I think the essay has the implicit assumption that if you're 'better' at doing things, you'll end up happier. Since intelligent people aren't happier, the conclusion becomes that intelligence is mostly only relevant to a subset of well-defined abstract problems that don't represent the real messy problems of life.

I think the reality of human happiness is that being in better material circumstances, or being better at solving problems practical or not, doesn't necessarily improve long term happiness.

Adaption to circumstances, the 'hedonic treadmill', means most people aren't going to obtain some very high level of happiness and sustain it indefinitely, regardless of station.

Doesn't really make sense as there are already open weight models that are close to frontier models, and the compute to run them isn't outrageously expensive, and every indication that it will only get cheaper over time as has been the strong trend in terms of $/outcome.