Safe As YT week in review: Failed leader walkarounds, interview with Sidney Dekker, awesomeness of coffee, and how we trained AI to hallucinate

Did you check out my Safe As YouTube channel this week?

Link: https://youtube.com/@safe_as_pod?si=nrKljK0AeUZXTOgd

We covered:

1. Part 1 interview with Prof Sid Dekker, discussing a definition of safety, perils of zero harm, blame, restorative culture and drift

2. Andrew Hopkins analysis of the leadership walkaround on the Deepwater Horizon rig and how it failed to address the major hazards (Shorts)

3. A study discussing how AI LLMs have been trained/rewarded to guess when they don’t know the answer–e.g. a focus on plausible-sounding answers (Shorts)

4. Sid Dekker discussing the need for compassion with accountability and learning

5. The awesomeness of coffee – and its (epidemiological) links with reduced early death.

Check it out and if you find it useful, please subscribe, share, like and comment directly on the videos on YT.

Leave a comment