Here we spot wild Bardfinn Bluesky activities.
Be valid and ping ! bardfinn for something worthwhile or create a new thread.
/h/bardfanns
BardWatch
She/Her
Anti-Hate-Brigade
27d ago
(text post)
89 thread views
#332402
earned 16 coins from votes
Bard spotting on Bluesky 11 January 2025 
- 21
- 3
Now playing: Gang-Plank Galleon (DKC).mp3
Jump in the discussion.
No email address required.
Orangesauce (@orangesauce.bsky.social):
Serious question: what do you do when you find an obvious MAGAt on this site?
I look at their followers with intent to block all of them.
Of the 3% I find myself following, are they just poorly vetted, following the MAGA to troll them, or just uninformed. Should I DM them and let them know?
USA_Patriot_76 (@usa-patriot-76.bsky.social):
Block:
OnlyFans
Illuminati
Vacant accounts
Other porn accounts
Illicit drug accounts
Misogynist accounts
MAGAs
Orangesauce (@orangesauce.bsky.social):
I'm referring to the people who follow these accounts tho.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
Understand. I just block leakage to my account
Orangesauce (@orangesauce.bsky.social):
I wonder if that works? Bot farmers have big automation and dozens/hundreds of accounts. Clearsky shows who someone is blocking, so it's public info for any bot farmer to pick up.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
Yeah, I had an experience a few days ago where a porn/onlyfans/prostitution account had 9 very similar clones. When I spotted them I went the "Report" then "Block" way. Exhausting but they folks at Bluesky can't do anything if they aren't aware.
Orangesauce (@orangesauce.bsky.social):
We need better cowtools to help moderators. The abuse problem just gets magnified tho.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
I suggested to Bluesky they build an algorithm identifying accounts with defined increased level of blocking from other accounts in addition to direct user reports.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
This detection of "crowd blocking" would speed cleanup of Bluesky membership.
Illicit accounts along with an incompetent billionaire killed Twitter.
Orangesauce (@orangesauce.bsky.social):
I'm trying out some theories right now with obvious MAGA accounts. One of the most important is curating my own follows by seeing who we have in common and removing them after checking in what's called "depth first search"
Orangesauce (@orangesauce.bsky.social):
It doesn't work perfectly, still requires discretion when unfollowing. So it's difficult to automate.
I guess my question is if I find my account has bad followers and I think I'm pretty good at vetting, do I really want to block others who made the same mistakes?
USA_Patriot_76 (@usa-patriot-76.bsky.social):
You just click on the offender's profile and block
Orangesauce (@orangesauce.bsky.social):
I'm talking about looking at the offernder's followers, ad infinitum, blocking the offender when I am done blocking all their followers.
Just found some good block lists at @skywitches.net.
Ms. Penny Oaken, SkyWitch (@skywitches.net):
Just so y'all know,
has been tried before and it gets promptly detected by bad actors and weaponised to get good faith / authentic users suspended.
Ms. Penny Oaken, SkyWitch (@skywitches.net):
It also ironically introduces a tonne of manual review and investigation, which most social media moderation departments won't do, and bsky avoids.
So the line is at "evaluate user reports in isolation", for the most part.
Ms. Penny Oaken, SkyWitch (@skywitches.net):
We can build and share blocklists though, based on any criteria we want, and increasingly I've been thinking about how those criteria can be clear, concise, and obviously necessary
Orangesauce (@orangesauce.bsky.social):
Thanks.. What I was just entertaining while subscribing to a few of your blocklists is to transitively follow all paths and unfollow people I accidentally follow. I got listed a few weeks ago and vetting 13K people in a few weeks apparently led to some errors...
Jump in the discussion.
No email address required.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
It's a solvable algorithm problem. Each Bluesky member is responsible for keeping site clean of bad actors. Start blocking domains if necessary.
Ms. Penny Oaken, SkyWitch (@skywitches.net):
the issue is that there's no workable algorithm for "is XYZ Bluesky User Authentic", only "is XYZ Bluesky User engaging in abuse of the service".
So bad actors will flood the site with fake accounts and use them to submit reports. Which is an incentive that shouldn't be rewarded.
Orangesauce (@orangesauce.bsky.social):
What I was referring to here... bsky.app/profile/oran...
Jump in the discussion.
No email address required.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
You can set a "number of followers" and "time on site" requirements for reporting.
Ms. Penny Oaken, SkyWitch (@skywitches.net):
Which also gets reverse engineered and exploited, as well. We saw it on Reddit a decade ago, where a network of pro-Trump astroturfers aged in their sockpuppet accounts for eight months in sports subreddits before using them to engage politics threads.
USA_Patriot_76 (@usa-patriot-76.bsky.social):
Sovereign state-level bots will continue escalating spy vs spy activities. But it will stop many OF and other abuses of the platform. My last round of followers was 2/3 bots.
Jump in the discussion.
No email address required.
More options
Context
More options
Context
More options
Context