[bonoboTP]

This used to be very common but nowadays it's more and more expected to release the code with the paper. This makes it harder to obfuscate big parts of the method. Arguably, this also has lowered the bar for doing incremental research. Just get the code of the sota from github, implement a new feature, test it, if it's 1% better then publish it.

Previously there was a natural gatekeep ing effect that you had to be able to understand and translate to code all the math and equations in the paper and be familiar with the dark knowledge details that everyone in the field knows but isn't spelled out explicitly in the papers. Boring bits like how to exactly preprocess, normalize, clean the data, how to precisely do the evals etc.