I was recently thinking about going back to the idea of Butter (https://butter.sonnet.io) and expanding it a little bit. Can we build a proof-of-concept of a universal ad blockerâ„¢ running on your phone? It uses AI (it's very fancy):
- model 1: use screen capture to detect ads
- model 2: once an ad is detected, hide it in a context specific way (e.g. if it's the native YT app, find and click the skip button, if it's a podcast, seek past the sponsored segment etc...)
both model 1 and model 2 gradually learn, perhaps share knowledge between users, in a manner similar to blocklists.
Perhaps in 3-5 years our devices (mobile G/CPUs and batteries) will get powerful enough to support this. Imagine the possibilities!
It's a ridiculous idea. I also think it's possible we'll see something like this in the future, because it feels like a simpler solution than dealing with the systemic issues related to advertising.
I wonder how this compares to the manufacture of the things being advertised.
Wait until you see how much physical advertising contributes. :D
Especially if you consider branding, which I don't know why it wouldn't be in this, it is tough to compete with shenanigans like how merchandising for the losing side of stuff like the super bowl are dealt with.
This doesn't pass a cursory sniff test: Web traffic is 17% of Internet traffic, so even if we accept 25% of that is ads (I'm dubious) that would be 4.5% of traffic. That would add up to 0.3% of emissions.
https://www.statista.com/statistics/271735/internet-traffic-...