DEV Community

PSBigBig
PSBigBig

Posted on

One Developer vs Ten AI Masters: My Open-Source Reasoning Engine Tournament

Image description

Hello DEV community,

I'm a solo developer and I've just completed a public experiment that I call the "AI Martial Arts Tournament."
I pitted myself and my open-source reasoning engine, WFGY, against the world’s ten top LLMs: ChatGPT, Claude, Gemini, Grok, Perplexity, Kimi, Merlin, HuggingChat, DeepSeek, Copilot.

Every prompt, answer, score, and line of code is fully public and reproducible.
Anyone can verify or break the results—no black boxes, no secrets, just a global open challenge.

🧙‍♂️ My “martial arts manual” boils down to four core moves, each inspired by Taiji (Yin-Yang) principles and formulated mathematically:

  • BBMC (Semantic Residue Formula):
    B = I − G + mc²

  • BBPF (Progressive Formula):
    BigBig(x) = x + ∑Vi(εi, C) + ∑Wj(Δt, ΔO)Pj

  • BBCR (Collapse-Reset Formula):
    Collapse → Reset(St, δB) → Rebirth(St+1, δB)

  • BBAM (Attention Modulation):
    ãᵢ = aᵢ · exp(−γσ(a))

Why does this matter?

  • Peer-reviewed results (95/100 by SciSpace)
  • Fully open-source methodology (free and ad-free)
  • Designed for any developer, researcher, or “meme scientist” to reproduce and improve

You can:

  • Test your favorite LLMs using the same reasoning toolkit
  • Suggest new edge cases or reasoning challenges
  • Audit, break, or improve the approach (I welcome all feedback and criticism!)

Links:
Main repo & docs: github.com/onestardao/WFGY

Discussion:
Would you trust an open-source “reasoning engine” to boost AI logic?
Which AI should face the next challenge, or what kind of “martial arts move” should I design for future tests?

All technical reviews, challenges, and memes are welcome!
Let the martial arts tournament commence! 🥋🤖

Tags:
AIChallenge, WFGY, OpenSource, LLM, ReasoningEngine, MartialArtsAI, PeerReview, PublicTest, NextGenAI, CivilizationStarter, MemeScience, Innovation, OneManArmy

Top comments (0)