Code that reads almost like plain English — the trade Python makes is raw execution speed for how fast a human can read and change what’s already written.
Guido van Rossum began Python as a hobby project over the 1989 Christmas holidays, designing it explicitly to prioritise readability — enforced indentation instead of braces, a small set of orthogonal features, “there should be one obvious way to do it.” That design bet paid off in a domain nobody was optimising for at the time: scientific computing, where researchers who aren’t primarily software engineers need to write and understand code quickly. SciPy, the foundational open-source scientific-computing library built on Python, has been cited over 39,000 times — a rare case of a piece of software itself becoming one of the most-cited works in its field.
That’s also the specific reason Python, not a faster compiled language, became the substrate the current wave of machine learning and AI tooling was built on: the bottleneck in ML research is usually a researcher’s time, not the CPU’s.
🔗 Related Concepts
Source: Virtanen, P. et al. “SciPy 1.0: fundamental algorithms for scientific computing in Python,” Nature Methods, 2020.