The 80:20 rule?
Will it be 80% of accounts are bots? Will they be wiped when Twitter officially finds out?
Computer boffins Juan Echeverria and Shi Zhou at University College London have chanced across a dormant Twitter botnet made up of more than 350,000 accounts with a fondness for quoting Star Wars novels. Twitter bots have been accused of warping the tone of the 2016 election. They also can be used for entertainment, marketing …
I was just thinking about crafting a tweetybot group the other day, and this gives some insight on how to have it not look like a single group of accounts. Why even bother showing location info? As a human tweeter I never did. I guess if you have a known, smallish group of followers that would be something, but to let random chaps and bots have your location seems a bit daft to me. I almost ended up working for T, but whilst having a good chat with one of the local tech leads at their data center that very same week the CEO left, so when I didn't hear back I knew what that meant. Still, I think they might have some interesting gear and racks to peek at. And they seem to be cleaning up the place and making it nice again. Who knows if they'll ever turn a profit. How can you monetize a bunch of chattering idiots? Fascinating.
"And they seem to be cleaning up the place and making it nice again."
What none of these outfits seem to realise is that once people realise that XYZ million users claims are shown to be vastly inflated by bots or spammers, credibilty is lost until they start actively hunting down and removing them.
You can part the new and naive with their money but not policing the network for abuse (network abuse) results in repeat business dropping off.
I was just thinking about crafting a tweetybot group the other day, and this gives some insight on how to have it not look like a single group of accounts. Why even bother showing location info? As a human tweeter I never did.
I think if I had a Twitter account as a human or a bot I would set the location data of my turdspurts such that over a period of time it spelled out 'Mind your own damn business' in the middle of the north Atlantic.
Why would Twitter want to disable the accounts?
Apart from usual reason (more users == higher value), once you spot an idle bot army it is better to classify and watch the behavior. That way you can identify all of them, and watch for the creation of new accounts to track evolution. If you kill them all immediately, the behavior may change too dramatically to track.
Once you spot a bot, you can compartmentalise them, so they don't affect other users.
And subtract the numbers from the usercount. Honesty in marketing will go a long way towards building up trust and winning repeat business. Most of these companies have been behaving in ways that would make an east end barrow boy blush in shame.
Vrh "...velocity to move between previous activity location?"
Geotracking can be useful, but may have odd exceptions.
My work PCs enter the 'net from any of several locations, including Texas, NYC, or any number of other odd locations along eastern USA. Another PC can be Ottawa, Montreal or Halifax. My phone often appears to be anywhere in eastern Canada, up to 2000km away. My home ISP moves around, but only a radius of 20km.
Transitions from one apparent location to the other will occur randomly. Instantaneously. Triggering false positives in the algorithms of those unaware of this location noise.
Obviously all this depends on the technology and techniques being used. Of course a botnet isn't built with 300,000 actual phones, so we can't assume GPS. Network Location is sometimes the only data.
It's easy to find something when you know what you are looking for - it's a lot harder when you don't. The fact that the bots are so old should be enough proof of this but I could be wrong, can you show some pseudo code that will find me something I'm not looking for please?
I see them as an evolution of Twitter ultimately.
The trouble is the content produced and placed on the Twitter platform has outgrown the functionality of the site.
Sure theres a lot of noise and crap but I see using Twitter as being a bit like tuning in a radio.
I personally have a couple of bits on Twitter and they have a decent level of popularity. They are automated and tweet out financial data...the FX bot pumps out the current ADX and RSI values and its current interpretation of those to detwrmine a trend and I have a BTC bot that does the same.
Each bot has approx 600 followers and about 80-90% of them are active and real people.
My assumption is that these are sleeper accounts that have built up some sort of provenance over a number of years. This means they are 'more' trusted than a brand new account. (For some value of more).
The accounts are dormant but could be put into effect for a large one time purchase. e.g. somebody wants to win a presidential election and needs 'credibility' through retweets and likes etc. Perhaps 350,000 retweets of accounts that are a couple of years old, that have a number of tweets already, have a number of followers (possibly a large subset of the other 350,000 accounts). I know this is far fetched but bear with me....
This passes all sorts of tests. Difficult to block as they look pretty genuine.
Rob
"I don't understand this at all. What are these "bots" doing? Why are they a nuisance?
I am not a twitter user and never have been."
AHA! But you're not saying a word about your future Twitter intentions, are you, Alan Sharkey? Eh? Well, eh?
This just in from Urban Dictionary:
Trump's Razor: The stupidest explanation is always the right one." Example:
Guy 1: "Did you hear that Trump said Chinese bees cause global warming?"
Alan Sharkey (via Twitter, no doubt): "Chalk another one up to Trump's razor."
http://urbandictionary.com/define.php?term=Trump%27s+Razor&defid=10914945