Thoughts on the AI Safety Summit company policy requests and responses — LessWrong