Séb Krier, Google DeepMind AGI policy lead, jokes that AI safety researchers operate like a secretive Rosicrucian order

Do you: - Believe your work is necessary to save the world? - Possess hidden knowledge ordinary people could not understand? - Attend confidential gatherings of like-minded brethren? - Collect esoteric texts on demonic behaviour? - Negotiate with the clergy and centres of power? - Defer to the authority of an inner circle? Then you're not an Al safety researcher. You're a Rosicrucian with compute.