person pointing out in the email/doc that "hey, our approach of evaluating numbers doesn't look right, and here's why....". Then, that person has been ignored.

Is is possible? Sure, of course. Would I put a bet on that exact scenario happening? Probably not.

Seems just as likely that someone sends the email that "our approach doesn't look right, and here's why...", and someone engages and responds with "I disagree, that methodology does X, and we chose to do Y because Z, but let's talk about it". Or, I could also see absolutely nothing internal on this because it's a box-ticking exercise as part of their SEC filings and nobody internal really cares.

> Coming up with a number like this will inevitably have multiple valid approaches.

The problem is that its not enough to show that someone else has a different valid approach. The need in court (and for the SEC) would be to show that Twitter's approach is not valid. It's really not enough to show that Twitter could have made different choices and got a different outcome. You need to show that Twitter's approach and outcome was unreasonable.

Anyway, my fundamental disagreement with your statement is that a "bot rate" of 30% automatically implies fraud and an SEC charge. I don't think it's that simple.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#552
post #538
post #505

Earlier quoted context omitted.

Twitter seems pretty confident about revealing their process to calculate that figure: They’ve described it multiple times and been explicit in their SEC filings. As they should have been, as it’s a sound enough process given constraints and doesn’t need to be 100% accurate. They have nothing to worry about.

I've been downvoted but what you state is exactly right: they have to decide exactly how confident they are in their own process holding up in court (which isn't exactly the same thing as it being accurate) - and to whatever degree they are not confident, back it out to a price change to de-risk the outcome I mentioned. I agree with you that they seem confident, but the entire back and forth around mDAU originated wi…

They’re going to court, because Twitter holds all the cards here and have the strongest legal case, so it’s in their best interests to have this adjudicated by a Delaware court.

Musk is trying to pretend the mDAU = number of active bots. It’s dishonest, and likely designed to hurt the company and tank its share price and therefore bully it into conceding. Given the strength of their case I don’t see that happening.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#553
post #64

Earlier quoted context omitted.

"He signed away any and all rights to due diligence in the agreement." Why do more fraudsters not use contracts If people can sign away due diligence, sign up for a Diamond, get a Lemon, and have no recourse, surely scammers would be doing this instead of structuring their scam in a way that has the potential of sending them to jail

You don’t deserve the downvotes here. Agreeing to buy Twitter as-is does not absolve them if they’re found to have misrepresented what they’re selling which is what the legal angle they’re gunning for is. I don’t buy it but it at least logically follows.

That’s not what Musk is claiming though, when you cut through the noise. He has zero actual evidence showing that Twitter made false misrepresentations. His entire argument really centred around the argument of “Twitter has lots of bots” which is orthogonal and immaterial to the question of mDAUs.

Twitter could be 75% bots and its SEC filing claim could still be entirely above board if it was able to filter them out.

Musk’s opportunity to verify that statement was during due diligence. He chose not to, and can’t legally try to perform due diligence after the fact because he has cold feet.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#554
post #150

Earlier quoted context omitted.

He can be popular due to Twitter and people can also not be deranged morons. Twitter is very popular with a few niche groups: Journalists and people in tech for example. The former is influenced by Elon's antics on Twitter and end up writing news stories or making documentaries [1] about him. People learn about Musk through these respectable publications. The latter group spreads the gospel of Elon's TED talks and pr…

Except that Musk has talked about how he had to work his way through college and basically didn’t benefit in any way from his dad. https://twitter.com/elonmusk/status/1211054942192119808?s=21...

While I agree that the emerald mine stuff is overblown, I don’t buy the idea that Musk was self made and independent. He didn’t always have free capital, but he seems to have a solid network through his extended family from early on that helped him.

That meant he had a readily available set of investors, advisors, and partners. That ranged from Greg Kouri, pivotal in forming Zip2 and X.com, to Roelof Botha. Roelof is the grandson of Pik Botha, the last apartheid South African foreign minister who had built up extensive connections amongst the US elite.

Success is so much easier when you go to knock on doors and find them already opened for you.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#555
post #510

Earlier quoted context omitted.

I think you and onos are not quite making the same point. Conservative voices are a bit suppressed in some places, like on Campus. Some don't like to accept that, but it's true. But it's a kind of Social Justice institutionalism that would be the issue there. Otherwise, 'classical conservatives' who voice their opinion against Trump, are also ostracized by that far right gang, which is very true. So that's a differen…

>Conservative voices are a bit suppressed in some places, like on Campus. I think it's a bit misleading though. Who would goto college and listen to somebody repeat ad nauseum their parent's conservative talking points they've already heard their entire lifes? Or listen to someone even more radically conservative than their parents? Also a lot of those "suppressed" people were just openly bigoted and claiming to feel…

Well you've proved my point for me.

Your assumption that 'conservative' voices lack in legitimacy and are merely repetitious of their parents views, and worse, that they are trying to 'persecute' people is itself obviously bigoted, and gives us a crystal clear example of the problem.

Even views such as 'people ought to be treated for who they are and not their intersecting identities' is oddly 'conservative' in 2022. Or 'Trans women are different than women' - some argue is a transphobic/conservative position. etc.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#556
post #543

Earlier quoted context omitted.

If you think his ratio of good to bad business decisions is even average then you haven't been paying attention. Tesla has nearly gone bankrupt repeatedly, and it's stock price has largely risen the most after Musk was removed from controlling the company as a result of a series of tweets that were considered illegal attempts at manipulating the stock market. In point of fact, Musk still controls SpaceX and internal…

You are basically making a claim that he is a bad businessman without mentioning that he is one of the richest people in the world. Why should anyone take your arguments seriously?

He signed a binding acquisition agreement in a $44 billion dollar deal without doing any due diligence. Bigger deals have been struck on similar timeframes, but rarely has the purchaser evidenced such obvious buyer’s remorse so quickly afterward, or embarked on the venture with as little thought as he put toward it.

That’s a pretty weighty fact on the scale of “Musk - canny businessperson, or remarkable streak of being in the right place at the right time.”

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#557

Earlier quoted context omitted.

Apparently Twitter subpoenaed Musk's friends and colleagues who were actively tweeting about the deal with what is believed to be an attempt to prove he encouraged his posse tweet negatively to force a lower price.

David Sachs, prominent VC was one of those subpoenaed and he definitely seemed a little sensitive about it: https://twitter.com/DavidSacks/status/1554258141876932608

My heart bleeds for him. He’s achieved tremendous success in a remarkably successfully economy based on dynamism and the rule of law, and now he’s asked to do his part, under oath, to maintain that system. Doesn’t he know that part is for chimps?*

*Tried to write “chumps,” but liked the typo better.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#558
post #417

Earlier quoted context omitted.

It's insane because Twitter has been very careful to explain that the 5% number is not some sort of "promise". > Additionally, our calculation of mDAU is not based on any standardized industry methodology and is not necessarily calculated in the same manner or comparable to similarly titled measures presented by other companies. Similarly, our measures of mDAU growth and engagement may differ from estimates published…

> They covered themselves very clearly, I am shocked, just shocked that a company with six billion dollars of cash on hand is able to afford good enough lawyers who make their SEC reports ironclad against shareholder lawsuits. Who would've thought.

Best part is that if Musk loses and the court enforces specific performance, he just bought a company that just paid for all those lawyers to fight him out of the company treasury!

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#559

Earlier quoted context omitted.

All that matters, against musk, is that however twitter claimed to count it's users was infact, how they counted. Musk wanted to invent some means to count users in an effort to demonstrate fraud. But that's not fraud. Just because musk provides an alternate means to count users, doesn't demonstrate fraud. Full stop. Unless he proves that they didn't count the way they claim, he has not claim. Waiving he due diligenc…

I was not defending Musk, just trying to explain what he might be “thinking”. I see you’re point though. I guess he’d have to prove they used a formula that they knew was wrong for it to be fraud?

> I guess he’d have to prove they used a formula that they knew was wrong for it to be fraud?

They could know the formula was wrong - inherently every measurement is wrong to some extent which is why you see a +/- error, that is not fraud.

For it to be fraud, they would need to measure it in the way they claim to, get a result they dont like, and report it as <5% anyway (with intent to defraud people). Which could happen in theory, but seems highly unlikely.

Re: Twitter says Musk’s spam analysis used tool that called his own account a bot

#560
post #476
post #451

Earlier quoted context omitted.

Quoted post unavailable.

You kinda can, though. https://www.scienceabc.com/innovation/why-do-hard-drivesusb-...

Those are reasonable.

If said I count GB as 512MB, the court would say “nobody else does that, you’re just trying to defraud people, your disclaimer is not in good faith”.

Post reply on HN