Why Kaonomi ?

Building Jarvis: Why I’m Still Chasing the 1956 Dream (and Why You Should Join Me)

"Fast forward 45+ years after I first fell in love with AI at age 11... Today, we have massive neural networks that can write poetry, debate quantum physics, and hallucinate fake citations with absolute confidence."

Back in 1956, a group of brilliant minds gathered at Dartmouth College with a delightfully optimistic idea: they figured that if they spent two months working together, they could pretty much solve artificial intelligence.

Yet, true Artificial General Intelligence (AGI)—the kind of seamless, context-aware, reasoning co-pilot Dartmouth envisioned—remains elusive.

Are we close to true AGI? Honestly? No, not even close.
Do I care? Not one bit. Because trying to build it anyway is the most fascinating engineering adventure of our time.

From Computer Vision to the Deep Waters of Enterprise IT

My story with AI didn't follow a straight line. I started hands-on in the early '90s specializing in computer vision. But then the AI winter hit: overblown promises gave way to reality, and the field went quiet. So I pivoted into enterprise architecture, exploring large-scale information systems—SAP, SharePoint, ECM—for over two decades.

That detour wasn't wasted time. It taught me how businesses actually operate, how data governance works, and how to build critical architectures at scale. When I returned to AI full-time in 2014, I brought that dual perspective: the theory of complex systems combined with real-world operational execution (finance, defense, energy, and currently Swift).

Being on the autism spectrum (Asperger's) gives me a rather raw, direct relationship with technology. I don't really do hype; I care about systems, architecture, and practical execution. On my YouTube channel (Le Lab IA), I break down local LLMs, agent frameworks, and enterprise architectures. But deep down, I’ve always held onto one ultimate geek fantasy: building a real, fully operational Jarvis.

> Not just another chatbot window where you type: "Please format this JSON."

I dream of a personal, highly capable, context-aware AI assistant—one that manages real-world tasks, interfaces with complex APIs, runs efficiently (often locally), and acts as a genuine extension of human intelligence.

The Kaonomi & M-A-R-I Journey

To move beyond theoretical discussions, I’ve been actively building open, practical foundations for this ecosystem:

• ➔ Kaonomi: A project designed to structure, connect, and orchestrate modern AI modules, enterprise tools, and local models into a coherent architecture.
• ➔ M-A-R-I: A collaborative exploration into modular multi-agent intelligence and human-AI interaction.

These projects aren't built to sell another proprietary SaaS subscription. They exist to answer a fundamental question: How do we assemble all these fragmented puzzle pieces into an actual personal assistant platform?

Why We Need a Community (And Why You Should Join)

Building an Iron Man-grade Jarvis isn't a weekend side-project for one person—it’s an ecosystem challenge. It requires:

1. Local & Hybrid Control: Keeping sensitive data local while orchestrating cloud models when heavy compute lifting is needed.
2. Context & Memory: Moving away from static context dumps toward semantic caching, dynamic RAG, and stateful memory.
3. Actionable Tool Use: Giving agents safe, well-governed execution boundaries to interact with real APIs and real workflows.