4 ms·
Eventually advertising will either 1.Merge with content: a single jpg/png render of a webpage + imagemaps for links. 2.Use canvas and emulate the browser outpu
by FrozenVoid 9y ago
Eventually advertising will either
1.Merge with content: a single jpg/png render of a webpage + imagemaps for links.
2.Use canvas and emulate the browser output in it(webkit.js which creates a browser-within-browser contraption)
3.Require loading ads to function: some already merge functionality with ad javascripts.
Since there is commercial pressure to have working ads, one of the methods will become commonplace eventually and will be really hard to filter(if at all possible).
For #1 machine learning to recognize what an ad looks like and replace it with black rectangles or if imagemap block its coords, for #2/#3 there will be some sort of javascript filter/hijacker that would remove the ad code from content/emulated browser. But with WebAssembly these methods will not works, as content will become obfuscated binary stream which has to be decompiled& deobfuscated to even begin searching for offending "javascript"(essentially something like a virus pattern in binary).
- eddyg 9y agoThe prospects of everything "bad" that will be done in the future via WebAssembly makes me cringe. It seems like it could be a major tipping point.
- orangecat 9y agoWebAssembly bytecode is relatively simple; it shouldn't be any worse than dealing with obfuscated JS today.
- FrozenVoid 9y agoIt isn't a single javascript can map to many bytecodes streams. i.e. inserting dummy instructions, alternative code paths, etc, all the tricks malware uses to escape detection.
- adrianN 9y agoEventually we'll find a better way of paying for online content than with our attention or ads will become promoted content that is actually interesting to people. I don't think that advertisers can win an arms race with users. Eventually ads will be annoying enough that people look for different content, like many people do with TV today.
- majewsky 9y agoAnother alternative adblocker implementation: 1. Browser window renders into a buffer instead of on screen (which it already does anyway because of how compositing window managers work). 2. Ad blocker grabs the image from the buffer, runs OCR, re-renders the useful content like in Reader Mode. 3. The final window is the browser chrome with the re-rendered content in it. So we're back at the arms race part. I'll much rather expect more Native Advertising. This would need much more than image recognition to block. (I mean, there are plenty of actual humans that cannot distinguish native advertising from legitimate content.)