Pull a weed? No, spray a crop. We could polish a thing thing but instead we will apply a coating. And so forth. Now, "pulling weeds" doesn't spread. It doesn't end up in water and going downstream. The act of polishing doesn't end up in food.
I am beginning to think that chemical design should include facets like "breaks down with such and such a halflife at STP" (and then the same rules apply to the breakdown products) and "everyone involved who wants to say it is safe, down to the CEO, needs to bathe in this crap for ten years before accepting any kind of financial benefit."
Then again, I also spend a lot of time wondering how much of the "cide market" (insecticides, herbicides, fungicides, et al) could be replaced by wee robots trundling their way through rows of vegetables, detecting weeds and yanking them, applying small and targeted blasts of fungicide here and there, plucking away insects and their eggs, and so on. Would such robots be affordable? Could they break even against the chemical approach? Or would they cost more and would we collectively decide to pay that price?