The reasoning treats a correlated characteristic as a causal mechanism.

successful person → unusual trait And infer: unusual trait → success

The most popular example of this in tech (that never seems to die) is when people notice moments where Steve Jobs was an asshole, or he said no directly to customers, and infer they need to be more like this because it's the unusual trait they're missing and need to emulate.

FDR's hatred was a byproduct of his consequential actions. But consequential actions are not the only things that produce hatred. In fact, rather petty actions can cause someone to feel hatred.

If you use hatred as a proxy for importance, you are effectively saying: "All impactful people are hated, therefore all hated people are impactful." This is logically equivalent to saying "All dogs are animals, therefore all animals are dogs." The metric has zero predictive power because the set of "hated people" is vastly larger than the set of "impactful people."

Friction is also terrible metric for progress. I would argue privacy has actually gotten worse due to the banners because if you decide you're not going to sign in to do something like a Google search (so it's not tied to your account), then you're immediately punished with a nag box. So you actually decide you'd rather stay signed in so you don't get nagged. Even if you're in favour of using the government, you should be using friction as a counter-signal. For example, switching to the Euro, reduced friction. Standardising to USB-C, you could argue this reduces friction for consumers.

TFA talks exclusively about being hated by the right people. If you are only hated by the right people (yes, I made an adjustment there), it's probalby not because you are a jerk. Come on.

google naggs you exactly once.