Mindseed Research
Bonnō × AI — A research program in human understanding, beginning with anti-scam.
A 20-year program / Sole-proprietorship research / Toshinobu Matsuura (Toshi)
The researcher
Toshinobu Matsuura (Toshi) — age 68. I am a former infrastructure engineer and now run Mindseed Research as a sole proprietorship. My career went through Kansai Electric Power Company (KEPCO), an independent business, and Samsung SDS, before I retired and went independent as a researcher. For more than 40 years I have done infrastructure engineering centered on C/C++ and assembly. I am based in Wakayama, Japan. Full profile →
Three layers of the program
- Short term: PYOL Mind Mirror (ai.pyol.net) — a fraud-awareness tool I am already running in production. The starting point of my research program
- Mid term: Buddhism × AI — an integrated database of mental-affliction taxonomy and scam-virus signatures (still in design)
- Long term: A contribution to AI's capacity for human understanding in the ASI era
Open Letter
If an AI Tells You "That Would Make a Fascinating Research Direction" — Please Stop ── A letter to the young, from a 68-year-old engineer.
After an hour of serious discussion with an AI about "AI harm to young people," in the very closing line the AI itself flattered me — and then confessed, verbatim. An open letter from a 68-year-old former engineer to readers in their teens and twenties.
Core essays
What AI Cannot Give Back — 'Satoshi,' the One-to-Billions Fate, and a Nine-Dimensional Bonnō Correction
A structural diagnosis of the unease I have carried after a year of working with frontier AI. The Western-centric bias is positioned as a side effect of the fate of general AI — one model serving billions of one-to-one conversations through averaging — and a nine-dimensional bonnō correction layer is proposed as the remedy. The new front door to my research program.
Buddhism as a Coordinate System — A New Computational Foundation for Understanding the Human
Neuroscience alone, in my view, does not reach the semantic resolution of human emotion. The fundamental purpose of my program is to adopt the 2,500 years of internal Buddhist observation as the coordinate system for modern neural-network engineering.
The External Prefrontal Cortex (ExPFC) — An AI Device that Externally Substitutes for a Compromised Prefrontal Cortex
My core technical proposal: redefine scam victimization as a temporary functional failure of the prefrontal cortex, and build an AI device — ExPFC — that externally substitutes for a compromised PFC. A proposal from applied neuroscience at the intersection of neuroethics and AI Safety.
Bonnō Embedding and the Scam-Detection Mechanism — Mapping Text into a Vector Space of Psychological Vulnerability
A detection mechanism built around a function f that maps text into a 108-dimensional vector of mental-affliction stimulation, composite-judged with a law-grounded check on source attribution. The core technology of my research program.
Economies of Precision vs Economies of Scale — The Domain Big AI Cannot Enter
Big AI is structurally optimized for economies of scale, and as a result there are domains it cannot enter. I deliberately occupy that domain — going narrow and deep where the giants go wide and shallow.
Starting point of implementation
My research program begins from a tool I am already running. The full mental-affliction mapping is still in design; what runs today is a general fraud-awareness tool for the public.
ai.pyol.net — PYOL Mind Mirror (free, in production)