Open source is a wonderful thing, but AI is exposing its weaknesses
Open source is the best way to create software: when a project brings together a community of developers actively working on it, documenting it and constantly improving it, we get a virtuous circle which, among other things, fulfills the Linus’s Law: “given enough eyeballs, all bugs are shallow”.
When a certain number of people, diverse and not belonging to a single organization, coordinate to develop a software project, the results are usually good. Much of the software we commonly use is based on open source projects, simply because it is the most appropriate way to develop software. Simply the fact of wanting to make software available in open source form obliges its creators, at the very least, to review it thoroughly and document it properly, something that few developers like to do, but which is essential if you want it to be inspected, understood and improved by others. The idea, moreover, of creating software so that anyone can benefit from it offers a glimpse of how society should work.
However, a recent study based on more than 1.2 million open source software projects found a worrying conclusion: only about 11% of them undergo minimal maintenance. The concern arises, logically, from what tends to happen when software, which is anything but isolated, stable and monolithic, is not updated…
