There are two comments worth making here.
1. The author is right that the effects of fossil fuel burning are non-linear. What that means is that there is no such thing as the "cost of a ton of CO2". Because the effects of carbon dioxide grow more than linearly, the cost of the ton of CO2 will continue to grow all the time. So the only reasonable cost of a ton of CO2 is what it will take to remove it. Optimistic estimates put that at 150 to 200 dollars. If you do the math, that means the current US subsidy to the fossil fuel people is about a trillion dollars a year.
2. There is another way to think about the non-linearity that's even simpler than what's described here. Look at the beaufort scale of hurricanes. With each linear increase in wind speed, the damage description makes the previous level of damage seem negligable. You’re that much farther from the world things were built for.
The same kind of relationship holds for almost any kind of climate damage you can think of. Sea level rise first affects marginal districts but then more and more of mainstream society. Droughts first affect marginal areas and gradually more and more of the breadbasket. Health threats first affect the most vulnerable but eventually everyone.
Accelerating costs are the rule, not the exception.