It's a purely mathematical counter to the argument that if plants suffer, people (who consume a roughly fixed amount of protein) could just eat animals instead. That's a contradictory position, because animal protein comes from animals eating plants.
The article is a little vague on why killing plants is bad. Whether it's because you believe that destruction is intrinsically bad, because they have souls, because they feel suffering, or because there's some potential for growth that's aborted, the various reasons are harder for me at least to understand and to build a coherent argument around.
It's philosophically very obvious why killing people is bad, this is so self-evident it's almost nonsense to write it down and inadvertently equivocates it with killing plants: it's trivial to imagine oneself in the position of the victim, or to enumerate the harms being perpetrated, the goals and values being denied, and the rights being violated in the transaction, and conclude that it's wrong. No reasonable person can reach a different conclusion.
It's a little more difficult to extend this empathy towards an especially intelligent animal being killed; while humans have vastly richer communication and more advanced tool use, some animals can have mental and emotional processes not that much different than us. In that case, measuring suffering is more difficult and sometimes leads to surprising conclusions when you do the math. (I said a couple days ago [1] that a reasonable metric for suffering is the number of cortical neurons possessed by the victim, see also [2]). People can disagree about the scale of the suffering or what behaviors are appropriate in light of this problem, but it's easy enough to consider.
Of course, plants, lacking a brain or any neurons at all, are undefined by this criteria. And it's really difficult for me to anthropomorphize a carrot.
[1] https://news.ycombinator.com/item?id=27492230
[2] https://astralcodexten.substack.com/p/moral-costs-of-chicken...