UNN Lab
The Unnecessary Lab
Serious Curiosity. Unnecessary Questions.
An independent archive for AI, decisions, complex systems, and untidy human judgment.
Questions that do not need to be useful yet.
The lab
UNN Lab is a place for independent curiosity and applied research. We follow an idea from source to interpretation, hypothesis, experiment, evidence, and unknowns.
We don't build software to make every question useful. We build a place to ask better ones. The public toilet is not currently a workstream.
How a file moves
From a strange question to a testable next step.
Source
Name the paper, observation, conversation, or lived question that started the file.
Interpret
Separate what the source says from the meaning we are tempted to add.
Hypothesis
Turn the interesting tension into a claim that could be wrong.
Experiment
Choose the smallest comparison, prompt, or observation that can teach us something.
Unknowns
Keep the limits, counterexamples, and open questions visible in the file.
Open research files
Current investigations
3 published files are on the table, with 4 recent notes showing where the work is moving. Each file is a working object; none are pretending to be finished because the page has a nice title.
Indra's Net: From Fazang to Relational Systems
Can missing structural properties between older conceptual models and modern systems reveal useful engineering questions?
The Cost of Being Wrong
How should a decision system represent uncertainty when the cost of one kind of mistake is much larger than another?
Opportunity Lenses: Seeing Connections Before They Are Obvious
Can looking at a problem through several deliberately different lenses reveal opportunities that ordinary market research misses, and can we tell honestly whether it did?
One file in focus.
UNN-001 / featured file
Indra's Net now asks what modern systems reproduce from an older relational model, and what they leave out.
One branch studies relational decisions. A second explores local/global information structure and computational efficiency. The metaphor remains a source of questions, not proof of prediction.
Open the research fileA diagram for thinking with relationships. It is a prompt for the experiment, not a map of the world.
Recently filed
Loose notes from active files
A research archive should show movement before it shows polish. These are the latest questions, boundaries, and counter-hypotheses added to the files.
The resemblance is not the result. The gap is the experiment.
Golden Mac Experiment 001 begins with a reproducible full-context baseline, not a speedup claim.
The discussion was triggered by real-world puzzles. None of those details are published here; only the questions are.
Opened as a map of regions, not a product. Candidate regions: multi-lens reasoning, a curiosity agent, market access, trust paths, demand signals. None is chosen yet.
Follow one thread.
A small live instrument
Choose a lens and watch the question change shape. This is a reading aid, not a model pretending to have measured the answer.
A fact is rarely alone
Context
What changes when a decision system keeps the relationships around an answer?
More context can preserve meaning. It can also turn a useful answer into noise.
Compare isolated retrieval with a small context-linked graph.
Visual notes, not evidence.
Field notes
The Lab borrows the quiet visual language of Sisi's existing working room. These images set the atmosphere; they do not support a research claim.




Research meets decisions.
The wider SisiTheFox system