The author more than understands this. In the first two paragraphs he address the much more interesting question about whether, in realistic causal DAGs, potent

The author more than understands this. In the first two paragraphs he address the much more interesting question about whether, in realistic causal DAGs, potential correlations grow at a faster rate than actual causal links—the idea being that if they did it would harm the notion that seeing a correlation improves the hope of causation due to pigeonholing. The author doesn't assert that this is true, but it definitely means he's thinking about this quite hard.