What is the energy source for that? Solar? Is it therefore possible to have enough solar energy to operate the machinery with enough to pull out and keep on going? And wouldn't you need a tremendous amount of batteries as well? Would seem there would be an environmental impact that could far exceed some kind of potential 'safe' chemical solutions.

It actually seems like a straightforward problem for solar energy. Just have a little "bot barn" with a solar roof and local storage, and the bots return to the bot barn for a recharge. That's almost a trivial problem.

Keep in mind the bots don't have to be very large; certainly not human-sized. Something the size of a roller skate, maybe, with specialized weed-killing tools. And if it encounters something too big or weird to handle on its own, it can mark it on a map and send an alert for human intervention.

Thankfully, labour is cheaper in India than many herbicides, so farmers didn't subscribe to them in big numbers. I can tell you with certainty that we don’t use glyphosate on wheat or sugar cane (at least in a large part of India that I know).

There's a danger these companies can make inroads though: I was in field few days back and there was an Agriculture University professor who quite paternalistically suggested to a farmer that he use this product to de-weed the field. Gosh, degrees and education aren't enough sometimes. I wasn't in the position to debate (as someone younger and a software engineer), it made my stomach growl.

We honestly need to have ethics trainings for Professors/Researchers across the globe.

Perhaps little automated dirt guns like mobile SAM sites that can shoot down pests as well? They can just pick up a few grains of dirt and shoot them at any pests that come near the plants.

That's what I was thinking of but salt doesn't do well with plants unfortunately so that's why I thought dirt might work.

At a smaller scale, Farmbot already does this.

https://farm.bot/

That would be pretty cool. I've seen some early prototypes in the news, so perhaps we'll get there in the not-too-distant future. It seems to me that chief challenges would be the AI of course, but also energy and mobility across a wet, tangly, clumpy terrain.

Even in that case, we would have the issue of Vavilovian Mimicry. This time being done by robots, which would certainly be a counter-argument for those that claim that this is not natural selection, but artificial selection.

or even better: abandon monoculture

Abandoning monoculture means completely upending the harvesting process. That's a much harder problem. Or abandoning annual crops in favor of perennials, which would have tremendous negative impact on productivity.

Don't let the perfect be the enemy of the good.

in the long term we will have to do this with perennials and find new harvesting techniques as current system is creating dustbowls long term