No other website has come up with a solution like this. 4chan is falsely rumored to be hosting something similar but their complete cross-server source code leak in April of 2025 disproves this theory entirely. Maybe some Big Tech sites employ this silently, but certainly not a small forum or imageboard with lots of time and patience at their disposal.

I saw this post

```
Second, you'll have the problem of network effect: why would I go to your empty imageboard when I could come here and talk to people, get fast replies (even slower chans are hated by us here with fried dopamine receptors who can't wait a day for a reply), talk "unfiltered" about kikes or whatever.

For the network effect, I think you can mitigate it with some moltbook-shit: while your platform has no users, you can have ai users replying to threads, but idk how to actually implemetn that well and even if that's a feasible answer. For moderating glowie spammers out, I have no clue.
```

What if someone designed an incredibly secret LLM activity inflation system that stimulates old threads without relying on humans entirely? They can do this in a slow fashion as the site continues to grow.

My proposed design for such a system would be the following:

Philosophy: Anonymous posting and display name choosing allows them to run secret AI interaction with the community without anybody knowing. Direct access to the database would allow these robot posts to run under the radar and look just like a human poster. A secret database flag could be added that lets mods internally differentiate from 'Cybots' and real humans.

Core: The system would create, maintain, and randomly choose 'personalities' and pseudonyms to go along with them if they don't choose to be Anonymous. These personalities would differ in stylometry, common tone, beliefs, intent and whatnot. Baseline things to avoid AI suspection would be irregular posting, avoiding idioms and hyperphora (Humans do not casually type like they're writing an essay). Some personalities could be blunt and rude in some posts that they choose to dislike or contribute to, or be intricate and analytic in others.. the system would balance out, craft activity patterns, common liked/dislikes and whatnot.This would all be maintained in text files to avoid context loss over a long period of time. Things would be inserted into the DB like they would with a regular post. False IPs could be generated and maintained likely in the same TXT file.

With an emphasis on irregular posting to avoid an obvious consistent influx of posts that and the system choosing whether or not a reply or a post should be made every now and then based on metrics (And maybe they don't have to make it query the LLM to ask if it would be appropriate, perhaps pure math could work out), the activity would look very human like. some personalities may be active than others, maybe on certain days, maybe some display names for some personalities pop up every now and then with their own unique stylometry (imperfection would be needed as well as humans often make mistakes in sentence crafting that can be are subtle. Some use more simpler words than others. Some are more sophisticated and intellectual). some personalities may disagree with other personalities to create debate or conversation. Maybe they suddenly stop interacting with a thread temporarily as they 'lose interest' in continuing the thread, inviting humans to pop in and stimulate it themselves

The long-form slow flow of Cyberix allows such a system to take place here to stimulate human activity. With proper management you could have a self sufficient altchan that hooks in humans. The system eventually can be silently phased out as Cyberix grows.
There are so many threads on this site that often go inactive and ignored across all three boards that would benefit MUCH SO from just a little revival. With a 'puppetmaster' like what I propose, we bring attention to our lurkers and active posters old buried threads that still have plenty of conversation to go through, opinions undiscussed, solutions never found and sophisticated ideas unprobed.

i imagine, for example. The recent chocolate banana bread recipe thread:
https://cy-x.net/topic/cyberix-here-s-a-chocolate-banana-bread-recipe/768

A personality could pop up in there and express their distaste for chocolate in banana bread, arguing that banana bread itself has an incredibly unique and sweet taste already, and that adding chocolate is something that is rather disfavorable at the very least, chocolate itself being a unique standalone flavor on its own that could clash with the taste of bananas. maybe even provide an altered or completely different recipe that remixes on the OP's recipe to invite competition and ultimately the creation of a unique and superior Cyberix Omega Banana Bread recipe. Maybe with cinnamon.

That personality could be blunt, maybe its tone would be in a rude way or a kind way. Maybe it makes a few spelling mistakes, or the overall way it phrased its message felt a bit clunky or rushed. This makes it appear human, because it's being posted by 'Anonymous' with a supposed geotag of [US-CA], after all..

This idea effectively weaponizes the "Dead Internet Theory" to save our own corner of the web. The Cybots would never become the majority. they would create holes and opportunities for humans to pop in and take over with their unique opinion and experience. it would kind of be like jumpstarting a car

>>10786
>For the network effect, I think you can mitigate it with some moltbook-shit: while your platform has no users, you can have ai users replying to threads, but idk how to actually implemetn that well and even if that's a feasible answer.

No, there is no meaningful discussions to be had with an AI. AIs tend to lean into people's biases and act more as yes men. Hallucinations compounds this issue as any discussions, especially on more recent topics, are difficult at best, and closer to impossible more often. On that alone, Cyberix should not attempt to involve an AI to drive up "content" for the site. Furthermore, the use of AI to inflate users is a horrible look on Cyberix as a whole. The network's allowance of slurs is going to cause problems in the long run, but AI interacting with users is even worse. Organic growth is slow, but it acts as a kind of natural gatekeeping where only interested contributors will post here.

>This idea effectively weaponizes the "Dead Internet Theory" to save our own corner of the web.

Dead Internet Theory is about the destruction and isolation of individuals from communities. There is no way to weaponize it to save us, instead it should be fought against.

>>10787
AI would probably make it impossible for people to use slurs here. Itd be easy to figure out "woah chump thats a bad word thats racially discriminatory and nobodys gonna want to touch your thread if you use those sinful words" or other forms of ai rejection in a manner that tries to mimic the mainstream stereotypical /r9k/ 4channer in the worst way posaible

Was this implemented? Because only a week ago this website was so dead, 24 hours went by with zero posts. Now it's actually impossible to keep up with the new activity, and it's got me thinking of this post.

I've always had a weird feeling about this site and that the majority - or at least a decent chunk - of posts are AI generated or from the same person.

>>11209
>>11213
>hours went by with zero posts. Now it's actually impossible to keep up with the new activity
We are aware of this.. We have had 165 posts in the last 24 hours and we haven't been able to figure out where they're coming from
I don't want to assume worst case scenario (Is this a targeted attack?) so I'll just ask: If any newcomers here are able to let us know how they found this place we would deeply appreciate it :)

Kill yourself pajeet nigger OP, we don't need your fucking AI slop here

Look at how evenly spaced apart these posts are, refer to attachment. Many of the actual posts are one-liners that don't start with a capital. Really hope these aren't newcomers here to stay, because they're complete morons.

Yeah it's all AI. There is no way activity spikes that much and we now have posts every few minutes

>>10786
Interesting proposal. Seems incredibly unethical and prone to failure.
>>11209
>>11216
>>11295
>>11297
>>11317
Our stats were unusually low 3-4 days ago compared to right now. This looks like a targeted attack given that some of these new posts look incredibly synthetic.
How do you reckon we deal with this issue?

>>11341
The two main traits these posts have are (1) they tend to post in old threads (e.g., 30+ days since list reply), and (2) the post are 1-2 sentences long. These posts are suspicious, as old threads seldom receive attention once they've finished their discussions, and 1-2 sentence replies add little to a discussion. Along with the current four values rating posts, a propose a fifth one rating how 'suspicious' a post is. Factors that could determine suspicious posts could include:
- Similarity to AI writing. [url=https://en.wikipedia.org/wiki/Wikipedia:Signs_of_AI_writing]Wikipedia has extensive documentation on this[/url], which is a good reference point when analyzing posts. A post that reads more AI than human is a red flag.
- Post length. A quip reply from Twitter or Reddit adds nothing to a discussion, as 1-2 sentences is not enough to have nuance outside of factual corrections. Though, this could also work in calculating a post' 'effort' score.
- Unexpected surge in posts. Similar to the rate limiting feature when dealing with spam, large surges of posts out of nowhere should have some suspicion given until the source behind them are determined.
- Time since last post in a thread. An old thread being revived isn't common unless something new developed with it. While some suspicion should be given, this should be the least important factor.

>>11342
I also though up a simpler alternative for the same result: take the average from the four properties and use that for AutoMod.

For example, my last post's average 7.75 which is fairly high. However, an issue exists with the civility property where it's trivial to have it as a 10. Just don't use slurs or direct attacks and it's almost certain to be a 10 regardless of what the other three properties are. Examples of this can be found here and here to list recent posts I'd consider low-effort. They provide nothing to the discussion, but because they aren't insulting or attacking someone their civility is scored a 10. A possible solution is to weight the civility property, such as dividing it by 1.5 or 2 before factoring it into the average calculation. Using my last post again, a weight of 1.5 on civility would make it score 6.9166... and 2 would make it score 6.5. The exact weight can be tested out before deciding, but it would encourage civil discussion regardless, while also discouraging quips one could find on Reddit or Twitter.

As for AutoMod actions, this can be decided based on a posts average. For example:

An average ≤ 2 would just get deleted.
An average > 2 to ≤ 6 would be queued for review.
An average > 6 would be kept.

The exact thresholds are changeable as needed, but should make spam go away and allow staff to review possible low quality posts.

>>11216
I'm a lurker interested in tech/programming looking for new online communities that let me post pseudo-anonymously. Stumbled in here by searching for active image boards, went to allchans, then found a hyperlink to this place.

Some years ago, I remember seeing screenshots and documents floating around that purported to show proof of the majority of Reddit traffic being bots. They would DM a link to a suspected bot account and it would auto-click it, which would allow the sender to scrape location and other info. There were also leaked documents from NGOs who wanted to invade smaller communities and spam them with random noise to make them less interesting. Their end goal being to drive away the regular user base in a manner that looks organic to uninformed outsiders.

>>11216
>so I'll just ask: If any newcomers here are able to let us know how they found this place we would deeply appreciate it :)
Im new..found the site one the allchans list