Intelligence is clearly useful for almost everything. And that's
exactly my point. The closer we look the more intelligence we find in nature. But there's a tremendous qualitative gap between human-type intelligence and everything else. When we believed other animals were dumb we were always underestimating the complexity of our own intelligence. What has been relatively clear, however, is that intelligence everywhere else in the animal kingdom is mediated by the lack of systemic, non-kin cooperation, because higher-orders of intelligence increasingly require the cooperation of other individuals to operationalize.[1]
Squid and octopus have jaw-dropping intelligence, clearly aimed at food acquisition. But they only mate once and then die, and so long as that's the case it's clear why cooperative behavior will never arise and why their intelligence maxes out at whatever is useful for the individual alone. (Note, however, such intelligence my incidentally promote limited types of cooperation. You have to be careful to distinguish the selection pressures to understand the limitations.)
Gorillas may learn how to disarm traps, but multiple gorilla troops are never go to learn to gang up together for their common defense and well-being such as by actively teaching other troops how to disarm traps, at least not in an evolutionarily meaningful sense, even though from our perspective it would be advantageous for the species as a whole. There's no genetic pressure for that. The genes that might promote non-kin cooperation are countered by genes for selfishness or laziness.
The fundamental dilemma is that benefits of cooperation partially inure to non-cooperators. If one gorilla disarms a trap then every gorilla benefits to some extent; to the extent they benefit then selection pressure is diminished for the trap-disarming intelligence trait to propagate as it's unnecessary to have the gene to reap those partial rewards. So what happens is that the species reaches an equilibrium according to how the costs & benefits are allocated.
Because of how genetics works, both the costs & benefits of cooperation and intelligence are moderated by degrees of kinship. The equilibrium reached--the selection pressure for stable, multi-generational cooperative intelligence--will always be a strict function of kinship. At least until a species can passthrough the bottleneck that humans did, after which things clearly get more complex. I don't know the mathematical function, but AFAIU game theory shows that selection pressures for cooperation basically drop off a cliff after the first or second generation.
Again, evolution doesn't have foresight nor intelligence of its own; it can't preference one trait over another simply because one trait leads to greater reproductive success down the road, or great success for the species as a whole. It can't disfavor genes for cheating or laziness as long as those strategies are equivalently successful for each individual's reproductive success considered in isolation.
Being able to disarm a simple trap clearly has tremendous benefit to the individual and the kin in his troop, no matter that almost every other gorilla in the forest benefits to a substantially similar degree regardless of kinship. So no surprise gorillas can do this, given other factors (e.g. in the abstract, slow rates of reproduction mean there's greater pressure for smarter individuals; whereas for species with high rates of reproduction other strategies will predominate). But legions of gorilla troops aren't going to start systematically disarming traps in the forest. Yes, they could if they spontaneously developed the certain types of intelligence, like complex communication and organizational behaviors, but those types of intelligence are predicated on being able to rely on a high degree of selfless behavior of non-kin, and that is unlikely to happen. As far as we know, in the billions of years of life on earth that qualitative kind of cooperation has only happened once. (Our degree of intelligence has also only happened once. That's unlikely to be a coincidence. My argument is that the former preceded the latter, at least at our branching point. Afterward intelligence may have driven greater cooperation in a virtuous cycle, but we shouldn't have much confidence that we can understand that process without understanding the genesis.)
My overarching point is that we can easily predict the types of intelligence exhibited in the article using basic, high-school level genetic theory. We should expect many more fascinating examples like this from gorillas, chimpanzees, orcas, dolphins, etc. And we can easily predict the limits of such intelligence similarly. Weird theories about sex or thumbs or w'ever, however, cannot explain how a species transitions to comparatively selfless, non-kin cooperation because they can't explain the mechanics of the selection pressures within the framework of Darwinian genetics; and because they can't do that they can't explain how a species is put on the path to human-level abstract intelligence.
[1] That the social groups of bonobos and naked mole-rates are organized around sisters is something that many research papers gloss over and often fail to mention at all. I actually had to dig through the literature to confirm this. (I'm not a biologist, FWIW.) Note this is distinct from a matriarchy. All that matters is that the seemingly atypical degree of social cooperation is most easily explainable by genetic relatedness. Occam's Razor says that Bonobo's non-reproductive sexual behaviors, for example, are ancillary to the core dynamic driving their cooperative social structure, rather than being a driver. And so analogizing Bonobo sex with human sex doesn't actually teach us anything; it doesn't suggest that social sex promotes cooperation, just that cooperative behaviors may correlate with social sex, which should be obvious.