News for hackers. Stuff that matters.
Re:Twitter says Musk’s spam analysis used tool that called h... by nomel on Friday August 05, 2022 @02:26PM (#32359943)
> I still can't get over the banana-pants insanity of the first count

As someone who's unfamiliar with this, could you explain what's banana-pants insane about a company inflating their hard-to-discover numbers, for profit? I thought that was, unfortunately, banana-pants standard? For example, the fiction that is Twitter or Reddit's user count.

Re:Twitter says Musk’s spam analysis used tool that called h... by phailhaus on Friday August 05, 2022 @02:54PM (#32360343)
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 by third parties or from similarly titled metrics of our competitors due to differences in methodology.

They covered themselves very clearly, and Musk has no standing here. He knew this, and decided to waive his due diligence and buy anyways.

[1] https://www.sec.gov/Archives/edgar/data/1418091/000141809120...

Re:Twitter says Musk’s spam analysis used tool that called h... by 8ytecoder on Friday August 05, 2022 @03:10PM (#32360577)
He claimed to buy Twitter, specifically, to clean it up; to get rid of bots. That was his stated goal.

"If our twitter bid succeeds, we will defeat the spam bots or die trying!"

https://twitter.com/elonmusk/status/1517215066550116354

Re:Twitter says Musk’s spam analysis used tool that called h... by kennxfl on Friday August 05, 2022 @04:10PM (#32361402)
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.
Re:Twitter says Musk’s spam analysis used tool that called h... by threeseed on Friday August 05, 2022 @05:17PM (#32362074)
David Sachs, prominent VC was one of those subpoenaed and he definitely seemed a little sensitive about it:

https://twitter.com/DavidSacks/status/1554258141876932608

Re:Twitter says Musk’s spam analysis used tool that called h... by nocoiner on Sunday August 07, 2022 @03:12AM (#32374542)
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 h... by madeofpalk on Friday August 05, 2022 @06:31PM (#32362729)
The weird part is that "twitter has lots of spam bots" is a completely independent claim from "5% of the users who use our website or 1st party apps are bots".
Re:Twitter says Musk’s spam analysis used tool that called h... by swores on Saturday August 06, 2022 @03:34AM (#32365862)
While your comment is technically true standing alone, you've also slightly misunderstood what the 5% figure is about.

It's not "users who use our website or 1st party apps", it's "monetizable daily active users", which may be similar to the group of users you described, but most importantly Twitter can choose not to show any ads (or not to charge money for ads shown) to bot accounts.

So even if Twitter only had 1 single non-bot user, as long as that's the only user they are "monetizing" they would be at 0% of mDAUs being bots.

Therefore, as jcranmer commented elsewhere in this thread, the 5% is actually their estimate of their own false negative rate of detecting bots (i.e. deciding which accounts can be monetised vs not).

Re:Twitter says Musk’s spam analysis used tool that called h... by rurp on Friday August 05, 2022 @07:35PM (#32363197)
Yep, which I'm sure Elon and his entire crew of bankers and lawyers understand quite well. Musk is lying and sowing confusion in order to mislead people.
Re:Twitter says Musk’s spam analysis used tool that called h... by chx on Friday August 05, 2022 @11:04PM (#32364429)
> 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.

Re:Twitter says Musk’s spam analysis used tool that called h... by nocoiner on Sunday August 07, 2022 @03:14AM (#32374552)
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 h... by the_duke on Friday August 05, 2022 @07:03PM (#32362932)
IANAL, but "we use our own metric" doesn't sound like an excuse for publishing numbers to shareholders that aren't a best effort estimation based on reasonable methods.

If the methods were reasonable is of course possible, but up for debate and at least questionable.

Re:Twitter says Musk’s spam analysis used tool that called h... by bialpio on Friday August 05, 2022 @07:36PM (#32363210)
IANAL, but I'm not sure if it matters, as long as the methodology is published? "Here's the number, here's how we calculate that number, don't buy shares if you think our methodology sucks" sounds reasonable to me, especially if they don't have a different number internally that they use to drive the decisions?
Re:Twitter says Musk’s spam analysis used tool that called h... by nonethewiser on Friday August 05, 2022 @03:15PM (#32360648)
That's not what the original poster implied. He implied it's crazy because of the duration and specific purpose of the supposed lie.
Re:Twitter says Musk’s spam analysis used tool that called h... by pavon on Friday August 05, 2022 @04:38PM (#32361709)
Doesn't seem strange to me. If I am suing a company for false advertising I'll claim that their behavior was aimed at inducing me to purchase the product at an inflated price. Not because I think they were specifically conspiring to target those lies directly at me personally, but because I am the one that is a party to the lawsuit. If Twitter was lying to inflate the stock price in general, then they were inducing everyone to buy at an inflated price and Musk is one of those people.
Re:Twitter says Musk’s spam analysis used tool that called h... by bostonsre on Friday August 05, 2022 @03:43PM (#32361002)
Twitter added a loop hole saying that their data might be incorrect. It seems disingenuous to absolve them of all sins if their data is hugely incorrect to the point where it's fraudulent just because they added that loop hole.
Re:Twitter says Musk’s spam analysis used tool that called h... by zinekeller on Friday August 05, 2022 @03:50PM (#32361100)
> It seems disingenuous to absolve them of all sins if their data is hugely incorrect to the point where it's fraudulent just because they added that loop hole.

If there's an actual industry standard then you could argue that Twitter is lying. However as everyone in the legal community knows there are two important concepts here: wilfulness and negligence. Did Twitter willfully do this or is this just because there isn't an actual standard methodology but tried to do their best-effort anyways? Is Twitter aware of problems in its methodology that can be solved but were left as-is? Conversely, did an external party intentionally misinterpreted those numbers so that that party can paint a different picture that what Twitter actually claims?

Re:Twitter says Musk’s spam analysis used tool that called h... by bostonsre on Friday August 05, 2022 @06:58PM (#32362901)
Yea, makes sense. Seems like it will be hard to prove malice over incompetence and vice versa.
Re:Twitter says Musk’s spam analysis used tool that called h... by jacquesm on Friday August 05, 2022 @04:38PM (#32361708)
Maybe, but that's just not how it normally works and Musks reality distortion field does probably not include the courts. You sign a terms sheet conditional on DD and only afterwards do you move to put in a binding offer. Doing it in a different order is a waste of effort, after all, what's the point of doing DD after you've already made a binding offer? You can't use it to change the deal parameters and you can't use it to break it up.
Re:Twitter says Musk’s spam analysis used tool that called h... by phailhaus on Friday August 05, 2022 @04:14PM (#32361451)
There's no standard or foolproof way to find bots, otherwise we'd be able to just ban them all. It all requires judgement. Twitter's statement is acknowledging this fact and legally covering themselves from people just like Musk who might go "well I have my own calculation metric and it says there's more."
[dead]
Re:Twitter says Musk’s spam analysis used tool that called h... by slavak on Saturday August 06, 2022 @02:37AM (#32365539)
Re:Twitter says Musk’s spam analysis used tool that called h... by refurb on Sunday August 07, 2022 @06:06AM (#32375271)
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”.

Re:Twitter says Musk’s spam analysis used tool that called h... by loceng on Friday August 05, 2022 @05:00PM (#32361923)
I imagine the argument is whether the method Twitter used, and was seemingly acceptable by SEC, is ethically, morally, or logically a reasonable measure.

E.g. Can they 100% say that there are less than 5% bot accounts, or is that 5% number being pulled out of their ass - and saying "less than 20% of accounts are bots" would hold as much water? That is therefore then a misrepresentation, fraud IMHO - and stockholders should be suing not only Twitter but also the SEC.

Re:Twitter says Musk’s spam analysis used tool that called h... by Sebb767 on Friday August 05, 2022 @09:43PM (#32364012)
You're missing that they never promised that there are only 5% bots. They said they estimate five percent based on their internal research, which is completely true as far as we're currently aware. Unless you can prove they intentionally faked their own internal investigation, you'll have a very hard time proving fraud.

Also, then there is the fact that Elon apparently even didn't believe this number. So even if he proves it's faked and the court agrees that this is actually a major dealbreaker (which is already near impossible), Twitter could still argue that he was completely aware of the bot situation when he made the offer.

Re:Twitter says Musk’s spam analysis used tool that called h... by loceng on Saturday August 06, 2022 @03:31AM (#32365849)
So if you have a home inspector inspect a house before you buy it, and they say "yeah, it looks good" - but it turns out they didn't actually investigate in a way that arguably involves an adequate amount of integrity or in a reasonable way to determine as accurately as possible that everything is sound - then you don't think that home inspector isn't liable? And what if the professional organization who has the responsibility of regulatory oversight of that profession, what if they're been lacking and not being responsible as they are supposed to be, allowing for poor analysis as the standard - are they guilt-free too?
Re:Twitter says Musk’s spam analysis used tool that called h... by JumpCrisscross on Saturday August 06, 2022 @04:14AM (#32366078)
> if you have a home inspector inspect a house before you buy it, and they say "yeah, it looks good" - but it turns out they didn't actually investigate

Home inspectors follow well-established practices. Bot estimation for Twitter has no such analog. The closest might be the seller saying they've seen rats but never looked into it. You say fine, I'm buying to clean it up anyway; waive inspection rights; commit and then discover a rat's nest.

Unless you can show the seller knew about the problem and lied about it, you're going to get your case thrown out of court. Now add multimillion dollar teams of lawyers and bankers advising each side prior to entering into the transaction and we have an approximation of the stupidity of this case.

Re:Twitter says Musk’s spam analysis used tool that called h... by loceng on Saturday August 06, 2022 @04:36AM (#32366187)
"Bot estimation for Twitter has no such analog."

And that's acceptable there isn't well-established/widely accepted practices?

Certainly Twitter and SEC will argue that to protect themselves from scrutiny and legal issues they likely deserve.

Re:Twitter says Musk’s spam analysis used tool that called h... by _djo_ on Saturday August 06, 2022 @07:20AM (#32367060)
Yes it is, because Twitter has no close analogues. There’s nothing to base an industry-wide standard on.

All company figures involve some level of trust that they’re based on correct processes as signed off by audit firms. For areas of material interest you’re supposed to do due diligence before signing final offers, but Musk explicitly waived that for some bizarre reason.

Re:Twitter says Musk’s spam analysis used tool that called h... by loceng on Sunday August 07, 2022 @11:04AM (#32377035)
If you can't with 100% certainty determine the number of real users you have you shouldn't be putting out any number, and second to that, you have to be open to challenges that "your"/their analysis isn't reasonable; from what it sounds like, based on what Elon had said or alluded to, the way Twitter determined bot % was they had staff manually and "randomly" select literally just 100 accounts, and based on their own observations, decided if accounts were bots or not. If that's true, the obvious problem with that, is bots are designed to mimic to blend in and look like humans - to have the same or similar behaviour patterns to not be outside an expected range of behaviours.

So, what might really be going on is: Twitter's able to detect that less than 5% of users are unsophisticated bots, but they're unable to determine how much of the remaining 95% are sophisticated bots that accurately enough mimic human behaviour; and so it's not possible to genuinely say "less than 5% are bots."

Re:Twitter says Musk’s spam analysis used tool that called h... by _djo_ on Sunday August 07, 2022 @11:17AM (#32377134)
You have not done enough work to understand this issue before picking a side, and it shows.

For one, this is not about what percentage of Twitter users are bots.

Nor have you accurately or fairly described Twitter’s methodology for verifying their mDAU figures.

Re:Twitter says Musk’s spam analysis used tool that called h... by vba616 on Saturday August 06, 2022 @01:11AM (#32365011)
I think your intuition that the SEC passes judgement on reasonableness of made up metrics is...doubtful.
Re:Twitter says Musk’s spam analysis used tool that called h... by jcranmer on Friday August 05, 2022 @02:35PM (#32360069)
Twitter is a public company (and has been for 9 years), and has maintained the 5% mDAU disclaimer in its required annual reports for years. I don't know off-hand how long it's been doing so, but probably at least 4 or 5 years.

The allegation is saying not only that Twitter has been lying about that number--for years, on documents that land it in legal hot water if it's been lying--but that it was doing so specifically so that Musk would buy Twitter at an inflated price. The insane part is really that specific intent, not the lying about the numbers.

Re:Twitter says Musk’s spam analysis used tool that called h... by flerchin on Friday August 05, 2022 @02:38PM (#32360115)
Well twitter would have been obscuring bots to obtain an inflated purchase price _by anyone_.
Re:Twitter says Musk’s spam analysis used tool that called h... by thr0wawayf00 on Friday August 05, 2022 @02:39PM (#32360145)
FWIW, academic research has questioned the 5% number for years. USC and IU put out a report 5 years ago that believed the upper bound for fake accounts was 15%[0]. This isn't a new allegation.

0: https://www.cnbc.com/2017/03/10/nearly-48-million-twitter-ac...

Re:Twitter says Musk’s spam analysis used tool that called h... by kergonath on Friday August 05, 2022 @02:45PM (#32360228)
From your source:

> our estimates suggest that between 9% and 15% of active Twitter accounts are bots

From Twitter: “roughly 5% of our monetizable daily users are bots”.

Both can be true at the same time, and it does not mean that Twitter lied.

Re:Twitter says Musk’s spam analysis used tool that called h... by dahdum on Friday August 05, 2022 @04:27PM (#32361589)
Twitter's lawsuit mentions this, stating only they know who is monetizable. IIRC Musk's lawyers deliberately push the same false equivalency in their filing.
Re:Twitter says Musk’s spam analysis used tool that called h... by sangnoir on Saturday August 06, 2022 @01:52AM (#32365275)
Musks lawyers aren't fans of being disbarred or other legal sanctions, so they carefully couched the language in their filing to say "Musk believes [the intentionally confusing, possibly false narrative about bots]" without providing evidence, which is weak sauce.
Re:Twitter says Musk’s spam analysis used tool that called h... by jcranmer on Friday August 05, 2022 @02:41PM (#32360180)
As explained many times by many people, the 5% is not the number of bots on Twitter. It is the false negative rate of Twitter's bot detection algorithms.
Re:Twitter says Musk’s spam analysis used tool that called h... by jjoonathan on Friday August 05, 2022 @05:26PM (#32362170)
5% of mDAU is a false negative rate.

5% of users is not a claim that Twitter made.

Re:Twitter says Musk’s spam analysis used tool that called h... by loceng on Friday August 05, 2022 @05:10PM (#32362024)
... "but that it was doing so specifically so that Musk would buy Twitter at an inflated price."

The insanity is people being perfectionistic in that it was specific intent to target him to buy it - even if it says "him" - it's absurd that people are taking that literally, and not simply to "make it seem valuable for someone to buy."

Re:Twitter says Musk’s spam analysis used tool that called h... by jjeaff on Friday August 05, 2022 @02:30PM (#32359998)
Whether they inflated their numbers or not, the burden would be on Musk to prove that and it would be completely impossible to do that with a 3rd party tool. You are a monetizable user on Twitter if you just log in and read tweets. Something that no 3rd party tool could possibly track.
Re:Twitter says Musk’s spam analysis used tool that called h... by zinekeller on Friday August 05, 2022 @02:36PM (#32360089)
> You are a monetizable user on Twitter if you just log in and read tweets.

Actually, even logged-out users (until they forced everyone to log in to read timelines) are monetizable, so Musk is really trying to make something to get out.

Re:Twitter says Musk’s spam analysis used tool that called h... by eastbound on Friday August 05, 2022 @03:00PM (#32360438)
I’m surprised that someone clever enough to run a business, takes that excuse to get out of a deal. He must have been tired.

Any businessman who gets into such a position, even by mistake, would have at least taken this opportunity to require Twitter’s methodology on counting bots an expose its flaws. The claim would be something tangible like “This account is a bot and your criteria falsely counts it as a user.”

Not “Really there are 5% bots? Everybody, look!” He looks more like an engineer who can’t stand the fact that ballpark numbers are a decent way to do business in a lot of cases.

Re:Twitter says Musk’s spam analysis used tool that called h... by evan_ on Friday August 05, 2022 @04:32PM (#32361647)
Another explanation could be, he’s not actually that good at running a business
Re:Twitter says Musk’s spam analysis used tool that called h... by plorg on Saturday August 06, 2022 @06:31PM (#32372202)
A large portion of Twitter's complaint is describing how many times they tried to explain their methodology to him, including in written documentation that he admits he didn't read.
Re:Twitter says Musk’s spam analysis used tool that called h... by lokar on Friday August 05, 2022 @11:47PM (#32364643)
No, the burden is not on Musk to prove anything. He waived all rights to due diligence. There is nothing to prove or disprove.
Re:Twitter says Musk’s spam analysis used tool that called h... by daed on Friday August 05, 2022 @03:29PM (#32360848)
One point not mentioned (and I’m just paraphrasing Levine from Money Stuff here):

It’s crazy because it would probably be BETTER if you had more bots. Then you could say that your revenue per user was higher and you could argue that you had more room for growth. So, if anything, lying and OVERestimating the bot count would help your valuation more.

Re:Twitter says Musk’s spam analysis used tool that called h... by Ferret7446 on Friday August 05, 2022 @04:41PM (#32361737)
Not really, because the revenue comes from ads not users. Twitter makes profit from ad buyers thinking they are reaching many users. The "revenue per user was higher" you mention is actually "cost per ad impression" and ad buyers want that to be low.
Re:Twitter says Musk’s spam analysis used tool that called h... by Sebb767 on Friday August 05, 2022 @09:47PM (#32364036)
Ad buyers want this to be lower, but investors would love for this to be higher. I think that's what the parent was aiming at.
Re:Twitter says Musk’s spam analysis used tool that called h... by benj111 on Friday August 05, 2022 @02:50PM (#32360293)
The fact that musk tweeted that he was buying Twitter to fix the bot problem.

So the fact that he's using his reason for buying to try and get out of buying seems... Insane.

Re:Twitter says Musk’s spam analysis used tool that called h... by fooey on Friday August 05, 2022 @03:52PM (#32361132)
There's a whole assumption that bots are a net negative to the platforms value besides any arguments over how many of them there are.

As a human, it's annoying to get get a reply notification only to discover it was bot spam, but from the platform perspective that bot just created user activity and content for you for free.

I have a theory that the most embarrassing thing to come out of this might be that Twitter doesn't really care about bots much more than issuing platitudes that they care about bots to assuage the user base.

Re:Twitter says Musk’s spam analysis used tool that called h... by ryantgtg on Friday August 05, 2022 @05:13PM (#32362053)
I have like 9 twitter bots and they’re all cool and interesting. I wonder if Musk’s tool differentiates between cool and uncool bots.

In the past I tried to write bots that do things like reply to people, but that is a violation of the TOS and twitter would immediately identify and ban the account.

Re:Twitter says Musk’s spam analysis used tool that called h... by fooey on Friday August 05, 2022 @05:24PM (#32362143)
The tool he used to identify bots identified his own account as a bot, so I'm gonna go out on a limb and guess it's not very intelligent (or is hyper-intelligent)
Re:Twitter says Musk’s spam analysis used tool that called h... by hef19898 on Saturday August 06, 2022 @04:19AM (#32366108)
At that level of intelligence I bet it can drive a car.
Re:Twitter says Musk’s spam analysis used tool that called h... by blackoil on Friday August 05, 2022 @04:56PM (#32361883)
If there are too many bot activities, user may completely turn off notifications and reduce the usage.