Mention is made of “using AI” and other data sources, and that’s what I’d like to read far more about.

I wonder if the new future is writing MCPs so agents can access the data.

Is this describing use of something like GitHub copilot?

If you look at hydroponics/aeroponics, plants basically need water, light, and fertilizer (N (nitrogen) P (phosphorous) K (potassium), and a few trace minerals). It can be the most synthetic process you've ever seen, and the plants will grow amazingly well.

The other elements regarding soil health, etc, would be much better framed in another way, rather than as directly necessary for plant health. The benefits of maintaining a nice living soil is that it makes the environment self-sustaining. You could just dump synthetic fertilizer on the plant, with some soil additives to help retain the right amount of drainage/retention, and it would do completely fine. But without constant optimal inputs, the plants would die.

If you cultivate a nice soil, such that the plants own/surrounding detritus can be broken down effectively, such that the nutrients in the natural processes can be broken down and made available to the plant, and the otherwise nonoptimal soil texture characteristics could be brought to some positive characteristics by those same processes, then you can theoretically arrive at a point that requires very few additional inputs.

After talking to fellow natural hobby farmers I realized the soil quality was garbage (lack of earth worms and insects), and there were severe drainage and water holding issues: weirdly the soil didn't hold water but it drained way too slow too. So, ehen it rained it was swamped for days but when it got dry none of that water stayed at the top 1 meters of the soil. I'm lucky to find amazing help from local natural farmers, so I got natural green compost (no animal products/byproducts). I have been introduced to no-dig farming too. So first year I started by applying 20cm thick compost on top soil, after putting a layer of old paper boxes against weeds. Then planted my seedlings on these, with worm poop and for some phosphate loving plants bat guano as fertilizers around the plants, topping of with hemp mulch and cacao shell mulch as topping. When this soil has sunken enough, topped off with 2-3 cm compost and mulched again. I have sprinkled insect friendly flowers to attract insects too. This was an amazing succes with not only plants flourishing, fighting diseases much better and resulting in an amazing yield. I didn't need to water as often as before (4x less frequent than before in the soil, 8x less frequent than in the pot). After year 3 I stopped all fertilization and introduced cover crops that could be used as mulch and fertilizer at the same time.

This process though is not linear. I still have plants which are not successful at all. I can grow juicy tasty watermelons in a northern European country but no parsnips or carrots or cauliflowers yet. This is what I love though, I'm interacting with a living microbiome rather than executing lab experiments. Failures are keeping it interesting and improving learning.