Does nuclear power have a negative learning curve?

Nuclear has what Elon Musk calls a "high PITA factor" where PITA stands for "Pain In The Ass."

A tech with a high PITA factor is one that looks simple and great on paper, but when you actually get into implementation and get a few iterations in you hit a combinatorial explosion of hidden complexity. Elon used this term to describe liquid hydrogen rockets in explaining why SpaceX instead went with kerosene for Falcon and is using methane for future rockets.

The reasons behind the high PITA factor for hydrogen rockets and nuclear power are similar in that they're both rooted in extreme difficulties in handling the materials. Hydrogen embrittles many metal alloys, likes to leak, and is a colorless odorless invisible explosion hazard. All that makes it much more expensive to handle than RP-1 or CH4. Likewise nuclear (fission) materials are nasty and metal-eating (via neutrons) and hard to handle, especially once subjected to a fission chain reaction. If something goes wrong you end up with a lump of hot mixed super-radioactive isotopes called "corium," famously described as "the nastiest crap in hell." The whole process demands explicit engineering of each step and is incredibly unforgiving.

This article makes kind of a novel point though: that high PITA factor technologies might indeed exhibit what looks like a negative learning curve. The learning curve isn't actually negative. It's just that what's being learned is mostly in the "oh shit this is way harder than we thought" category rather than the "wow! another way to make this more economical!" category.

I don't mean to assert that nuclear (fission) can't be "solved," but I don't think anyone has done it yet. Fusion might be different but we don't know yet.

There isn't any real way to contain or process nuclear waste (most of it is literally kept in swimming pools) and this article fries nuclear economics without even taking waste into account. And don't forget decommissioning, another waste process that costs billions and every plant eventually requires (at ratepayers' expense).