Hacker News Reader

Comment by retpirato | Hacker News Reader

So Google somehow knows which apps block certain instances, which generally get reputation among other instances & are quickly blocked? That's not believable.

View on HN

Why isn't that believable?

How does Google know this unless they have some access to the databases of each of those sites/instances? Why would google have that kind of access?

Such apps tend to advertise that they block instances. Tusky, for example, blocks all Gab instances and says so right in the FAQ [1].

[1]: https://github.com/tuskyapp/faq bottom of the page

Couldn't this get done as part of the manual review of an app's source code? It seems like this wouldn't necessarily have to be automated

And right after that we can remove any FTP client that uses the FTP protocol to download content Google doesn't like. We should scan all apps that use a common, published protocol to make sure the protocol is not being used to consume objectionable content. /s

The app is not the service; the protocol is not the platform.

I think you might have misread my comment; I wasn't suggesting whether a course of action was correct or not, but just explaining how it could technically be feasible. I interpreted the comment I responded to as not understanding how it would be possible for Google to have done this a certain way, and I was theorizing one possible way they might have done it.

Ah, then yes, apologies -I did not mean to put words in your mouth. Technical feasibility is likely easier than imagined; most Mastodon services use the auto-generated list that appears on their "about" page - easily scraped if not available through the API - here's the list on the instance I moderate on for example:

https://toot.wales/about/more#unavailable-content

I currently think this may be exactly what is happening. If I'm wrong, I'd love to know about that!

Reply on news.ycombinator.com