In a series of posts on X, Steven Adler - who has been working on AI safety for four years - described his journey as a "wild ride with lots of chapters".