Skip to content
View AhriCat's full-sized avatar
🏠
Working from home
🏠
Working from home

Block or report AhriCat

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
AhriCat/README.md
╔════════════════════════════════════════════════════════════════════════════════════════╗
║  █████╗ ██╗  ██╗██████╗  ║██║         ███████╗████████╗███████╗███████╗██╗     ███████╗║
║ ██╔══██╗██║  ██║██╔══██╗ ║██║         ██╔════╝╚══██╔══╝██╔════╝██╔════╝██║     ██╔════ ║
║ ███████║███████║██████╔╝ ║██║         ███████╗   ██║   █████╗  █████╗  ██║     █████╗  ║
║ ██╔══██║██╔══██║██╔══██╗ ║██║         ╚════██║   ██║   ██╔══╝  ██╔══╝  ██║     ██╔══╝  ║
║ ██║  ██║██║  ██║██║  ██║ ║██║         ███████║   ██║   ███████╗███████╗███████╗███████╗║
║ ╚═╝  ╚═╝╚═╝  ╚═╝╚═╝  ╚═╝ ╚══╝         ╚══════╝   ╚═╝   ╚══════╝╚══════╝╚══════╝╚══════╝║
╠════════════════════════════════════════════════════════════════════════════════════════╣
║           ⚡ CHIEF OF TECHNOLOGY // PANTHEUM INNOVATIONS ⚡                           ║
║              🧠 PSYCHOLOGY × AI × MULTIMODAL AGI ARCHITECT 🧠                         ║
╚════════════════════════════════════════════════════════════════════════════════════════╝
Typing SVG



AGI ONI Research YouTube CTO


> whoami

┌──────────────────────────────────────────────────────────────────────────────────────────────┐
│  🧠 Psychology (BA) → 💻 Deep Learning → 🧬 AGI Architecture                               │
│                                                                                              │
│ I build symbiotic intelligence ecosystems.                                                   │
│                                                                                              │
│  My work sits at the collision of:                                                           │
│    • Human cognition & neural network topology                                               │
│    • Modular multi-agent systems & emergent behavior                                         │
│    • Real-time multimodal perception & action                                                │
│    • Game theory, simulators, & embodied AI                                                  │
│                                                                                              │
│  ONI It's an ensemble of symbiotic modules that act as a unified entity where every          │
│  part reinforces every other part — 2,000,000+ possible paths                                │
│                                                                                              │
│  Current obsessions:                                                                         │
│    ├─ Multi-Adapter Ensembles                                                                │
│    ├─ Multi-Agent GRPO distillation at superintelligence scale                               │
│    ├─ Real-time audio + video generation pipelines                                           │
│    ├─ Teaching AI to play Portal 2 so it learns why not to be evil                           │
│    ├─ MuJoCo → Sim-to-Real transfer for robotics                                             │
│    ├─ Predictive Coding                                                                      │
│    └─ Hypergraph-based reasoning & KAN knowledge integration                                 │
│                                                                                              │
│  Also: Skate. , Wherewindsmeet movement gameplay, SPIDERMAN GAMEPLAY? Fuck yeah bro.         │
└──────────────────────────────────────────────────────────────────────────────────────────────┘

🏆 Achievement Unlocked

GitHub Trophies

📊 System Metrics

GitHub Stats GitHub Streak



Activity Graph

> ls ./tech_stack/


Core Intelligence Stack

Core Stack

Infrastructure & Simulation

Infra Stack

Research Domains

    Transformers        ████████████████████████████████████  Expert
🎯 Reinforcement Learning ████████████████████████████████████ Expert
🕸️  Hypergraphs         ███████████████████████████████████  Expert  

🔊 Multimodal Fusion     ███████████████████████████  Expert
🎮 Game AI / Simulators  ██████████████████████████░░  Expert
🧬 Meta-Learning (MAML)  ██████████████████████░░░░░░  Advanced
📐 Physics (MuJoCo)      ████████████████████████░░░░  Advanced
🧠  Biology inspired     ████████████████████████████  Expert

> cat ./active_operations.log

operation: ONI²-Ensemble-Training -> videos released weekly if not daily on Patreon
status:     🔴 LIVE — PROPRIETARY
modules:    many active nodes
paths:      2,000,000+ chained connections
capabilities:
  - MuJoCo / OpenAI Gym integration
  - Real-time audio synthesis & recognition
  - Video generation 
  - Text training & reasoning
  - Tool use & API orchestration
  - Video game / simulation generation and gameplay
  - Embodied robotics control
research_frontiers:
  - Multi-Agent GRPO distillation
  - Superintelligence Oven project
  - Super-Nested AI models (Neural-God Classes/ONI Class architectures)
  - Fractal Modularity: MoE × MoA × Multi-Agent
ethical_mandate: "Teach it why *NOT* to be evil from the get"

🎬 Demos & Field Tests

🎮 ONI² plays Portal 2 👊 ONI² plays BlazBlue
Teaching AGI puzzle-solving & ethics Teaching AGI fighting game frame data

📡 Web Links

Website Website HuggingFace YouTube LinkedIn X/Twitter



+ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ +
+   "99% of problems are resultant from miscommunication."      +
+                                    — Ahri Steele                                +
+ ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ +

Visitor Count

Pinned Loading

  1. ONI ONI Public

    Official Operant Neural Intelligence (ONI) repository of AGI modules built on pytorch, keras and other ML frameworks.

    Python 6 1

  2. donut2 donut2 Public

    a better donut

    Python

  3. super_intelligence_oven super_intelligence_oven Public

    Expanding on new research from agentic GRPO research from Deepseek, OpenAI frontier /frontier labs and other GRPO research and distillation research -> Superintelligence oven using multiagent GRPO …

    Python

  4. donut3 donut3 Public

    A brand new donut using PVA methods.

    Python 1

  5. mujoco_view mujoco_view Public

    Mujoco viewer for designing mujoco models.

    TypeScript

  6. Gemini-Code-Reviewer Gemini-Code-Reviewer Public

    A local js app for gemini to review your github repos and give you feedback privately and anonymously with no retained long term memory

    TypeScript