@Jorge Candeias Bad idea. (Hubzilla user here.)

Hashtags are not only for discoverability (and critically so on Mastodon). They're also the preferred way of triggering the automatic generation of individual reader-side content warnings.

Content warnings that are automatically generated for each user individually based on keyword lists have a long tradition in the Fediverse. Friendica has had them long before Mastodon even existed, much less before Mastodon hijacked the summary field for content warnings. Hubzilla has had them since its own inception which was before Mastodon, too. (streams) has them, Forte has them.

On all four, automated reader-side content warnings are an integral part of their culture. And users of all four (those who are not recent Mastodon converts at least, i.e. those who entered the Fediverse by joining Friendica in the early 2010s) insist in automated reader-side content warnings being vastly better than Mastodon's poster-side content warnings that are forced upon everyone all the same.

Oh, and by the way, Mastodon has this feature, too. It has only introduced it in October, 2022, and since the re-definition of Mastodon's culture in mid-2022 pre-dates it, it is not part of Mastodon's culture. But Mastodon has this feature.

However, in order for these content warnings to be generated, there needs to be a trigger. The safest way is by hashtags: If you post content that not everyone may want to see, add corresponding hashtags, enough to cover as many people as possible. If you don't want to see certain content right away, add the corresponding hashtags as keywords to NSFW (Friendica, Hubzilla, (streams), Forte) or a CW-generating filter (Mastodon).

In fact, hashtags can also be used to completely filter out content that you don't want to see at all. And they can be used to trigger such filters. This should work everywhere in the Fediverse.

I myself post stuff that some people don't want to see all the time. Hence, I need a whole lot of hashtags.

Let me explain the "hashtag wall" at the bottom of this comment to you.

• #Long, #LongPost
  This comment is over 500 characters long. Many Mastodon users don't want to see any content that exceeds 500 characters. They can filter either or both of these hashtags and at least get rid of my content with over 500 characters.
  Why two hashtags? Because I can't know beforehand which one of them people will filter. And because I can't know beforehand which of one of them people will search for or follow.
• #CWLong, #CWLongPost
  The same as above, but making clear that it's supposed to stand in for a content warning ("CW: long (over 8,300 characters)"). Also, filtering these instead of the above has less of a chance of false positives than the above.
  Why two hashtags? Because I can't know beforehand which one of them people will filter. And because I can't know beforehand which of one of them people will search for or follow.
• #FediMeta, #FediverseMeta
  This comment contains Fediverse meta content. Some people don't want to read anything about the Fediverse, not even as by-catch or boosted to them by someone whom they follow or even only on their federated timeline. They can filter either or both of these.
  Why two hashtags? Because I can't know beforehand which one of them people will filter. And because I can't know beforehand which of one of them people will search for or follow.
• #CWFediMeta, #CWFediverseMeta
  The same as above, but making clear that it's supposed to stand in for a content warning ("CW: Fediverse meta" or, in this case, "CW: Fediverse meta, Fediverse-beyond-Mastodon meta").
  Why two hashtags? Because I can't know beforehand which one of them people will filter. And because I can't know beforehand which of one of them people will search for or follow.
• #Fediverse
  This comment is about the Fediverse. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic. Also, the hashtag helps people looking for content about the Fediverse find my comment.
• #Mastodon
  This comment touches Mastodon as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic. Also, the hashtag helps people looking for content about Mastodon find my comment.
• #Friendica
  This comment touches Friendica as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic, especially if you don't know what the hell Friendica is, but you're curious. Also, the hashtag helps people looking for content about Friendica find my comment.
• #Hubzilla
  This comment touches Hubzilla as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic, especially if you don't know what the hell Hubzilla is, but you're curious. Also, the hashtag helps people looking for content about Hubzilla find my comment.
• #Streams, #(streams)
  This comment touches (streams) as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic, especially if you don't know what the hell the streams repository is, but you're curious. Also, the hashtag helps people looking for content about (streams) find my comment.
  Why two hashtags if they're the same on Mastodon? Because they are not the same on Friendica, Hubzilla (again, that's where I am), (streams) itself and Forte. If I have to choose between catering to the technologies and cultures of Friendica, Hubzilla, (streams) and Forte and catering to Mastodon's, I will always choose the former.
• #Forte
  This comment touches Forte as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic, especially if you don't know what the hell Forte is, but you're curious. Also, the hashtag helps people looking for content about Forte find my comment.
• #MastodonCulture
  This comment touches Mastodon culture as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic, including critical views upon how Mastodon users try to force Mastodon's 2022 culture upon the users of Fediverse server applications that are very different from Mastodon, and that have had their own culture for much longer. Also, the hashtag helps people looking for content about Mastodon culture find my comment.
• #Hashtag, #Hashtags
  This comment touches hashtags as a topic. If you don't like it, you can filter it out. Otherwise, click it or tap it to find more content on the topic. Also, the hashtag helps people looking for content about hashtags and their implications find my comment.
  Why two hashtags? Because I can't know beforehand which one of them people will filter. And because I can't know beforehand which of one of them people will search for or follow.
• #HashtagMeta
  This comment contains hashtag meta content. Some people don't want to read anything about it, not even as by-catch or boosted to them by someone whom they follow or even only on their federated timeline. They can filter either it.
• #CWHashtagMeta
  The same as above, but making clear that it's supposed to stand in for a content warning ("CW: hashtag meta").

By the way: Hashtags for triggering filters are even more important on Hubzilla in comments when Mastodon users may see them. That's because Hubzilla cannot add Mastodon-style content warnings to comments (= everything that replies to something else; here on Hubzilla, it's very different from a post that isn't a reply). What's a content warning on Mastodon is still (and justifiedly so) a summary on Hubzilla. But from a traditional blogging point of view (Hubzilla can very much be used for full-fledged long-form blogging with all bells and whistles), a summary for a comment doesn't make sense. Thus, the comment editors have no summary field on Hubzilla. Thus, I can't add Mastodon-style CWs to comments here on Hubzilla.

#Long #LongPost #CWLong #CWLongPost #FediMeta #FediverseMeta #CWFediMeta #CWFediverseMeta #Fediverse #Mastodon #Friendica #Hubzilla #Streams #(streams) #Forte #MastodonCulture #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta

@Fabian :antifa: AI must never be mandatory, and AI must never be forced upon anyone or everyone.

For example, image descriptions. I write them myself. I write the longest, most extensive, most detailed image descriptions in the whole Fediverse because I'm largely free from Mastodon's restrictions. I spend hours researching for and describing and explaining one picture in tens of thousands of characters.

The full image description goes into the post itself, but I always have an additional actual alt-text with a short description which also mentions that there's a full description with lots of explanations in the post. I have to do this because I have to hide posts with image descriptions behind content warnings for being over 500 characters long.

Now imagine what'd happen if AI-generated image descriptions were forced upon everyone. The AI would first throw my own alt-text away. Result: Hardly anyone would even discover the actual image description because the alt-text would no longer explain where it is. Or that there is one in the first place.

Besides, the AI would write a completely mangled, non-sense description of an image which it fails to understand because the picture was not taken in the real world. The AI wouldn't even understand that the picture wasn't taken in the real world because it wouldn't expect that.

I wouldn't let my own hashtags be forcibly replaced by AI-generated hashtags. AI doesn't necessarily understand my posts, it doesn't have the competence in the super-niche fields I write about. How is an AI supposed to know what the Hypergrid International Expo is or what the Discovery Grid is? Worse yet: How is it supposed to tell from a picture which OpenSim grid name to hashtag?

Besides, I have my own way of using hashtags which is not rooted in Mastodon's culture. After all, I'm not on Mastodon. I'm on something that has existed since before there was Mastodon. It does things differently.

I always add hashtags with the purpose of triggering automated, reader-side content warning filters which are standard where I am, and which became available to Mastodon itself last year with the 4.0 release.

For example, if a post exceeds the 500-character mark like this one, I add the hashtags #Long, #LongPost, #CWLong and #CWLongPost. Or if I talk about something concerning the Fediverse which many people don't want to read about either, just like I'm doing here, I add the hashtags #FediMeta, #FediverseMeta, #CWFediMeta and #CWFediverseMeta.

Or if I ramble about image descriptions, and I'm pretty sure that many many Mastodon users don't want to read about that, I add the hashtags #AltText, #ImageDescription, #ImageDescriptions, #ImageDescriptionMeta and #CWImageDescriptionMeta.

You can easily make these trigger the automatic generation of a content warning just for you. Or you can have any and all posts with them deleted automatically. That's what they're there for.

But if some AI that's only built against Mastodon and Mastodon's culture comes and deletes them and replaces them with its own ones, it takes away this feature.

And don't get started about opt-out. You can add opt-outs for whatever you want on Mastodon. But I'm on Hubzilla. I can't opt into or out of anything that Mastodon and only Mastodon does.

#AI #Hashtag #Hashtags #HashtagMeta #CWHashtagMeta