I think a fragile world hypothesis is more interesting and scary.

Instead of a superintelligence killing any emerging intelligence, superintelligence is so smart that it can perturb spacetime in ways that easily kill itself.

Kind of like the idea that LLMs will give anyone with a garage the ability to create killer viruses (wild claim), a superintelligence might trivially create black holes that engulf it. If the intelligence is distributed across many agents (or nations), someone will try something stupid that kills the entire population.

It'd be even scarier if they could accidentally collapse the supposed vacuum state. Universal annihilation at the speed of light just because you got too smart.

Fragile world / fragile universe meets "too smart for your own good".

Leaving aside the fact that there's no evidence superintelligence is even possible, no amount of intelligence alone will let you "trivially" create black holes. Creating black holes requires a ludicrous amount of energy. Planet-scale energy, maybe solar-scale energy. Energy is conserved. You can't just think really hard and have it. At minimum you have to already have the energy and materials to build a machine that will let you (mostly build another machine that lets you) harness that much energy.

It's at least not something you'll do by accident.

Sounds logical to me. It's already remarkable that we haven't done this with fission weapons. It's not unreasonable to assume that within a century we'll have antimatter weapons. Destructive potential only goes up.