Anshuman comments on What were mistakes of AI Safety field-building? How can we avoid them while we build the AI Welfare?