So I have a contentious one. Quantum computers. (I am actually a physicist, and specialised in qunatum back in uni days, but now work mainly in in medical and nuclear physics.)
Most of the "working": quantum computers are experiments where the outcome has already been decided and the factoring they do can be performed on 8 bit computers or even a dog.
https://eprint.iacr.org/2025/1237.pdf "Replication of Quantum Factorisation Records with an
8-bit Home Computer, an Abacus, and a Dog"
This paper is a hilarious explanation of the tricks being pulled to get published. But then again, it is a nascent technology, and like fusion, I believe it will one day be a world changing technology, but in it's current state is a failure on account of the bullshittery being published. Then again such publications are still useful in the grand scheme of developing the technology, hence why the article I cited is good humoured but still making the point that we need to improve our standards. Plus who doesnt like it when an article includes dogs.
Anyway, my point is, some technologies will be constant failures, but that doesn't mean we should stop.
A cure for cancer is a perfect example. Research has been going on for a century and cumulatively amassed 100s of billions of dollars of funding. It has failed constantly to find a cure, but our understanding of the disease, treatment, how to conduct research, and prevention have all massively increased.