@viq@social.hackerspace.pl @mina@berlin.social do you apply the same reasoning also to the various “Trending/Hype” bots that boost any post that gets more than a certain amount of interactions?
Moreover, services like tagpush.app have been around for longer than tags.pub, and I have never seen such a backlash against them.
A good gentlemen’s agreement has always been for users to add
#nobot to their profiles if they don’t want automated accounts of any kind, and for bots to respect it.
This episode is a replay of many other incidents on the Fediverse where disagreements surfaced between those who want to improve discoverability and those who want discoverability to be minimized.
It’s always good to listen to everyone’s points of view, but some pragmatic considerations should also be made before charging windmills with our spears:
Right now, nothing prevents me from writing a bot in 30 lines of code of Python that does something similar to what tags.pub does, hook it to e.g. relay.toot.io or some other big relay, and achieve exactly the same results. Simple because toots are public.
If someone posts something publicly on the Internet, with hashtags attached, it probably means that they want to increase the visibility of their content under that hashtag, and that they accept to lose control over which hashtag aggregators end up processing their content, in exchange for greater visibility towards anyone who follows that hashtag. If this assumption is not correct in a specific case, then perhaps using a hashtag in that specific case is a bad idea.
Right now, dozens of scrapers per minute are running against my instance. And it’s just an almost single-user instance, imagine what something like mastodon.social must see. All kinds of bots - LLM scrapers, search engine bots, security scans…most of them don’t even bother to respect robots.txt, let alone the
#nobot hashtag on someone’s profile. You can assume that anything you post publicly online (especially if both the discoverable and indexable flags are enabled on your profile, which AFAIK is the default on Mastodon) will end up processed by some faceless bot regardless of your
#nobot, and that it will end up on search engines, fed to the next iteration of ChatGPT, analyzed by some State-sponsored organization or private company for sentiment analysis, etc. And none of those usages are something that you can control, regardless of your consent. Given this background, do you think that something like tags.pub is the greatest concern here - especially considering how much
@evan@cosocial.ca is sensitive on some topics and open to feedback?
I personally believe that the fight for limited visibility and sharing of content that is already public (and not only, but also already easily accessible through machine-friendly formats like JSON-LD) is a lost fight, and that those who feel uncomfortable with any non-consensual usage of their public content should just not make their content public (let alone index it under a hashtag). But of course I’m open to be proven wrong.
p.s. The only actual real issue I can see for privacy isn’t much for those who post their content publicly anyway, but because the list of followers for a @tags.pub account seems to be publicly visible.
@evan@cosocial.ca do you think it’s worth addressing this and hiding the followers list for these bots? I wouldn’t want that someone with bad intentions decides to scrape all the followers of a sensitive tag…
