Nothing helps the information overload, we hoard and hoard, but the models will have all of that in their training in a few months time, now you can refer back to things that are just at the tip of the tongue, but can't quite recall the name fully, let the AI fill in the blanks.
I collected so many interesting things, blog posts, articles, papers, I learned about many authorities in their respective fields because of that, and mapped who's who, which has been valuable. But the AI itself is so convenient, and search engines (and the state of the Web) sucks so much now, that the time saved just by starting from an AI with search capabilities is worth it.
I'm downsizing my personal wiki, keeping my zotero collection only with the best or most serious papers, culling the small effect size / speculative stuff.
For fun stuff, hobbies and nostalgia, blogs are still valuable. Youtube is better than ever. But as far as huge note systems, nope. The only thing that stays growing so far is the browser bookmarks collection, separated by YEAR-SEMESTER, then topics inside. Topics that stay permanent staples in my life, not one-off projects or casual dives, go outside that hierarchy, into folders with actual categories.
Maybe the only thing I'm missing to sort and keep checking the frequently, high signal blogs and news sites, is some sort of Netvibes revival clone. The RSS reader with its typical e-mail layout doesn't cut it for me. Netvibes' layout was great because it was a table grid for each feed.
And as far as joining all of that into one super dashboard or app, I couldn't come up with anything after years, and every attempt at syncing collections of data from different domains in a unified interface, meant simply more clutter in the long run. Keeping it simple worked for years.