Dennard Scaling
en.wikipedia.org
en.wikipedia.org
This is interesting because it brings time scopes to mind. Sometimes you see a significant mean-reversion happening on one timeline, let's say a five-year-resolution timeline. Meanwhile the decades-scoped timeline is far from over, in terms of the principle no longer applying.
So, I wonder if we might be about to enter a period of relative volatility with regard to this principle. It would make some sense looking at some of the other relatively impressive progress we are seeing in computing as a whole.