They want the same thing as every other organism wants - maximal exploitation of a niche by a lineage. Each adaptation that survives overwhelmingly tends toward advantage in the exploitation of a niche - fending off predation, establishing control over resources, symbiotic support, parasitic drain, and a myriad other capabilities that are highly environment dependent.

Just look at antelope in north america - they evolved incredible speed and agility in order to outrun and evade megafauna predators, but there's nothing left nearly fast enough to be a threat to them. Environments can change, and leave an organism with features that are no longer necessary or even beneficial in terms of overall quality of life and energy efficiency. The slightest noise can disturb a herd of antelope into bolting as if there were prairie lions or sabertooth tigers on the prowl. They don't need to be hypervigilant in the same way, and it burns a lot of calories to move the way they do, so whitetail deer and other slower species that aren't quite as reactive or fast are better at exploiting the ecosystem as it is.

With mushrooms that have mysterious chemistry, there will be a lot of those sorts of vestigial features. Extinct species of insects and animals and plants will have been the target of specific features, or they might end up in novel environments where other features are particularly suitable, but some become completely counterproductive in practice.

As far as psilocybe mushrooms go, in lower quantities, they actually provide a cognitive advantage sufficient to make a symbiotic relationship plausible between mammals and the mushrooms, albeit indirect. Animals under low levels of psilocybin influence have better spatial perception, can better spot movement in low light conditions, and there's a slight reduction in the neural influence of trauma inspired networks. Large quantities can be beneficial in a number of abstract ways. Any animal that sought those mushrooms out could thereby gain adaptive advantage over competitors that didn't partake.

Having an extremely toxic substance might be useful for killing large organisms and their decomposition either feeding the fungi directly, or feeding the organisms beneficial to the fungi. This can be plants, other fungi, or the feces of scavengers. Horizontal transfer might occur if there's an initial beneficial relationship, animals like the smell and taste of a thing, and then the fungi picks up the killing poison, and the consequences are sufficiently beneficial to outbreed the safe ones.

If too many become deadly, animals get killed off, and the non-deadly ones tend to gain the upper ground, since they aren't spending any resources on producing any poisons. Where there's a balance of intermittent similar but poisonous mushrooms, they take down enough animals to optimize their niche.

There are dozens of such indirect webs of influences and consequences that spread from seemingly simple adaptations, and it's amazing that things seem so balanced and stable as they do. It's a constant arms race of attacks and temptations and strategies.