I wonder why everyone seems to hate the tags.pub bot so much.
-
I wonder why everyone seems to hate the tags.pub bot so much.
I mean: Until now, there was no way no get a notification if somebody posts something with a hashtag you're subscribed to.
This is a great feature, especially for hashtags with a low number of occurrences, greatly enhancing the value of the Fediverse for information purposes.
Just asking.
Edit: There's some good discussion below, worth reading. Also including some arguments for being sceptical about this thing.
@mina porque es el enésimo listo que hace algo y se lo mete a todo el mundo, y si no lo quieres, entonces tienes que poner unos tags en tu perfil o bloquear instancias. O como decía una colega, esto puede ser un problema con gente que tu tengas bloqueada. Perfectamente podría haberlo hecho opt-in, pero claro, los listos no entienden el consentimiento.
-
I wonder why everyone seems to hate the tags.pub bot so much.
I mean: Until now, there was no way no get a notification if somebody posts something with a hashtag you're subscribed to.
This is a great feature, especially for hashtags with a low number of occurrences, greatly enhancing the value of the Fediverse for information purposes.
Just asking.
Edit: There's some good discussion below, worth reading. Also including some arguments for being sceptical about this thing.
-
I wonder why everyone seems to hate the tags.pub bot so much.
I mean: Until now, there was no way no get a notification if somebody posts something with a hashtag you're subscribed to.
This is a great feature, especially for hashtags with a low number of occurrences, greatly enhancing the value of the Fediverse for information purposes.
Just asking.
Edit: There's some good discussion below, worth reading. Also including some arguments for being sceptical about this thing.
@mina Here is a thead where it is explained why this is a bad idea
-
Yet, I don't really get it.
You get a notification if your (anyway public) post gets boosted.
If you don't like that, you block the bot.
That should be it, or not?
-
@mina Here is a thead where it is explained why this is a bad idea
I'll think it through.
Need to sleep now.
Answers are coming in. I will read them all and possibly update my thoughts.
That was the whole point of me asking. To gather information.
-
Yet, I don't really get it.
You get a notification if your (anyway public) post gets boosted.
If you don't like that, you block the bot.
That should be it, or not?
@mina no. You should not have to find all the things that one day decided to include you in their processing (boosting, scraping, whatnot) of your information, and opt-out one by one.
You should opt-in into those that you want to participate in.
That's the principle of consent that people are offended about. For the moment I'll skip on the graphic parallels. -
I wonder why everyone seems to hate the tags.pub bot so much.
I mean: Until now, there was no way no get a notification if somebody posts something with a hashtag you're subscribed to.
This is a great feature, especially for hashtags with a low number of occurrences, greatly enhancing the value of the Fediverse for information purposes.
Just asking.
Edit: There's some good discussion below, worth reading. Also including some arguments for being sceptical about this thing.
@mina Para algunos esa situación es una característica útil, no algo para arreglar.
-
I wonder why everyone seems to hate the tags.pub bot so much.
I mean: Until now, there was no way no get a notification if somebody posts something with a hashtag you're subscribed to.
This is a great feature, especially for hashtags with a low number of occurrences, greatly enhancing the value of the Fediverse for information purposes.
Just asking.
Edit: There's some good discussion below, worth reading. Also including some arguments for being sceptical about this thing.
@mina Subscribing to a hashtag on a Mastodon server only shows you tagged content that arrived at your server for other reasons -- you follow the author, someone on your server follows the author, relays, boosts, replies, a couple of other reasons.
-
@mina Subscribing to a hashtag on a Mastodon server only shows you tagged content that arrived at your server for other reasons -- you follow the author, someone on your server follows the author, relays, boosts, replies, a couple of other reasons.
That means on small servers, you don't get much content on the hashtag feeds. Even on medium-sized servers like berlin.social, you don't get everything. You can compare this feed:
https://berlin.social/tags/fediverse
To the same tag on mastodon.social:
-
That means on small servers, you don't get much content on the hashtag feeds. Even on medium-sized servers like berlin.social, you don't get everything. You can compare this feed:
https://berlin.social/tags/fediverse
To the same tag on mastodon.social:
@mina tags.pub tries to bring content from different servers together and redistribute them by hashtag, so that people can find content they're interested in. This makes smaller servers and non-Mastodon platforms, some of which don't have hashtag feeds, more viable alternatives to mastodon.social.
There's a little more info here: https://tags.pub/#why
-
I wonder why everyone seems to hate the tags.pub bot so much.
I mean: Until now, there was no way no get a notification if somebody posts something with a hashtag you're subscribed to.
This is a great feature, especially for hashtags with a low number of occurrences, greatly enhancing the value of the Fediverse for information purposes.
Just asking.
Edit: There's some good discussion below, worth reading. Also including some arguments for being sceptical about this thing.
-
@viq@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. Simply because toots are already public and fan out by the relays.
-
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
discoverableandindexableflags 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 issue I can see for privacy isn’t much about 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…
-
-
@viq@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. Simply because toots are already public and fan out by the relays.
-
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
discoverableandindexableflags 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 issue I can see for privacy isn’t much about 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…
-
-
@mina tags.pub tries to bring content from different servers together and redistribute them by hashtag, so that people can find content they're interested in. This makes smaller servers and non-Mastodon platforms, some of which don't have hashtag feeds, more viable alternatives to mastodon.social.
There's a little more info here: https://tags.pub/#why
This makes total sense to me.
For me, the main reason to follow such a bot is not to get more content into my feed (at best, I read perhaps 2% of it), but to get posts with certain hashtags into my mentions in order to actually see them, as there is no standard way of getting notified when a new post under a certain hashtag is published.
This is gold!
-
@mina Para algunos esa situación es una característica útil, no algo para arreglar.
A mí también me parece.
-
@mina no. You should not have to find all the things that one day decided to include you in their processing (boosting, scraping, whatnot) of your information, and opt-out one by one.
You should opt-in into those that you want to participate in.
That's the principle of consent that people are offended about. For the moment I'll skip on the graphic parallels.In reality, this is not an opt-in/opt-out thing.
Ask yourself:
Why would you make a public post with a hashtag?
Is there another reason then to enable people to find it?
Hashtags are propagated through the Fediverse anyway.
With the bot you make it easier for people on smaller instances to actually get the posts published under the hashtags they're interested in.
-
@mina Here is a thead where it is explained why this is a bad idea
Seriously:
I still don't understand it.
You publish a public post under a hashtag.
It gets propagated through your network on the Fediverse. Every person can subscribe to that hashtag and find your post in their feed.
The bot only shares your post with people who would subscribe to the hashtag anyway but who want to miss out less.
-
I do believe, this is a misunderstanding how the Fediverse and bots actually work.
Yes, you can already subscribe to hashtags, but on small instances you only see a fraction of the posts published under the hashtag you're following.
With the bot, you have the chance to see posts published on distant instances.
It's not a question of consent: You already posted something on the pulic internet and made it discoverable on purpose by adding a hashtag. Every search engine can index it.
-
Seriously:
I still don't understand it.
You publish a public post under a hashtag.
It gets propagated through your network on the Fediverse. Every person can subscribe to that hashtag and find your post in their feed.
The bot only shares your post with people who would subscribe to the hashtag anyway but who want to miss out less.
@mina The idea of Fedi is a human-interaction system, we do not expect our posts to be boosted by bots, especially not without prior consent.
This is a bot-only system where bot-accounts are created automatically; there is zero interaction between the bot (and the bot owners) and the post's author.
This system assumes that a major goal of posting is to have a large reach, but this ain't Twitter, this ain't Threads. Most people use hashtags as a search-help, not as an audience-reach tool.
I have e.g. now switched to Followers Only-posts anymore, as I had enough of techbros like him arguing like this. I don't really want reach, I just know that there are people who watch some of the hashtags I post, that was the reason for using those hashtags at all.
Now, no hashtags anymore because I can't know if another techbro comes around the corner doing whatever with my posts.
There are tons of reasons why we love Fedi, and of which is that there is no algorithm and no automated system to gather reach.
-
Seriously:
I still don't understand it.
You publish a public post under a hashtag.
It gets propagated through your network on the Fediverse. Every person can subscribe to that hashtag and find your post in their feed.
The bot only shares your post with people who would subscribe to the hashtag anyway but who want to miss out less.
@mina @ics The „service“ added an unasked for turbo boost to your hashtags and created accounts on their instance which could look like yours, so Alice posted their nudes under #alicepics and suddenly there is this account „alicepics at tags.pub“ that only boosts the nudes.
This looks shitty and breaks boundaries. Alice has made a few posts about that.
The biggest problem is that this happened without consent. If there was an explicit opt-in step, nobody would have complained. But, as Evan said, then they wouldn’t have as many users.
