There are some specialized applications (e.g. RTGs for space probes) where nuclear power is still very useful. I agree that “regular” NPPs aren’t that great anymore.

There is an arguable case for fast breeders as part of a long-term waste management strategy. That one works by breeding high level waste into a more radioactive form that will take a mere 200 years to decay to the level of natural uranium ore as opposed to 20000 years.

The upside: Disposal is doable with technology that exists today as opposed to technology we may at one point possess in the future. We also don’t need to design facilities that last longer than all of recorded history. We don’t need much beyond fast breeders and a few guarded and well-maintained warehouses.

The downside: It still involves guarding and maintaining warehouses full of extremely dangerous high level waste for 200 years and breeders inherently pose a nuclear proliferation risk.

It’s by no means a panacea but as one of the very few feasible ideas for nuclear waste management I think they’re at least worth talking about.