I think leftypol should introduce a stricter limit on the length of posts. We live in the age of AI, and AI-generated posts are unavoidable. However, we need to find a middle ground so that the forum doesn't get flooded with AI-generated walls of text

Posting mode: Reply [Return]
Name
Options
Subject	Spoiler Image
Comment
Verification
Flag
File
Embed
Password	(For file deletion.)

File: 1766933241545.jpeg (83.02 KB, 576x408, images.jpeg)

Anonymous 28-12-25 14:47:21 No.43764

I think leftypol should introduce a stricter limit on the length of posts. We live in the age of AI, and AI-generated posts are unavoidable. However, we need to find a middle ground so that the forum doesn't get flooded with AI-generated walls of text. That's why a strong character limit for posts is important. It shouldn't be too short, because that would prevent genuine effortposts. But it also shouldn't be as generous as it is now, since that enables wall-of-text bombing. A compromise is necessary in our current era.

Anonymous 29-12-25 15:43:28 No.43771

No.
Some of the best schizo posts are super long.

sage sage 30-12-25 00:53:36 No.43774

Your post is LLM-generated

Miraculous 31-01-26 18:36:38 No.44140

>>43771
This
Fucker just wants the ai ran reddit where it's just bots and Indians posting back to one another

You already can't say shit without takedowns on normie slop apps like fedbook or tikcock

Or look at X. That's always had mass character limits and not only is it the most compromised of places it's shit. The people are dumb as bricks.

Anonymous 31-01-26 20:23:46 No.44143

>character limit to filter AI
Twitter being ground zero for AI spam should be evidence enough that doesn't work.

Maybe a formatting minimum? Like require a bit of italicization or a list or something. Like LLMs can do those but they're bad at knowing when to do it.

If a thread doesn't do anything past the boilerplate substance-wise it can be treated as AI regardless.

Anonymous 01-02-26 01:09:00 No.44149

Reading a wall of text is the only way I can confidently tell whether something is AI slop or not :/

Miraculous 01-02-26 05:31:27 No.44151

>>44149
Basically
Schizo is praxis

Anonymous 01-02-26 12:49:48 No.44155

>>44149
no you can't. i have posted countless ai generated wall of texts here and if i take the deliberate effort to create a text which appears like it was made by a human, nobody of the anti-AI idiots cares. in most cases, it is totally enough just to delete all em dashes and replace them with , ; : people will swallow the bait, believing a schizo wrote this. sprinkle in some weird typosd by hand and voila: nobody knows if your text was made by a human or machine.

Anonymous 01-02-26 20:41:21 No.44159

>>44155
Non-interaction isn't not being clocked. The jannies just didn't bother to do anything about the reports.

Anonymous 01-02-26 20:47:44 No.44160

>>44155
Also that's the full extent of it's believability. AI can only make text that looks like a twitter user wrote it due to the practicalities of how LLMs generate text in the first place.
It doesn't matter if a believable block of text was made by AI or a human, because the kind of human you'd think wrote such a thing is already a sophistic stochrastic parrot trained by social media algorithms to say nothing to make line go up. AI merely automated the twitterlerite.

Anonymous 01-02-26 20:57:04 No.44161

>>44155
Are they those texts that go like:

You should do revolution, this issue is bourgeois
<6 paragraph of Marx
Bla bla bla
>4 paragraphs of Marx
Bla bla bla

Anonymous 01-02-26 21:04:26 No.44163

>>44161
Nah that's just ban evader that lacks object permanence and projects that onto everyone else, thinking that people don't click links.

Anonymous 01-02-26 21:12:42 No.44165

>>44163
Can you post an example of some AI-generated text?

Anonymous 01-02-26 21:19:46 No.44166

>>44165
If it's conceptually and executionally low effort and doesn't attempt escape the boilerplate, then it doesn't matter if it is AI or not, and may as well be treated as AI.
>>>/leftypol/2675276 may as well be and therefor can be assumed to be AI. There's not going to be any artifiacts to drive yourself schizo over. It's always going to be a "may as well be" type deal.

Unique IPs: 9