David Daniel

daviddaniel.bsky.social

23 August 2026 22:14

ICYMI: The institute's "data reports" have footnotes, tables of contents, and a neutral tone, helping them appeal to chatbots like Claude or Gemini. The website says it "author(s) content engineered for how LLMs evaluate credibility." Others refer to it as "LLM poisoning." (h/t @rvawonk.bsky.social)

David Daniel

daviddaniel.bsky.social

23 August 2026 22:14

More via @rvawonk.bsky.social: Researchers say the campaign is designed to discredit the media, and force journalists and fact-checkers to spend resources repeatedly disproving fabricated content. It's basically a DDoS attack against cognitive infrastructure.

Mirth

mirthdal.bsky.social

16 August 2026 01:21

Clues to likely bots and botnets on Bluesky. Patterns I've seen: - #Resist (unfortunately, as also used by legit accts) - #FollowBack - bio: NO DMS - graphic memes A lot use generic "lefty" descriptions, like supporting LGBTQ+, pro-trans, but otherwise appear to have no actual interests

Mirth

mirthdal.bsky.social

16 August 2026 01:20

Is there a unique hashtag you use? That I can start using to tag accts I find suspicious?

Mirth

mirthdal.bsky.social

15 August 2026 20:04

5/ it's disconcerting just how many inauthentic accts I've come across via a very casual inspection And how many large accts are following and promoting them. Even among AltGov, who, at least the ones from Twitter, actually had training on recognizing fake / bot accts More training needed!