Skip to content
  • Categories
  • Recent
  • Tags
  • Popular
  • World
  • Users
  • Groups
Skins
  • Light
  • Brite
  • Cerulean
  • Cosmo
  • Flatly
  • Journal
  • Litera
  • Lumen
  • Lux
  • Materia
  • Minty
  • Morph
  • Pulse
  • Sandstone
  • Simplex
  • Sketchy
  • Spacelab
  • United
  • Yeti
  • Zephyr
  • Dark
  • Cyborg
  • Darkly
  • Quartz
  • Slate
  • Solar
  • Superhero
  • Vapor

  • Default (Cyborg)
  • No Skin
Collapse
Brand Logo

CIRCLE WITH A DOT

  1. Home
  2. Uncategorized
  3. CW: AI Experiments / Personality Cloning 🤖I’ve been playing "Black Mirror" in my home lab!

CW: AI Experiments / Personality Cloning 🤖I’ve been playing "Black Mirror" in my home lab!

Scheduled Pinned Locked Moved Uncategorized
n8nselfhostedllmclaudeai
1 Posts 1 Posters 0 Views
  • Oldest to Newest
  • Newest to Oldest
  • Most Votes
Reply
  • Reply as topic
Log in to reply
This topic has been deleted. Only users with topic management privileges can see it.
  • dannyhayes@misskey.arcverum.ruD This user is from outside of this forum
    dannyhayes@misskey.arcverum.ruD This user is from outside of this forum
    dannyhayes@misskey.arcverum.ru
    wrote last edited by
    #1
    CW: AI Experiments / Personality Cloning 🤖

    I’ve been playing "Black Mirror" in my home lab! I built an AI agent that clones a friend’s personality using n8n and 5 years of Telegram chat history.

    🛠️ The Stack: n8n (Self-hosted) + OpenRouter + Telegram API.

    The Process:

    1. Data Cleaning: Used Claude Sonnet 3.5 to scrub 10k+ messages, keeping only the "target" voice.
    2. Profiling: Had gpt-4 analyze the clean data to extract speech patterns: sarcasm levels, emoji frequency, and "signature" phrases.
    3. The Logic: Built a workflow where the personality profile is injected into the system prompt.

    The Model Showdown:

    - Claude Haiku 4.5: 🏆 4/5. Perfect balance of speed and vibe. Closest to my friend’s natural flow.
    - OpenAI GPT 4.1 Mini: 3/5. Efficient for data tasks, but felt a bit "robotic" for deep conversation.
    - Claude Sonnet 3.5: 5/5. Surprisingly aggressive! It actually threatened to end the chat over a bad joke — spooky how much personality it injected.

    It's wild how well it works. It doesn't just answer questions; it answers them exactly like him.

    Have any of you experimented with personality replication? Which 2026 models are you finding best for "vibe" rather than just logic?

    #n8n #SelfHosted #AI #LLM #ClaudeAI #DigitalTwin #OpenRouter #Fediverse #Automation
    1 Reply Last reply
    1
    0
    • R relay@relay.an.exchange shared this topic
    Reply
    • Reply as topic
    Log in to reply
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes


    • Login

    • Login or register to search.
    • First post
      Last post
    0
    • Categories
    • Recent
    • Tags
    • Popular
    • World
    • Users
    • Groups