We Don't Depend on Google (or Search Engines in General)
Most projects and sites become extinct within way less than a decade
Last year, culminating in October/November (we got interrupted by a trial where we were unrepresented), we finalised and took live our own 'search engine', which not only helps visitors but also helps us, the team which runs the site and must sometimes research and find old articles. One cannot rely on GAFAM to index all the pages (the sister site is truly massive and spans a 22-year period). Yesterday the sister site served 605,400 Web requests and this site served 540,568 without any reliance on search engines (a lot of visitors come owing to RSS feeds; they're regulars).
But there's a lesson here and it extends beyond sites. Relying on third parties - be these parties social control networks or search engines - is not wise and definitely, inevitably, ultimately unsustainable. YouTubers learned the hard way that outsourcing to "free hosting" means penalties, shadowbans, "strikes" etc. Trying so hard to appease Google (Alphabet), they lose sight of who they are and end up brown-nosing not only GAFAM but also its biggest advertisers (who include some highly despicable politicians).
In software, we've become accustomed to the same trajectory of surrender. CoCs, opinionated licences, and capitulations to online mobs can transform a project, even if the project itself doesn't have those issues, only modules or communities it depends on do. Problems can be inherited and difficulties can be imported. Historically, judging by whatever projects survived 50+ years, more vibrant communities (robust projects) have less dependencies on third parties, they're basically far more independent and free-thinking. Projects such as WordPress are perishing, a prison like GitHub is losing users, and famous brands like Linux with Rust invite endless controversy because they become slop-ware full of headers and and code nobody wrote (machine-generated slop, barely audited, expecting others to waste time auditing despite it being authorless plagiarism).
If Linux carries on adopting code that 90% of Linux developers cannot read and won't bother reading, it'll just be a massive bag of "legacy code" (like Windows). No reason to be an apologist for that sort of strategy.
For Linux, the slop won't improve security (fuzzing disguised as "AI" is mostly a time-wasting, energy-wasting yoke that discourages human contributors). Also "simple is more secure," as an associate reminds us, as "if it is not there, it cannot break."
Moreover, "code which nobody on site understands is by definition (or in practice) legacy code and thus a liability and, again, complexity is a liability and since more code equals more complexity, code is itself a liability [so] K.I.S.S." (Keep It Simple, Stupid)
People have already begun forking Linux as a project with just "15k lines of code for the kernel" and no dependence on "modern things" (except they add Rust as a dependency, which is self-defeating), set aside the lousy plagiarism engines. █
Image source: American Museum Of Natural History