IAA™ · Assisted Artificial Intelligence
A place to experiment with thinking systems
Delegating decisions to AI is risky
Reasoning with it is powerful
MindLoop is an experimental space to explore how humans and AI systems can think together about complex problems.
Ideas before products. Reasoning before automation.
Each MindLoop project is a documented experiment. The goal is not to deliver answers, but to investigate how decisions can be better structured.
IAA™ — Assisted Artificial Intelligence
AI as an object of study
Assisted Artificial Intelligence is a concept created to guide the conscious use of AI. At MindLoop, AI is not given responsibility nor does it make decisions. It acts as an amplifier of human thinking and as a means of accessing the knowledge accumulated by humanity.
Human
Responsible for meaning and decision-making. Formulates hypotheses, interprets results, evaluates consequences, and fully assumes responsibility for choices made.
IAA™
Acts as cognitive support. Organizes information, retrieves knowledge, suggests reasoning paths, and makes relationships explicit to support human thinking.
Experimental Loop
A space of continuous learning between human and AI. The value lies in iteration, reflection, and refinement of reasoning — not in the final answer.
What makes MindLoop?
An environment to responsibly test complex ideas.
Projects as portfolio
Each experiment is a documented artifact of learning.
Context engineering
We investigate how context shapes AI-assisted decisions.
HITL as a premise
Humans remain at the center of all decisions.
Responsible exploration
No promises of automation or human replacement.
Our Culture
Thinking before building, always grounded in real problems.
Applied research
Each project starts from a concrete problem and is treated as a practical experiment.
Technical humility
AI is not an authority or a ready-made solution — it is a tool to support human decisions.
Verifiable learning
The value lies in documented processes that are tested and applicable to real contexts.
LoopynLATS
Decision trees that know when they are uncertain
LoopynLATS is a probabilistic decision tree engine built for real-world systems — where decisions are rarely black or white.
It measures uncertainty explicitly, escalates ambiguity to humans, and records every decision for full auditability.
Core Idea
If a system is uncertain, it should say so — and ask for help.
LoopynLATS turns uncertainty into a first-class control signal using probability, entropy, and Human-in-the-Loop governance.
Who builds MindLoop
A portfolio built on real ideas.
Here you access ideas in progress: projects, texts, articles, reflections, and prototypes I develop over time. MindLoop works as a living extension of my thinking.