Igor Ivanov comments on Problems of people new to AI safety and my project ideas to mitigate them