>Breaking adblockers is a technically solved problem; just serve web pages as server-rendered blobs.

Can you elaborate on this a little? How does this win the ad-block arms race?

If your webpage is a binary blob - for example, several megabytes of webassembly that unpacks an encrypted payload onto a randomly-generated DOM - then adblockers would have to disassemble the thing and run heuristics to figure out which part of the page is ads and which are innocent images and text.

Which is impossible to do without horrible false-positives. The arms race between ever-increasing heuristics complexity on the adblocker side and the webpage cloaking ads as content is a race that the adblockers cannot win.

Ok. So, at that point, you are practically downloading a binary executable, which uses browser API's as it standard lib, per site.

You're right I think, it would become very difficult to create client side heuristics to identify the advertisement elements in the way it is currently done.

But who knows? Maybe machine learning based solutions will become necessary? Maybe dedicated recognition FPGAs?

The war will probably continue for some time.

Not really, because a) people aren't willing to put up with false positives (i.e., adblockers hiding the meme cat photos you were looking for), and b) people aren't willing to pay money to hide ads.

The real solution is getting rid of intrusive and/or annoying ads, but that would mean sites with tightly curated audiences; no clickbait, SERP spam and other stupid audience-inflating crap.

Ultimately intrusive ads happen because publishers are in a rush to monetize traffic that advertisers view as cheap/low quality. It becomes a downwards spiral of attract low-quality eyeballs -> ad spendings plummet -> devote more screen share to annoying ads -> viewership falls -> attract more low-quality eyeballs -> etc.

Maybe at some point we'll hit rock bottom and things will start improving.

> a) people aren't willing to put up with false positives (i.e., adblockers hiding the meme cat photos you were looking for)

You can estimate the actual ratio of false positives by examining existing automation solutions based on image recognition. As a very rudimentary example, I have successfully used Sikuli [1] for various tasks including a program running "binary blobs". Unless you mix contents and ads altogether (and I personally think that it is fine as long as the ad is properly marked! It is at least a kind of contents, just that what I don't want generally though.) you can't ramp the false positive ratio up beyond a certain level.

[1] http://www.sikuli.org/

Well, this assusmes that the ads are served from the origin site, not some other domain right?