I might be wrong but is this really a solution for RSI? I interpret it's as a way to reducing wasted tokens and compute on paths that don't yield better results. It's an optimization. It's a faster way to get to RSI though. what's wrong?
MSFT, dell, NVIDIA. These guys will make money no matter who uses what AI since they're the hardware providers. Even though they all own stake in frontier labs, chinese open weight models will commoditize the intelligence layer.
Do you make this decision after doing tons of research about how you could implement a solution for the gray areas or do you just decide not to pursue the gray areas that you don't know how to implement yourself?