There were definitely still very impressive engineering breakthroughs.

Also it’s pretty good confirmation that synthetic data is a valid answer to the data wall problem (non-problem).