HappyDemon
May 1, 20261 min read
I just uploaded my first project to GitHub: HappyDemon.
It's a desktop chatbot that learns from you. The more you interact, the more "demonic" it becomes, and yes, that's intentional.
What it does:
- Continuous learning (persistent in JSON)
- Multiple answers to the same question (random choice)
- Manual fuzzy matching (understands sentence variations)
- Security blacklist (violence, racism, pedophilia)
- Dark mode interface with color coding
- Custom pixel art icon
Why did I do this?
Because SimSimi taught me that chaos and unpredictability are fun, and current AIs are so full of filters that they've lost their charm and don't learn anything from the user.
I'm a Design and Cybersecurity student. This project is my living laboratory for understanding:
- How phishing exploits persuasive design
- How collective learning can be used for defense
- Why blacklists alone don't work
Repository link: https://github.com/CyberCrip22/IA.dev
Feedback, issues, and contributions are welcome. Want to teach me something new? HappyDemon is all ears.
Do you prefer AI with rigid filters or with personality? Let me know.
Source: Dev.to
