Stop trying to review AI's code faster: bet on rollback instead

the amount of cope required to say we should just embrace mistakes because it's "more efficient" is staggering