AI Research & Cognitive Infrastructure

Building the bridge between
how humans think and
how models reason

Iron Sun Works develops original methodology for AI calibration, human-AI reasoning divergence research, and cognitive infrastructure. The work sits at the intersection of alignment, cognitive science, and practical system design.

See the research
Research

Original Work

Current AI alignment approaches capture what humans prefer but discard why they prefer it. Iron Sun Works develops methodology that preserves the reasoning — and uses it to build richer signal for model calibration than preference ranking alone can provide.

01

Calibration Corpus

A structured methodology for capturing how human cognition diverges from model reasoning — preserving both sides' logic, not just the outcome. Where RLHF captures that a human preferred output A over B, the corpus captures why. That reasoning is where the information about human cognition lives. At population scale, individual corpora surface generalizable patterns in model miscalibration.

02

Model-Side Process Visibility

Applying process trace analysis to model generation — branching probabilities, confidence patterns, and hesitation signatures the model can't self-report. Skilled interpreters reading these traces identify sycophancy, confabulation, and over-trained caution at the process level. Complementary to mechanistic interpretability — phenomenological rather than architectural.

03

Metacognitive Cadre

A self-selecting recruitment mechanism for the humans the research requires — people who can articulate their own cognition and diagnose model reasoning failures. The system identifies its own future interpreters from people already using it, inverting the typical platform/user relationship from extraction to recognition.

04

EF Support Platform

An AI-assisted executive function support system grounded in neuroscience. The platform infers cognitive state and provides scaffolding — task decomposition, context restoration, thread reconstruction — without requiring the user to self-report or configure. Cognitive modality as a first-class design constraint.

Product

AI-Native Knowledge Infrastructure

AI enables cross-domain thinking at a scale that exceeds human organizational capacity. A single conversation can traverse neuroscience, software architecture, and philosophy — and the AI keeps up. The human can't. Not because they lack capability, but because working memory has limits and sessions are stateless.

We close that gap. AI-native cognitive infrastructure — externalized memory architecture that gives AI context persistence across sessions, captures unstructured thought with zero friction, and builds a knowledge graph that surfaces connections human memory can no longer hold.

Cross-Session Continuity

AI sessions are stateless. The system gives them memory. Context files, session protocols, and handoff mechanisms preserve project state across any number of sessions.

Zero-Friction Capture

Structure at the point of capture loses knowledge in proportion to the cognitive cost. The system captures everything unstructured, then routes and connects it after the fact.

Knowledge That Compounds

Every piece of information becomes more valuable through its relationship to what's already there. Bidirectional links and concept nodes surface connections across domains and time.

AI-Native Architecture

Designed for AI orientation, not human navigation alone. Every structural decision is optimized for an AI collaborator to pick up where the last session left off.

About

Iron Sun Works

Founded by Andre Foti and Jana Cook. Ohio LLC. Iron Sun Works builds cognitive infrastructure and develops original AI calibration methodology — systems and research that make AI-augmented thinking sustainable and AI reasoning more aligned with human intent.

The name comes from spectroscopy. When sunlight passes through the sun's atmosphere, iron atoms absorb and re-emit specific wavelengths — creating dark lines in the spectrum that reveal the sun's composition. NASA's Solar Dynamics Observatory captures these emissions directly, producing images of the corona at iron wavelengths that show magnetic structure invisible to the naked eye.

The parallel is the work: making invisible cognitive structure visible. What's happening in the gap between human intent and model output. What's happening in the gap between a thought and its organization. Iron in the sun, doing the work you can't see.

NASA SDO image of the sun at 171 angstroms (Fe IX) with magnetic field lines overlaid

171 Angstroms — Iron IX

The sun's corona imaged at the spectral line of iron atoms that have lost eight electrons. False-color gold. The magnetic field lines — overlaid from helioseismic data — show structure that's always there but only visible at the right wavelength.

Image: NASA/SDO/AIA/HMI — Public domain

Get in Touch

Contact

Interested in the research, the methodology, or potential collaboration?

andre@ironsunworks.com