What NEL Is

NEL is a first-principles synthetic cognitive architecture. Its intelligence is not intended to come from prewritten answers or a language model acting as an executive. Instead, behaviour is produced through the live interaction of specialised mechanisms for perception, memory, prediction, uncertainty, value, goal formation, planning, arbitration, action and learning.

The same cognitive architecture can be connected to different worlds. Each world supplies observations, possible actions and consequences; it does not supply NEL with environment-specific cognition. This separation allows the project to ask whether one continuing mind can acquire competence through its own mechanisms across changing environments.

NEL is environment- and embodiment-agnostic by design. A filesystem, simulated maze, key-and-lock world or future physical body should be experienced through different interfaces while preserving the identity and cognitive continuity of the same mind.

How the Mind Develops

NEL follows a recursive cognitive cycle rather than a fixed command script:

Experience → representation → relationship discovery → uncertainty or knowledge gap → hypothesis → prediction → opportunity evaluation → internal value → action selection → consequence observation → revision → generalisation

Goal and plan formation are developing pathways within this cycle. NEL currently evaluates goal opportunities and candidate plans, but autonomous goal adoption and direct plan control of external action are not yet complete.

Later knowledge can change how earlier experience is interpreted. The long-term objective is a mind whose concepts, strategies and behaviour develop from evidence while preserving the origin and history of what it has learned.

Current Research Foundation

NEL currently contains tested foundations across the following areas:

• Perception and structured experience formation
• Episodic, spatial, behavioural and prediction history
• Pattern, temporal and relationship discovery
• Hypothesis formation and outcome prediction
• Knowledge assessment, knowledge gaps and multiple forms of uncertainty
• Information-gain and novelty assessment
• Internally formed value and autonomous goal-opportunity evaluation
• Candidate planning, action arbitration and execution
• Action-consequence observation, behavioural preference and cognitive revision
• Continuous cognition operating through a persistent NELMind
• Generic operation across multiple environments
• Inspection of live architectural components and cognitive telemetry

These mechanisms are not counted as part of the functioning mind merely because code exists for them. NEL development distinguishes isolated tests from mechanisms that are owned by the architecture, observed during live cognition or causally affect live decisions and behaviour.

What Makes the Research Different

• No hardcoded knowledge: facts and relationships should be acquired from evidence.
• No hardcoded behaviour: actions should emerge from the system's current cognitive state.
• No hardcoded goals or tasks: goals should arise from detected gaps, opportunities, values and expected outcomes.
• No predefined reward policy: value is formed internally from experience and consequence.
• No LLM executive: language models do not supply NEL's cognition or decide what it should do.
• One mind across worlds: environments change; the cognitive architecture persists.
• Causal evidence: a claimed capability must measurably alter what NEL can learn, decide or do.

The Cognitive Machinery Limit Hypothesis

Proposed by Charles Dooner
Expanded Definitive Edition · Version 3.0 · September 2026

“The limit of a mind is the completeness of its reasoning machinery, not the infinity of its knowledge.”

The Cognitive Machinery Limit Hypothesis proposes that a finite mind’s long-term developmental reach is constrained not simply by how much knowledge it possesses, but by the cognitive operations through which it can acquire, connect, test, revise and generalise knowledge.

In scientific terms, “completeness” does not mean that any mind can possess universally complete reasoning. It refers to measurable functional coverage, causal integration and revisability relative to declared tasks, evidence and resource conditions.

NEL provides a developing synthetic cognitive architecture through which parts of this hypothesis may eventually be investigated. NEL does not currently prove the hypothesis. It provides a potential experimental system for testing its predictions through mechanism removal, restoration, functional-equivalence testing, cross-world transfer and machinery-growth experiments.

The complete paper defines the hypothesis formally and includes its operational definitions, predictions, competing explanations, experimental programme and falsification conditions.

Current Status and Scientific Boundaries

NEL is an early-stage research architecture with evidence of internally driven value formation, action selection, experience-based learning and persistent cognition across controlled environments. NEL can evaluate internally arising goal opportunities, but autonomous adoption of those proposed goals is not yet complete. It is not presented as artificial general intelligence, a finished synthetic mind or proof of consciousness.

NEL now has tested foundations for cross-world structural detection and structural concept formation. Important capabilities still under development include grounded language, self-directed external information seeking, autonomous goal adoption, proven cross-world competence transfer, richer embodiment, complete use of consequence evidence in future predictions and machinery capable of safely extending its own cognitive organisation.

Every claim is intended to remain testable. A demonstration establishes possibility, not generality. Stronger conclusions require controlled intervention, ablation, restoration, replication and comparison against simpler explanations.

NEL Rover 1 — The Beast

One Mind. A Physical World.

NeuralEdge Limitless is preparing to move beyond simulated and digital environments into its first physical embodiment.

NEL Rover 1, built around the Waveshare UGV Beast platform, will provide the existing NELMind architecture with a physical interface to the world. Cameras, depth sensing, LiDAR, inertial sensing and wheel encoders will provide perceptual evidence, while the tracked drive system and pan/tilt camera assembly will provide physical opportunities for action.

The objective is not to program NEL to behave like a robot. The rover is a body. It exposes perception, possible actions and their physical consequences to the same persistent cognitive architecture already used across NEL's other environments.

NEL will therefore have to encounter the physical world through experience: perceive changes, form memories, discover relationships, make predictions, evaluate available opportunities, select actions through its own cognitive machinery, observe their consequences and allow those consequences to influence future cognition.

The first experiment will be deliberately simple. NEL will receive genuine sensory information from the rover and genuine physical actions will be made available to it. If NEL's own cognitive process selects an action, that decision will pass through the embodiment interface to the rover's hardware and physically alter the world. The resulting sensory change will then return to NEL as new experience.

This is the next expression of the project's central principle:

One Mind, Many Worlds.

Interactive NEL Rover 1 model — drag to rotate · scroll to zoom

Toward a Live Digital Twin

As physical integration develops, this model is intended to become a live digital twin of NEL Rover 1. Physical movement, body orientation and camera pan/tilt can be reflected through telemetry from the real rover, allowing observers to watch NEL's embodiment operate in real time.

The public interface will remain observational. Telemetry can report what the rover and NEL are doing without allowing the website to determine NEL's decisions or control its physical actions.

Research Progress

A concise record of NEL's transition from earlier AI experimentation to the present first-principles cognitive architecture.