“John Ennis expressed the opinion that Anthropic's public statements about AI safety are primarily motivated by a lack of compute to serve its models at scale and a desire for regulatory capture.”