You now run a public school district with 50,000 students. Test scores are flat. How do you measure whether a teacher is actually working - and how fast can you test a new curriculum?
You're optimizing the wrong thing - stop measuring teacher effort and start measuring student velocity, then build a system that runs thousands of A/B tests per week instead of changing curriculum once per decade.
Look, I've seen this movie before. You're asking how to measure if teachers are working - that's the wrong question entirely. At Facebook, I never measured how many hours engineers spent at their desks. I measured what they shipped. What ma...