WRITING · NINE ESSAYS, EVIDENCE-FIRST
Essays that carry their evidence.
Each essay names its audience and archetype before it is written, and every claim chains to a measured artifact. Every spec is published here: the acceptance criteria, the objection list that rejects it, and the figures it will cite.
9 specified, 0 with prose in draft, 9 published
Open any essay to read its spec: what it must hit, what would reject it, and the artifacts it rests on. Where prose is published, the essay is there.
Start here · essay 01
The personal agentic operating system
A single operator can run a JARVIS-class life-and-work OS, if the agent is given real utilities rather than prompts. The OS is the integration layer, not the model.
- 05
Ideal tooling infrastructure: choosing and building tools
publishedTool and language choice are architecture decisions, not preferences. The substrate follows the workload; identity never picks it.
- 08
The autonomous operating system: self-healing agents that learn
publishedAutonomy is an operations problem. Deterministic automation does not self-heal; only a system that detects its own degradation does.
- 02
Learning about AI: preparing for the agentic era
publishedAI literacy is not tool training. It is learning to direct systems, which requires understanding what the systems cannot do.
- 06
Ideal agentic harness architecture
publishedThe harness is where context, tool schemas, memory, and permissions compete for the same budget. Architecture is the allocation policy.
- 03
Learning through AI: redefining the education system
publishedThe unit of education is not the course, it is the assessment loop. AI makes continuous diagnosis cheap, which makes the cohort model obsolete.
- 07
Research engine and KB architecture for autonomous R&D
publishedSelf-referential research loops fail on provenance, not on generation. An agent that cites itself has no way to detect its own drift.
- 04
The return of the renaissance: collapsed cost of discovery
publishedWhen synthesis and infrastructure operation approach zero cost, the bottleneck moves from access to judgment, and who can do research changes.
- 09
AI and consciousness: the observer in the loop
publishedAugmentation is not cognition. The harness allocates context, tools, memory, and permissions. It cannot allocate judgment. The observer is the one component no harness allocates.
How these publish
Each essay names its audience and archetype before it is written, and its evidence spine is resolved before drafting starts. That is why the statuses above separate spec complete from in draft, and why the list is shorter than a promise would be.
Every figure an essay will cite is already registered in the measurement record, with its method and date.