It's almost like they're trying to sell these updates without realising that an LLM can't tell what it's outputting no matter how much "fact checking" they think might help...