Why are we going through all this trouble to make something that in the best case is not worse than existing IDEs?

I can't help but feel that a visualization window + a cacheResult(data, fileName) function would be superior to what Jupyter can ever hope to be