Emily M. Bender @emilymbender.bsky.social · 15/10/2026But in fact the bulk of the paper was about other harms: 1. Environmental harms 2. Barriers to entry into a research field hyper-focused on expensive methodology 3. Encoding & reproducing bias/systems of oppression >> 117719
Emily M. Bender @emilymbender.bsky.social · 15/10/20264. Unfathomably large datasets precludes careful curation, documentation, & other mitigation strategies Also known at the time but not covered were harms of labor exploitation and normalizing surveillance. >>9:52 AM · Oct 15, 2026 115818
Emily M. Bender @emilymbender.bsky.social · 15/10/2026So even in the very paper where the phrase "stochastic parrots" was introduced, the critique was never only "this tech is a bad fit for the tasks it's being marketed for". >> 213810
