Follow
fuchsiaaaaaaaaaaaaaaaaa
is there like a concept that describes the pitfalls of ascribing improvements to a specific innovation, while it's really due to smaller changes that could've much better been implemented in the existing system?
@f0x a few years ago I worked on a project to try to port a computationally bound program to the multi-processor Phi. The changes that we made to improve the vectorization of the code to perform well on the Phi also improved performance on the normal x86, and we ended up making it so much faster that we abandonded the port.
someone noticed this once in the companies that tout blockchain as a great improvement, while the actual change was due to finally using a better specified data model, which would've worked even better without any blockchain garbage