Ask HN: How Can We Programmatically Assess if an LLM Feels ‘Too AI’?

🌟 Transforming AI Content Creation with Aaptics 🌟

As a founder on a mission, I’m excited to introduce Aaptics, an innovative tool designed to empower fellow founders in drafting authentic content.

Key Highlights:

• Overcoming Challenges: The biggest hurdle isn’t the tech; it’s making AI sound human—not like a corporate robot.
• Our Approach: We utilize a custom Retrieval-Augmented Generation (RAG) setup, combining:
  • Users’ past writings
  • Heavy negative prompting
  • Few-shot examples to refine tone

Call for Collaboration:

For those in the AI field:

• How do you evaluate the “humanness” of your outputs?
• Do you use LLM-as-a-judge frameworks or tweak parameters?

Your insights are invaluable as we gear up for our mid-April launch. Let’s connect and share strategies to enhance AI-generated content!

📬 Join the conversation and sign up on our waitlist: Aaptics Waitlist

Source link