Tomoko Mitsuoka comments on Why AI Safety’s Measurement Frameworks Are Blind to the Risks Already Happening