It's blocking almost nothing, and many of the few videos it does block are videos from channels I actually subscribe to which are made by actual humans. So I tried to get a sense of how its blocklist is compiled, and basically it looks like anyone can submit something to be blocked. Which seems like it could be gamed, but it has assurances that it goes through a strict review process. Except I click through to the blocklist on github and it has updates every few minutes, round the clock. Who are all these people rigorously reviewing these hundreds or thousands of submissions a day?

There's other things too. The extension asks for permission access my user data at github. Why? I get that it needs to access github for its list, but it has no reason to know *who I am* on github.

And if you look too hard at the website you notice sketchy stuff, like it says it's MIT licensed, but you click through, and there's no source code in that repo, just the blocklist.

So, while the idea is nice, this extension is super sketchy and doesn't work for me.

Their website is clearly AI, a little ironic. The actual code also seems to be generated by AI, or at least with AI assistance. If you go to their github you can see some comments reminiscent of AI code.

Claims to be open source, but has not updated the corresponding repo in 11 month as of writing. Repo currently contains 0 code, but the block list is in a github repository too which is constantly updated (which makes no sense from a design standpoint, you don't store that kind of stuff in a repo) so you know that the author knows how to use it and it isn't *just* incompetence.

Author, why are you hiding the code? You know that anyone can just paste "about:support" in the address bar, open the Profile Folder, navigate to the extensions subfolder, and then open "ai-channel-blocker@aiblock.extension.xpi" in 7zip to find your unobfuscated code. Its all there plain to see, just put it in the Github repo. What are you trying to hide, that there are signs of AI usage to write the code in the first place (among other things mixed comment styles among all the files including in sync.js where you have a suspicious TODO line to "replace with your actual GitHub repository URL", who is "your" talking to here, the chatbot to you)? That you are using github to store content instead of loading it from a more proper place? That this is all just hastily thrown together? none of that matters, just throw the code up on github, its already accessible.

I love the idea of this and hoped someone would create it, but i have massive reservations.

1. The website is AI generated... ironically
2. The Extension is not OpenSource, yet someone advertising this Extension on reddit and claiming to be the developer has promised to make it open source 7 months ago.
3. The development workflow seems unstructured; bugs are reported on multiple platforms, (mostly discord, sometimes reddit) even though the core functionality of this extension (reporting ai-content) is primarily done and managed through github, somehow.