What We're Building

What Makes humanity.ai Unique?

humanity.ai is, at its core, a 3-part system comprised of an Orchestrator, Verifiers with domain-specific experts, and an Architect that identifies gaps in knowledge and learns new skills on demand, all on its own — no human interference required. The result? A self-governing AI ecosystem.

Key features of humanity.ai include:

• Self-evolving
• Hallucination-proof through iterative verification
• Hardware-agnostic
• Customizable domain expertise with automatic module fine-tuning
• Dynamic RAM usage for resource-constrained devices
  • Can be run on the cloud via distributed compute, on-prem, or on an individual user-owned device
• Shared global context management
• Always current: Leverages open-source models
  

For a detailed review of humanity.ai’s capabilities compared to other modular AI systems: humanity.ai vs. Other Modular AI

Benefits & Use Cases

There are three key benefits to the humanity.ai system:

1. Proprietary foundational tech allows for highly efficient compute, which is ideal for real-world constraints, such as cost-restrictions, on-prem device compatibility, and even limited global cloud-based compute capacity
2. Use of smaller dedicated domain experts increases veracity and reduces hallucination
3. The Architect's role in guiding autonomous self-evolution provides a framework for AI that can adapt to solve any and all challenges.

There is a wide range of potential use cases for this technology, including but not limited to:

• Cloud-based public AGI or locally hosted private AGI
• Enterprise-grade on-prem AI or cost-effective personal AI systems
• Compute-efficient distributed AI
• Autonomous robotics and unmanned aerial vehicles with seamless control interface integration

Our Technology

At the core of what makes humanity.ai so powerful is Chariot Technologies Lab's proprietary iCon (interpretable containers) architecture, which allows for the dynamic swapping in and out of different models on a single device.

In iCon, the Conductor processes input and develops a plan to complete the request, then delegates subtasks to relevant domain experts accordingly. If a relevant domain expert does not exist in the system, the Conduct sends a request to the Architect, who makes and oversees the execution of a plan to find, download, and add a new domain expert module (either pre-trained or trained on the fly via available datasets). This could hypothetically be expanded to collecting new data and building a new dataset from scratch.

Each domain expert's output is checked by a corresponding verification module. If verification of output fails, the Conductor sends a request to the Architect, which will develop a plan to fine-tune the expert in question to improve the output.

The iCon architecture is made possible by two of Chariot Technologies Lab's prior patent-protected innovations:

• FIONa, a mathematical method of converting conditional statements into full arithmetic equivalent that allows for accelerated execution of any program code with multiple Boolean conditions as a set of arithmetic operations
• A dynamic processor (made possible by FIONa) configured to automatically adjust its architecture to best fit instructions on a clock-cycle level, increasing execution speed and energy efficiency

The humanity.ai platform built on this technology operates multiple LLMs as well as ML-models and rule-based algorithms: a fine-tuned LLM as a conductor, another open-source model for general knowledge, and 100+ more precise models for math, reasoning, video analysis, and much more, along with verification mechanisms.

Learn more in Chariot Technologies Lab's paper on the innovation:
Modular Hybrid AI Architecture: A Pathway to AGI