jacob_cannell comments on Embedding safety in ML development