I follow a more traditional literature-search method, though it's augmented with online sources.
I start with a question. Mine initially was "What are the Big Problems?" I then dive recursively into that.
Starting reading at virtually any point, you'll discover that knowledge is a web, and that the best authorities reference others. Then the spidering begins.
Follow an author's references. If you find a quip or fact or quote or reference which seems especially germain, then look it up. Given today's Internet, this oftem means you can have a specific reference in front of you in seconds. I had this experience re-reading James Burke's Connections a few years ago, in which he mentioned Agricola's De Re Metallica (which is not about the band), and found that the English translation of this 16th century work (not completed until the early 20th century) was at the Internet Archive. (The translators also have interesting biographies.)
If you find an especially good source, look to see what other sources reference it. That is, look for citations. This is slightly less powerful than the first method, but 1) serves as a check to see what works are truly significant (they'll have high citation counts) and 2) will lead to more current treatments of a concept (which aren't always better or improvments, mind).
Those are the two principle methods.
Once I find an author or topic (subject heading or keywords) of interest, I'll use a traditional catalogue, almost always Worldcat, to look for additional materials. If you find an author of interest, this is a good way of finding their other works. Worldcat indexes both books and articles.
https://worldcat.org/ DDG bang search !worldcat 'au:' == author, 'ti:' == title, 'kw:' == keyword
I don't have a good catalogue for popular magazine or newspaper articles, though there are several commercial options. Some libraries (public, community college) will provide access to these. Google Books captures some of this material, at least for searching.
Google Scholar, Archive.org, Open Library, LibGen, and ZLib are also useful for both searching and sourcing documents.
General Web Search has become all but useless over the past 5 years or so.
Finding an idea, especially one that seems to be universally accepted and unquestioned, and seeking out its source can be profoundly interesting. Google's Ngram Viewer is your principle tool here, as you can see specifically when a specific word or phrase (up to five words) emerges. Quite often "accepted wisdom" is found to emerge with very little empirical foundation. It can be tricky to identify where the breakout occurs and through what work, but this approach seems to work better than others.
Online sources are another option, though what I increasingly find is that more-recent online content tends strongly toward lower value, and less use of these is better. This depends greatly on the field. Among the best options is to not read the current submissions, but to do a specific search for top items within some time bound.
On HN, you can effectively see the top submissions from the past week, month, or year. I've addressed that here:
https://news.ycombinator.com/item?id=28806795
Other sites, notably Reddit, have similar date-bounded search options. Incidentally, if you're assessing whether or not a subreddit is worth subscribing to, reviewing its top posts by week / month / year is useful.
In general, I find that identifying a good author or publication, and "stalking" their output, is superior to virtually any user-generated content site (FB, Reddit, Twitter, HN, etc.).
Books and articles have higher hurdles to publication than online articles do. The Internet's editorlessness is becoming more of an obstacle than a benefit as there is simply so much crap online.
Track your references. Zotero seems to be the gold standard here, though I don't use it myself. Calibre has its uses. Avoid Mendelay like the plauge it is.
Consider a Zettelkasten or equivalent. I'm referring to pen-on-paper index cards as the most robust option here, though there are digital versions. Of these, I'd strongly recommend Emacs org-mode or a flatfile ASCII / UTF-8 reference as the most robust, possibly a wiki. The simpler and more robust this is, the better, as it will quite possibly last your entire life. The problem with hot new software is it often does not.
I strongly recommend an e-book reader or tablet with the absolute most onboard storage you can manage. I'm pretty happy with the Onyx BOOX line, and have their largest device, the 13.3" Max Lumi. It's been updated recently to 128 GB onboard storage (mine is 64 GB, and I'm bumping up into that), and I'd prefer that were bumped to 1 TB (some Apple iPads reach this).
I strongly prefer e-ink to emissive displays.
For size, 6" is about the minimum size you should consider, 8" is comfortable for most straight text or e-pubs, 10--13" is much better for scanned-in PDFs of older works and articles in particular. (That was my thinking in buying the Max Lumi, and it's largely been validated.)
My usual problem isn't to little to read, but far too much, and setting (and sticking to) priorities on that.