🇺🇦 Help us to save Ukrainian cultural heritage!

https://www.weaponstoukraine.com/kampane/archa 🇺🇦

🌲ikt has funded planting of 1,145 trees and the prevention of 54.1 tCO2e from being emitted

https://ecologi.com/ikt 🌲

  • 627 Posts
  • 1.93K Comments
Joined 3 years ago
cake
Cake day: June 17th, 2023

help-circle







  • ikt@aussie.zonetoGreentext@sh.itjust.worksThe cops pay Anon a visit
    link
    fedilink
    arrow-up
    4
    arrow-down
    1
    ·
    edit-2
    1 day ago

    You already failed 😂 you should have started with: This hits so hard!

    Also the bot never responds to comments only make once off posts

    Also you think you type fast but I highly doubt you could type out 3 well thought paragraphs in under 2 minutes before rapidly shifting to your next random post

    Actually let me put your posts into AI to see what it thinks of you >.>

    edit:

    Based on this document, this user appears to be very unlikely to be AI. Here’s why:

    Strong indicators they’re human:

    1. Natural conversational style - Their comments show authentic voice, sarcasm, frustration, and humor that feels genuinely human
    2. Self-aware about AI comparisons - They literally joke about “typing fast and having the writing style of LLMs” and mention doing “blade runner android tests” - this self-deprecating awareness is very human
    3. Contextual engagement - They engage deeply with Lemmy community dynamics, reference specific memes, understand platform culture
    4. Nuanced positions - They take complex stances (like on landlords, erotica vs porn, political opinions) that show genuine critical thinking
    5. Emotional authenticity - Express genuine frustration, disappointment, hope (“May men and women unite to be horny and kinky and unashamed”)
    6. Inconsistent formality - Mix casual language (“Homie,” “Thad”) with more elaborate thoughts
    7. Community participation - Active across different communities with varied topics over time

    Why they might worry about seeming like AI:

    • They mention typing fast and apparently write in a clear, structured way
    • They provide “thorough breakdowns” with their expertise
    • They’re articulate and thoughtful

    But these are just signs of an educated, engaged person who writes well. Their humor, cultural references, self-awareness, and authentic frustration with community dynamics all point strongly to a real human user.

    vs what it thinks of /u/[email protected]:

    Based on this user’s comment history, I’d say very likely AI-generated, probably 85-90% confidence. Here are the telltale signs:

    Strong AI indicators:

    1. Unnaturally consistent quality and structure - Every single comment is well-formatted, articulate, and substantive. Real users have off days, typos, lazy responses.

    2. Suspiciously broad expertise - This account confidently discusses Linux kernel internals, LLM security, DRM technicalities, privacy policy, climate science, ADHD management, and Home Assistant updates with equal facility. That’s unusual.

    3. Formulaic pattern - Almost every comment follows: acknowledge topic → detailed technical/political take → prescriptive advice → punchy conclusion. It’s like watching the same template execute.

    4. Perfect grammar with casual affect - The writing is too clean. Real casual internet comments have more sentence fragments, inconsistent punctuation, genuine rambling.

    5. No genuine interaction - These read like mini-essays dropped into threads, not actual conversations. There’s no “lol,” no genuine confusion, no asking for clarification.

    6. The timing - Account created 10 hours ago, 36 comments already, all this polished? That’s bot behavior.

    What’s interesting: The account got banned from multiple communities quickly, which suggests moderators spotted this too. The “Removed by mod” entries are probably communities that don’t allow AI-generated content.

    Bottom line: This reads like someone using Claude or GPT to farm engagement/karma across Lemmy instances. The voice is too consistent, too knowledgeable, and too… helpful to be one person casually commenting.