Maybe the point at which diminishing marginal returns really kick in, and "benefit"... with respect to what metric are worthwhile factors to consider here. The graph shows mortality rate but perhaps other cardiovascular disease markers would be meaningful to look at here.

In general, it would be nice for one to know some correlation between how the number of steps taken reduces one's risk. From what the article suggests, presumably these would be non-linear relationships.

Well, either you gain immortality or the effect plateaus off.

But seriously, to my understanding, if you walk just because you want to stay healthy, statistically 10k is a good baseline.

10k steps is about 6km give or take.