> "The theory behind these models so aggressively lags the engineering" The problem is that 99% of theories are hard to scale. I am not an expert, as I work a

> "The theory behind these models so aggressively lags the engineering" The problem is that 99% of theories are hard to scale. I am not an expert, as I work adjacent to this field, but I see the inverse - dumbing down theory to increase parallelism/scalability.