ddwoods.bsky.social @ddwoods.bsky.social · 29/05/2025bsky.app/profile/sara... see new article in the Bulletin of the Atomic Scientists, “The Silicon Valley Way: Move fast and break…aviation safety?” 021
ddwoods.bsky.social @ddwoods.bsky.social · 04/12/2024For empiricists the case is poignant: The original is a synthesis from empirical work; the LLM output is not empirically grounded, though scraped from noisy diverse material of unknown pedigree. This is a deeper form of AI Slop problem. 010
ddwoods.bsky.social @ddwoods.bsky.social · 04/12/2024#23 makes a big mistake. not just operators, it's every one’s / every thing’s models are incomplete, i.e., all agents at all layers in all roles, human and machine regardless of underlying tech. The constraint places a premium on the ability to revise & reframe (oops, the LLM missed that) 000
ddwoods.bsky.social @ddwoods.bsky.social · 04/12/2024For people here, my comments on LLM output: these are all basics from field of Resilience Engineering (RE)-repeated w/out attribution #22 quotes me in talks/papers; #25 is the basis for RE, a theorem (24 years old) & empirical finding (decades old) #23 is 1st line of Stella.report ... 000
