Graeme Ford comments on Do LLMs know what they’re capable of? Why this matters for AI safety, and initial findings