AI that works is just software; a robot that works is just a machine.
A modern tractor is a phenomenally complex machine, capable of fully autonomous operation with millimetre precision. The system is guided by a complex set of 3D soil maps based on moisture holding capacity, compaction and texture, root formation and a multitude of other factors. Each square foot of field receives precisely controlled amounts of water, fertiliser and pesticide. Nobody calls it robotics or AI, because it works.
It's also not unique to AI. One person with a tractor can outharvest the amount of corn it used to take a whole village to harvest. This feels just like a dressed up luddite complaint.
Farmers use a lot of technology. So if he is talking "AI" think like combines keeping the "lane" like your car and tractors sowing basically on their own adjusting seed and fertilizer levels as needed as they go. And none of that needs actual "AI". Of course we are talking large industrial farm companies here in many cases. Not your mom and pop farm.
Yeah this isn't terribly exciting. People have been demoing self driving tractors and farm machinery for decades. Ford had one in the 1950s. Most of the hard work like actually picking the crop has already been solved, and making the tractor self driving just saves a few hours of work at most.
What would be really cool is automating crops that still require lots of manual labor. Like vegetables. That's the reason they are still so expensive. An automated greenhouse would be enormous news.
Without using AI there have been substantial improvements already - I was home for harvest on my family farm last year, driving a JD S970 (last time I drove was a CTSII about 10 years ago).
40 ft front, gps auto steer, radar controlled front height - cognitively it was 10% of what it was 15 years ago.
My uncle had a new X series with a 60ft front that also slaves the chaser bin for unloading so you’re basically automating the unloading as well - it’s wild.
Our country is all to undulating I think to be even close to being fully autonomous anytime in the next 2.0 years but it will be interesting to see how this all develops.
10 years ago the talk of remote tractors was to have one lead manned machine with 1-3 ‘slaves’ machines following behind so the one operator can manage a much bigger area, it’s interesting that this doesn’t seem to be the way they’re going
I agree, there is a lot of delicate and manual agricultural work that I would like see AI tackle. But this doesn't seem to be one of them.
The article is a little light on the details, but it doesn't look like this robut does any manual work on the flowers. Rather, it seems like it uses image recognition to plot the GPS coordinates of potential sick plants (with the assumption that a human will follow up to remove them manually if necessary).
It's not that common. I don't know even one farmer who uses autonomous tractors. Like literally..
Except..Right now, blue river tech's thinning machine in salinas but that's kinda hitched to a tractor, iirc.
And here is why..1. They are hundreds of thousands of dollars and it is not accessible to smaller farmers. 2. Current tech is more useful for commodity farmers and not the kind of farmers who bring your tomatoes or kale to the kitchen. 3. Corporate farms that cultivate a couple of thousand acres will have more use for it. Small farmers like me won't do it. 4. Even large farmers will be suspicious ..because as a farmer ..if something malfunctions(large machinery and threat of malfunction is a package deal)and it is a few tonnes heavy and your entire harvest over a thousand acres is dependent on machinery you didn't tool or fashion with your own hands, you tend not to sink in a lot of money in it. 5. It beggars belief that they won't come up with smaller robots for the field that is affordable. Because it dilutes risk. Most importanly, we need it. One can harvest a couple of hundred acres of sugar beets with a small crew and machines within a short period of time but when it's time to pick tomatoes, it has to be done by hand only..only with back breaking labour..perishables must be picked right ON that day.
This is important to us. Labour issues are not going to get any better. It was never in a good spot in the first place anyways.
The problem is that AI isn't AGI. So a robot that can pick strawberries (exists, is in commercial use) cannot also pick apples (also a robot in commercial use). The apple robot needed its own R&D cycle. Grapes are different again, and olives, and lemons, and cherries, and almonds... And that's just horticulture.
In a highly diversified economy, it is going to take multiple decades and person-millenia of effort for the AI we have to show much impact.
Automated tractor is not as complex as automated car since the environment is much simpler. I think it's closer to iRobot Roomba rather than self driving car.
reply