[hedgehog]

Like the rest of industrial age development, farm automation has historically been built around treating everything uniformly (even distribution of seeds bred for easy harvest in evenly spaced rows evenly fertilizied with no rocks etc etc). Moving away from that introduces all kinds of complexity, mechanical problems, data problems, etc, which are not easy to solve even when you have historical existence proofs of potentially better ways to do things. It's happening though.

[lotsofpulp]

It introduces trying to sell your crops for 5x what your competitor is and no buyer is going to pay 5x for “sustainable” produce.

[hedgehog]

What I'm getting at is things like applying fertilizer based on estimated need from imaging data, using robots with lasers to kill weeds, mechanized intercropping, etc. Not something consumers will necessarily see any direct impact from but that improve the quality and economics of production.

[lotsofpulp]

> which are not easy to solve even when you have historical existence proofs of potentially better ways to do things. It's happening though.

What I am getting at is if these “proofs” existed, they would already be proven by lower priced goods at the market.

[hedgehog]

A historical existence proof of a system or technology working has essentially nothing to do with whether it's being practiced in industry today, and whether it's being practiced today has very little to do with the product or pricing you see when you for example stop by KFC to eat a piece of chicken that's largely created out of corn, soy, and methane. That's the beauty of capitalism and commodity markets, John Deere can roll out technology for say computer-controlled planting that improves efficiency and the dividends get spread across the value chain without everyone having to know about it or change what they're doing.