Derivation Logs
Every claim in QIF is traceable to a specific derivation step. These logs document the reasoning chains, decision points, and discoveries as they happened, in real time, append-only, and never edited after the fact.
QIF Derivation Log
Framework reasoning, hypothesis derivation, equation proofs, and architectural decisions. The complete intellectual timeline from OSI rejection through hourglass validation, BCI limits equation, and neurorights mapping.
→ View on GitHubQIF Field Journal
First-person research observations at the intersection of neurodivergence, synesthesia, and computational neuroscience. Raw voice, unpolished, documenting the experience of thinking rather than the thoughts themselves.
→ View on GitHubNeurowall Derivation Log
Signal pipeline engineering decisions for the neural firewall simulation. Covers coherence metric design, multi-band EEG generation, attack simulation, and ROC analysis.
→ View on GitHubEntry Classifications
Every derivation log entry is tagged with its epistemic status. This tells you how much weight to put on the claim.
Grounded in established physics, peer-reviewed literature, or empirical data. The reasoning chain is complete and checkable.
Logical extension from verified premises. The inference is sound but the conclusion has not been independently confirmed.
Novel conjecture with theoretical motivation. May be wrong. Explicitly labeled so readers know this is speculation, not established fact.
The insight has been translated into code, configuration, or documentation. The implementation can be inspected directly.
A previous entry was found to be wrong. The correction entry explains what changed and why. Original entries are never modified.
The Append-Only Policy
These logs follow strict rules designed for academic transparency and intellectual reproducibility.
Never edited after the fact
If an entry is later found to be wrong, a new CORRECTION entry documents what changed and references the original. The original entry stays untouched.
Never deleted
Wrong ideas are part of the intellectual record. Seeing how thinking evolved is as valuable as the final conclusions.
Timestamped in real time
Entries are written as insights occur, not retroactively. This preserves the actual discovery timeline and prevents hindsight bias.
AI collaboration is transparent
Every entry documents which AI systems were involved and what the human decided. AI cannot author entries; all final decisions are human.
Entry 7
The Hourglass Model
Where the 3-1-3 architecture was first conceived
Entry 35
The Black Hole Security Principle
Hawking/Susskind/Maldacena applied to BCI security
Entry 50
TARA: Therapeutic Atlas
Registry reframed from threat catalog to dual-use mechanism atlas
Entry 60
BCI Limits Equation
Unified physics-constraint system for BCI hardware limits
Entry 70
Wearable Neural Firewall
Secure edge architecture for non-intrusive BCI
Neurowall 001-010
Coherence Metric to ROC Analysis
10 entries building a neural firewall simulation from scratch