While the general thrust of this comment is accurate, especially if we only consider relatively recent 4chan, it does miss some of the subtleties of variation in anonymous-imageboard discourse quality. There are two primary differentiators between absolute cesspool and actually fairly decent boards, one of which the parent hints at: size and quality of moderation.
First, size. Although I can't speak to "modern" (say, post-2012) 4chan due to lack of firsthand experience, prior to then many of the smaller boards were actually quite nice places for serious discussion of their subject matter. /out/ (outdoors), /m/ (mecha), /n/ (transportation), and even some of the text boards like /prog/ were all fairly high quality due to their strong, persistent board cultures that discouraged pointless shitposting and encouraged actual contribution and dialogue. In each case, the ratio of new contributors to old hands was low enough to allow these board cultures to perpetuate and maintain quality. For example, when a community is small enough that many participants will have personally witnessed the debunking of a set of claims, reposting the same crap will be minimally effective. This dynamic breaks down at size, and it's easy to see how increased traffic over the years leads to the erosion of communities that rely on their small size for their quality.
That's where the other factor comes into play: quality of moderation. Despite being much larger and faster than the aforementioned slow boards, /a/ was quite good for a long period of time, largely thanks to very aggressive moderation. Low tolerance for off-topic content (especially anything with the scent of /pol/ on it) enforced a strong culture of topical, relatively thoughtful discourse. As one of the handful of places where you could discuss anime way too seriously without being hampered by those troublesome Naruto-watching non-elitists, old-/a/ produced a mind-boggling amount of in-depth textual analysis, especially of popular and controversial works like the F/SN VN. This was in large part due to the differences between the predominant pseudoanonymous forum-based model and the anonymous discussion model that aimed to address the deficiencies of the former. For example, a blight of old forums was circlejerking around the opinions of established members. Some guy, possibly a mod, would have a pet theory or position and everyone else had damn well better fall in line because he's got 20k posts and/or will ban anyone who disagrees. Crap like that wasn't universal, but it was a major issue with a lot of forum communities. Minus the persistent egos, discussion was allowed to progress. I think we let nostalgia blind us to a lot of the problems that afflicted old forum communities - the reasons that led folks to anonymous imageboards in the first place. Eventually, though, the pattern of strong moderation that enabled high-quality anonymous discussion fell by the wayside.
What caused that, though? Having spoken with a variety of janitors over the years, the common thread is that the ability to trivially ban evade via mobile connection (especially ipv6) DHCP pools basically destroyed the mod team's ability to quash bad actors. Ban evasion was always a problem, but prior to the emergence of phoneposting it took a lot more effort and was easier to combat. Banning known proxies and tor exit nodes is a pain, but banning mobile carrier IP ranges is a non-starter. This led to a general sense of resignation, which in turn translated to less-aggressive moderation and a decline in quality.
What I'm really getting at is that discarding the lessons learned about what makes an anonymous discussion system good or not just because the overall experiment didn't pan out is unfortunate, and that all hope isn't necessarily lost. For instance, an anonymous community with a barrier to entry (say, a SA/metafilter-type entry fee, where your ability to post anonymously is contingent upon not getting your paid account banned) could address the issues of size and moderation while maintaining the freedom-of-expression benefits of the anonymous model. The hard part there is bootstrapping - nobody wants to pay to participate in a yet-unpopulated community whose moderation model/community culture isn't a guaranteed personal fit, but opening it up initially to develop a userbase exposes it to the Gab/Voat problem.