The data Musk has cannot be used to determine the %age bots as it does not have the private information of each account.
But it is under 5% anyway, they have no reason to lie about it.
The ongoing Elon Twitter saga: "insert demographic" melts down
- Thread starter Energize
- Start date
- Status
But it is under 5% anyway, they have no reason to lie about it.
Permabanned
- Joined
- 25 Oct 2004
- Posts
- 9,078
They absolutely do have a reason to lie about it, if a large number of active users on the site were fake/bots the share price would tank.
Commissario
And that is the whole thing, please define "active" users or indeed if Twitter claimed less than 5% of their total userbase or 5% of their posting user base, 5% of "browsing" users, or 5% of their userbase that interacts are bots?
There are at least three different possible ways to look at 5% of users or even "active" users, and that's before you start to narrow down your definition.
For example is someone who browses Twitter "active", is someone who once a week "likes" something an "active" user, how about someone who posts a comment or a retweet once a fortnight?
I suspect Twitters number is probably high by their definition (lawyers will have been involved in it, and they tend to be very cautious creatures that are easily startled by noises such as "you lied in legal documentation"), but you need to actually know what their definition is before you can even start to try and work out how accurate it is, and by the sounds of it Musk signed his deal waiving the right to examine that under an NDA.
It's almost certainly going to be extremely hard to show who is a bot or not on Twitter a lot of the time, as there is probably not much difference in a lot of the bot accounts and say "passive" users who might only like a comment once in a while until the bots go fully active and start doing the Tory MP thing of dozens or hundreds of identical posts defending the PM within minutes.
Even on something like a computer forum the definition of "active" users can vary a lot, as IIRC the software may consider anyone looking at the forum as an "active" user, or it may only count those who are logged into accounts, whilst others might only consider how many posts are made, or how many accounts make a post in a set time frame (and that time frame might be daily, weekly, monthly or more).
Permabanned
- Joined
- 25 Oct 2004
- Posts
- 9,078
I wouldn't underestimate just how much bots can make up the numbers, twitter has what? 230 million users, 5% of that is what 11.5m that alone is a startling number and enough to make a difference in what ever your trying to do. Now what if that number was actually closer to 20% or even 30% then your looking at a bot population that outnumbers most small countries, and it would certainly be enough to influence all manner of situations.
It's entirely possible that there are only 5% or fewer, but there should certainly be a way for twitter to track that sort of thing, considering bots generally just regurgitate what others are already saying and likely come from linked accounts it should be relatively easy to account for them. Throw an AI bot at the problem and no doubt you'll have a good idea.
It's entirely possible that there are only 5% or fewer, but there should certainly be a way for twitter to track that sort of thing, considering bots generally just regurgitate what others are already saying and likely come from linked accounts it should be relatively easy to account for them. Throw an AI bot at the problem and no doubt you'll have a good idea.
You only have to look at the replies to any popular thread to see the scale of the bot problem. Numerous cut and pasted messages.... Particularly bad on political posts.
most of the accounts on there might as well be bots, celeb accounts with PR people posting crap just label them bots, same as all the brand accounts and other crap that are only their to advertise themselves
Permabanned
Not sure I see many bots on twitter but I barely use it. Only follow a few comedy posters.
Soldato
Sure bit that's where they go, they don't post on Joe blogs twitter feed. You've essentially walked into the house of commons and said you think that there's MPs in every party of the country because all around you you see MPs.
^
People have extremely skewed perception because they see popular things getting botted (because that's where the traffic and attention is).
Permabanned
- Joined
- 25 Oct 2004
- Posts
- 9,078
I wouldn't be too dismissal of bots, they can have a huge impact on any platform when deployed en masse. Take for example the game Lost Ark, a week ago it had nearly 900,000 concurrent players on Steam, following a huge bot ban wave, it's down to <300,000 concurrent players, where there's a financial incentive, expect bots in their tens or even hundreds of thousands.
Soldato
Dangerous thing to say, they have plenty of incentives to lie about the actual number. Share price, reputational risk to both the company and it's directors, questions from government officials and committee's, potential SEC violations the list goes on.
which is why their definition will be extremely robust and supported by their legal team smd unlikely to be wrong by any significant factor
Permabanned
- Joined
- 25 Oct 2004
- Posts
- 9,078
Yet that statement about bots only being 5% could have been made under the knowledge that musk wouldn't be able to verify it as he didn't have access to the firehose at the time. Lets not forget that they also were relucant to even give him access attempting to prevent it happening using multiple different legal angles before it was eventually ruled in Musks favour that he was legally entitled to the information for the purpose of verification before the sale could even complete which I believe were part of the terms anyway.
Commissario
The statement about the number of bots was from memory in line with their official filings to the FEC (and which they've done every year as part of their legal obligations as a publicly traded company on the stock exchange), which means if they've lied about them they're not just looking at a small fine, but potentially a very large fine and from what I recall potentially jail time as I believe it's one of those legal documents where people are personally held responsible for putting their signatures on it.
Also IIRC Musk waived the right to the information initially in the original pre purchase contract he signed, that information is routinely given under an NDA if someone is a serious potential buyer, so I'm not sure what the court case was about.
Caporegime
As Twitter pushes back, Musk inches closer to retreat:
This made me chuckle:
Still hard to say which way the chips will fall, but it seems increasingly likely that Musk will head for the nearest exit he can find.
This made me chuckle:
Still hard to say which way the chips will fall, but it seems increasingly likely that Musk will head for the nearest exit he can find.
Was he really trying to buy twitter just to shut down accounts that exploit his image to sell crypto?
Then again that one dude that tracks his flights seemed to annoy him greatly so maybe he wanted to purchase it just to get revenge on him.
Then again that one dude that tracks his flights seemed to annoy him greatly so maybe he wanted to purchase it just to get revenge on him.
Last edited:
Commissario
I think it was more likely he got upset by the people that kept pointing out how he wasn't the best thing since sliced bread, and that the build quality on his cars was worse than a cheap Kia etc (seriously, I had a £6k when NEW Kia that had better external QC than I've seen Tesla owners accept).
Caporegime
- Joined
- 30 Jul 2013
- Posts
- 29,418
How's that meltdown going?
Yeah but could your KIA project the word "KIA" on the floor from its headlights?
Commissario
It barely had a stereo
but the panels were aligned properly and the paint was well applied*
Seriously it was about the most basic car you could imagine in the early 2000's, but it was put together better than many tesla's from what I've seen*, probably because Kia were using robots and a functional production line with proper QC so it was consistent (whilst Musk seems to have wanted to throw out 100 years of lessons in mass production and QC across all industries in his attempt to reinvent the car).
*I ran it for 9 years, then my sister had it for another 5 until the labour cost for a relatively small repair was going to far exceed the value of the car (a part had gone, but the bolts had seized so it was going to require a lot of manhours).
Last edited:
- Status