• Anna Riedl: “The amount of prompt engineering needed to not make medical professionals have wild irrelevant hallucinations is truly extraordinary”

  2026/01/30

• Using AI to sweep a social media reply chain to actively block abusers and bots…

  2026/01/15

• I am totally here for accusations that ChatGPT & other LLM/AI are demonstrating woke liberal bias: Part 2

  2026/01/09

• I am totally here for accusations that ChatGPT & other LLM/AI are demonstrating woke liberal bias

  2026/01/06

• The coming war against local LLMs will be wrapped, as ever, in a choice between Child Protection and Anti-Terrorism

  2025/12/19

• 2015: “AI Images Look Ridiculously Fake” | 2025: “AI Images Look Ridiculously Real” | The solution to risk is education

  2025/09/15

• Center for the Alignment of AI Alignment Centers | …this is a wonderful and absolutely necessary tool for AI policy experts

  2025/09/13

• Teaching AIs to know when they don’t know something

  2025/09/10

• We’re Walling Off The Open Internet To Stop AI—And It May End Up Breaking Everything Else | Techdirt

  2025/09/09

• 2023: Reminder: the FDA permits 9x as much rat faeces in food than there is likely abusive imagery in an AI image training set

  2025/09/04

• The possibilities that would be raised by a coordinated prompt injection civil society campaign, are limitless…

  2025/08/19

• Today in “…just wait until the UK Ofcom censors find out that you can talk to AI Chatbots over DNS…”

  2025/08/10