Then, What Happened to the Model Predictive Control?

I was trying to do some adaptation where like you have this environment that would be self resetting. And I was looking for an algorithm, like a backbone algorithm that I could use for this problem and just add my contributions of top of that. But the more it looked, it just seemed to not exist. So what we ended up with really was an algorithm that would do planning in action spaces without any kind of video prediction or anything like that. It turned out that we could actually get better at long horizons than people had thought.