This is amazing, and I'm glad we're now starting to see the benefits of these startups and companies come to Ag.
I was in my last year of apprenticeship in Switzerland when New Holland was testing their autopilot planting system, something John Deere had already deployed, and it pretty much had to be corrected every few meteres. But the overall sentiment was that this tech is great, but is still a long way from becoming a reality.
Having been in tech and from CA I knew that was partly true but that you cannot dismiss the rate of progress if bright minds and resources are coupled with determined and skilled workers.
I had access to one of those John Deere, it had a realtively comfortable cabin with a radio and seat warmers etc... but after seeing the sticker shock (500k CHF) I saw how they were closed sourced and had to be serviced by appointment only--the brother in law owned the nearby JD dealership in town. I never bothered to use them more than one or twice when the older manual tranmission and for more spartan Renault, Masey-Fugerson, Case, Fiat tractors weren't available or had to haul much larger cargo (100s of tons of potatos to the fulfillment depot, or take 20 cows out of pasture into the barn under 15 mins) due to snow storms.
With that said, this has unit has seemingly limited applications: as you can see this was done in the early phases of growth (looks like very early germination/sprouting) and while you can do this in later stages for things like greens and possibly tubers and other lower growth crops with spaces in between this won't work for things that like grains and corn which are much taller, more densely sowed and have thicker canopies and are what accounts of the most mono-crop and pesticide use outside of cotton.
It's amazing, and I hope we see more of this instead of smart phone 'innovation' but this is why I think rather than shoot for full autonomy in Ag we should just have equipment that can focus on the more menial and labor intesive things like weeding, planting, harvesting because the farmer will always need to be on sight for things like pH test, irrigation, nutrient fluctuations, changes in weather (hail, frost) etc... And have the farm manager be paid well, like 70K+ and that they get trained from an early stage in their careers to rely and use this kind of technology wherever possible and have direct feedback with technologists to direct them to other uses.
I advocated for this as much as I did for sustainability and renewable Ag when I was farming, and it was looked down on in the Biodyanmic community that relies on lots of 'woo' to mask that it's a back breaking, manual labor intensive job that is necessary to create a superior product and justify the +30% mark up in price.
This alone would free up so much more Human capital and allow one to do more of their time that can help sustain the business and help in creating auxiliary business models to support the core instead of focusing on weeding and lower yields due drops is soil fertility due to overuse of fertilizers and pesticides and a direct result due to lack of labor, which other than the ongoing water crisis, and climate change is the biggest issue for farmers in the Western/developed World.