This optimization is premature in that he's trying to optimize for time and effort when we haven't yet nailed — in terms of fully understanding — the whole "optimal nutrition" thing.

Optimizing for performance before you have a fully working implementation is exactly what Knuth was decrying.