Fortunately this is knowable because sticking magnets in cows is something that we've done for many, many years. So we know what it costs and how effective it

Fortunately this is knowable because sticking magnets in cows is something that we've done for many, many years. So we know what it costs and how effective it is. Whereas a robot to solve the problem doesn't yet exist. And the important bit here is to realize what the word "solve" means here. I'm not saying a workable solution will never exist, but in order to be as effective it would need to be orders of magnitude better at its job than a roomba is at its job. And what the roomba does isn't very hard to begin with. You would have to cover a very large area of sometimes pretty rough terrain every N days, and you would need to have a very high success rate at both detection and removal for the solution to be able to compete with the established practices. To understand the problem I can only recommend spending some time on a farm where they keep grazing cattle.