> The variables required in agriculture are just too many for something like this, is my opinion.
I'm certain there is a lot of complex knowledge locked up in farmers' brains that most folks wouldn't even think of. The same is true for doctors though, and it doesn't seem foolish that folks are creating expert systems to help with complicated diagnostics and such. Surely the wisdom of farmers can be codified and emulated reasonably well by software.
Further, maybe this is only limited to a subset of agriculture. It doesn't need to apply to every crop type to be useful. It's probably dumb to grow 100 acres of wheat this way. Perhaps it's useful for highly-controlled small beds of high-value high-labor-input vegetables or such?