It was submitted to HN 2 times already but unfortunately it flew under the radar: https://hn.algolia.com/?q=https%3A%2F%2Fwww.youtube.com%2Fwa...
A claim in the video that I can't verify but makes economic/logistic sense is that the speed problem isn't the panels but the controllers. The current crop of controllers are optimized for low power, which fits the e-reader use case but that is not optimal for the interactive use case.
Eink is awesome for low-power use. Why try to contort the technology for something it's not good at, instead of using a more appropriate technology like transflective LCDs? Eink isn't the only option for reflective displays. If you increase the power use of eink to get better refresh rates, I imagine you'd end up using more power than (and still end up with lower refresh rates than) an MIP display.