Hormesis is heavily debated. But just in general, I'm not a fan of "what doesn't kill you makes you stronger." What doesn't kill you generally makes you very sick.
Epidemiology didn't prove smoking causes cancer, btw. It cannot. But the study in question that I presume you're referring to, Doll and Hill's, had a very significant relative risk finding:
It showed a 45 fold increase for men smoking more than 35 cigarettes a day. In comparison, our favourite study of colorectal cancer by IARC had a RR of 1.18. In the same figures, that's a RR of 45 for smoking vs RR of 1.18 for processed meat. In percentages, that is 4500% vs 18%. And yet this is the most significant epidemiology that we have for studies in nutrition.
An interesting aside is that Austin Bradford Hill, the Hill in "Doll and Hill", is also the person responsible for the Bradford Hill criteria . Generally, anything interesting worth investigating should have a RR of > 2. I don't reject epidemiology, btw, I think it's a good way to generate a hypothesis. But it cannot prove causality and isn't science on its own.
You are free to draw your own conclusions, of course.