When communicating about policy and technical AI work, one should aim for intellectual honesty about timelines and risks; the 'slow world' compartmentalization is defensible in personal life but not in professional/policy contexts where truth-telling is primary