A quarter of the deep links in The New York Times’ articles are now “rotten”, or no longer accessible. The older the web page, the more likely it is that the articles it links to no longer exist. This chart makes it clear:
The internet is decentralised by design. That means no single entity decides whether a given article on the web is taken down.
But that also means that no single entity can ensure that that article can stay up. If the owner of the domain dies, forgets to renew, or simply chooses not to, it’s gone. The Internet archive can’t archive every single web page that ever existed.
That means it is up to each of us to preserve, privately, those parts of the web that matter to each of us.
I am personally a long-time user of both Instapaper and Pocket (from when they were personal projects of their creators), and have thousands of articles in each. Should either of these services shut down, I will be able to export my saved articles. For articles and web pages with more significant personal value, I also have a folder full of markdown-formatted versions of them. I ended up creating an iOS Siri Shortcut to automate this, which I use every day.
Other ways are to save the full text in Evernote, or OneNote, or Notion using their browser extensions, and they’ll be available to you as long as these services are active. You could also copy the web page, paste it in an email and mail it to yourself, creating a library within email. Which again is accessible – and searchable! – as long as you have access to that email address. There’s no perfect solution.
The important take-away here is that what makes the Internet resilient as a whole makes it fragile at a microscopic level. Saving bookmarks alone is no guarantee that you’ll be able to access something on the web later. You’ll need to save the page itself, and find a system for this that works for you.