X changes 'show more replies' to 'probable SPAM' – but also hides dissenting posts from paying X Premium users
The algorithm for X, formerly Twitter, is labeling legitimate replies as SPAM, even from users paying for a verified account
🗑️ X has changed “Show more replies” to “Show probable SPAM”
🤖 The X algorithm can trap dissenting negative opinions in this submenu
💸 This includes hidden posts from X Premium users paying $8/mo
💡 We uncovered a way to get more visibility for your legit dissenting opinion
☝️ Be sure to tap the “Show probable SPAM” on X to get a fuller picture
X, the social media platform formerly known as Twitter, has changed the way it labels replies that are hidden from posts. Its language is a little more specific now, but it also hides dissenting opinions, even from paying X Premium users, under the guise that it’s SPAM, according to the research of The Shortcut.
Where it used to say “Show more replies” the text now says “Show probable SPAM.” X certainly has a SPAM problem and owner Elon Musk’s pledge to “defeat the spam bots or die trying” didn’t completely fix the platform. In some areas, it’s better, and that may be due to tactics like charging a fee for verified accounts, charging a fee for new users during their first months on the platform, and routine bot purges. My own Matt Swider X account at 900,000 followers faces less SPAM than it did in 2022.
So it makes sense why X is taking steps to label SPAM. There’s just one problem…
X labels dissent as SPAM even by paying users
“Show probable SPAM” is a positive step for X, but it looks as if the algorithm is still too aggressive, trapping legitimate replies behind this hidden interface. Occasionally, these are dissenting opinions, with lengthy, fully formed replies that stand in contrast to the original post.
Paying for X Premium isn’t a guaranteed get-out-of-probable-SPAM-jail-free card
This includes accounts paying for the verified X badge under the X Premium plan (formerly known as Twitter Blue). The algorithm sees them as negative in tone. So paying $3/mo for X Basic, $8/mo for X Premium, or $16/mo for X Premium+ isn’t a guaranteed get-out-of-probable-SPAM-jail-free card.
This also happened when the language was “Show more replies,” but at least then the social media platform didn’t mislabel dissenting opinions as replies.
How to get dissenting replies seen on X
The X algorithm is aggressive and sometimes categorizes replies as SPAM, but it’s still an algorithm. There are ways to get around this, like beginning posts with positive language, and pointing out the parts (if any) of the original post you agree with. “I agree with you partially, but…” I’ve noticed when you use harsher language that’s not in agreement with the original post, your reply has less visibility.
There’s also another hidden section on X replies beyond "Show probable SPAM” that reads “Show additional replies, including those that may contain offensive content.” More times than not, the social media platform’s algorithm correctly rounds up SPAM and graphic tweets in this second layer. Only a few times have I seen standalone GIFs or 🫡 emojis without accompanying text in this portion of the interface.
This is just the latest lever X is pulling to combat its ongoing problem with SPAM. The fact that paying verified accounts are being trapped in the hidden layer of the UI is especially troubling. Clearly, the algorithm needs fine-tuning and this may require input from Grok, the first product from Elon Musk’s AI company xAI.
Next: See Gigabyte’s AI Arcade which combined Midjourney with Street Fighter
I am currently having as situation where my replies aren't even seen under "probable spam." They just aren't being shown at all.
The tricky, shady part is there is no notification of this and to the user (myself) it looks like my posts are...posting. There is no indication that I'm being hidden, no one ever told me anything. It just doesn't show my posts at all to other people which is BS in my opinion and no way to run a social media platform.
This isn't that accurate. The tweets after 'Show more replies'/'Probable spam' have always been labeled as "LowQuality". Those in the next section are labeled as "AbusiveQuality". The other tweets - the ones above 'Show more replies' are labeled a "HighQuality". Twitter has been smearing millions of their users each day but no one else outside Twitter besides me knew about it because they couldn't do something as simple as look at Twitter's AJAX.
In any case, their algorithm has always been incredibly flawed; more times than not it gets things wrong. E.g., just yesterday I looked a replies to Musk. On one tweet, someone had left about a dozen worthless meme tweets, all of which were labeled as "HighQuality". I keep seeing merch spammers. They even use the same lingo, but Musk is too grossly incompetent to pick up on that.
To understand what's really going on, see the description of my "More Speech" application on Github.