You may have read the article before it was updated, but there is a note and a couple of comments addressing this. Basically this is a somewhat simplified model, and the papers on which it is based view things more complex then impulsivness = hard-wired:
An important note is that the full version of TMT places a constant in the denominator to prevent the denominator from skyrocketing into infinity as delay approaches 0. Also, 'impulsiveness' here is a substitute for 'susceptibility to delay,' something which may vary by task, whereas 'impulsiveness' sounds like a stable character trait that might not help to explain having different motivations to perform different tasks.
An important note is that the full version of TMT places a constant in the denominator to prevent the denominator from skyrocketing into infinity as delay approaches 0. Also, 'impulsiveness' here is a substitute for 'susceptibility to delay,' something which may vary by task, whereas 'impulsiveness' sounds like a stable character trait that might not help to explain having different motivations to perform different tasks.