Email or username:

Password:

Forgot your password?
Darius Kazemi

I'm hearing Chris Bouzy has a new corporate social network site he is trying to start up, Spoutible.

My history with Bouzy is I know him to be the face of Bot Sentinel, a service that purports to automatically detect problematic Twitter accounts.

It does not work.

Bouzy claims it works because he defines his metrics of success as 'whatever Bot Sentinel says is problematic is problematic', which, cool tautology generator you built there. More criticism from me:

wired.com/story/bot-hunting-is

17 comments
Glyph

@darius oh great he turned that thing where MRA influencers and the obnoxious high schoolers in their audience call people "NPCs" into a business

Darius Kazemi

Spoutible is apparently a reskinned version of cheap social network software you can buy off the shelf, which matches with my experience of Bouzy and his hyperbolic claims to the media and general huckster nature.

mementomori.social/@rolle/1098

Mark Sample

@darius Not even sure why he or Spoutible are getting so much oxygen over on the other place.

waldläuferin unterwegs

@darius Thank you for mentioning this. I was wondering if it was worth checking out. Still missing Spooler, by the way 😔

Brian Gawalt

@darius I do find it frustrating to come across uncalibrated estimation routines; to see someone "correctly predict twenty of the last eight bot nets"

Brian Gawalt

@darius it's a cliche/joke in macroeconomics to say someone's predicted twelve of the last three recessions. The Rainmakers branch of the Frauds Family Tree.

Darius Kazemi

@bgawalt oh I was just wondering if someone claimed in a paper to predict 20/8 botnets because I would believe it

Brian Gawalt

@darius i dont read many papers any more, but by happy coincidence, the last one i read *was* about bot detection: OpenAI's "did this text from from ChatGPT" classifier. And they genuinely and helpfully *did* break their results into both false-positive and false-negative rates, rather than hiding behind the more typical reporting of "it is a classifer" or "it has 99% accuracy [on a dataset with 98:2 class imbalance]".

OldTurk🗽Focus: #FreeAssange

@darius 1 ) awesome you were quoted in wired !
2) I’ve said in the past, if Musk were serious abou catching bots, he’d add a CAPTCHA to the “send tweet” button

DELETED

@darius He defined my turdsite account as a bot because I called out the shit behind the algo years back when he started it.

Probably Paul 🌍

@darius I've read that he also made the start-up money for Bot Sentinel from crypto scams. He also sometimes engages in quite toxic behaviour on Twitter, effectively doxxing people to his large following. Most of the time, it's clear those people deserved to be exposed, but other times? Not so much.

On those occasions, we're expected to just take Bouzy's word for it that these are problematic characters, and that putting their personal safety at risk was justified.

Groove, M.D.

@darius I keep reading that as 'spoutbile' and I don't see the mistake.

Kurt Fliegel

@darius One or two firms in the marketing side of the business have had some solid success with evidence-based detection of bot networks, keeping their heads slightly above water against the daily code changes written specifically to counter their detection tools. Platform and publisher sides of the business have been reluctant to use them on content because of expense. Not an easy problem to solve but industry can do much better.

Go Up