I concluded a long time ago that most of the content of a modern paper is simply unnecessary. The joke often goes: "I only read the abstract and conclusions, a

I concluded a long time ago that most of the content of a modern paper is simply unnecessary. The joke often goes: "I only read the abstract and conclusions, and if they agree with mine, I check the methods to see if there's any trick I haven't learned yet". In many cases, a paper I read could be replaced with a decent github repo to hold a README.md that explains what the repo does, source code that implements the scientific discovery (does not apply to lab experiments) and an S3 bucket to hold the raw data. A unit test/system test should ensure that the code builds periodically (many codes fail over time as dependencies change) and passes basic statistical tests; ideally, it would produce the entire final data set and the analysis. in the meantime, I just ask Gemini to do that step for me. It does a remarkable job turning papers into code.