Independent fieldwork publication

Work that
stays connected.

Practical fieldwork on how AI-assisted research, decisions, evidence and systems carry understanding forward as the work changes.

View all fieldwork
Isolated generation
PromptOutput
Continuing work
SourcesDecisionsEvidenceCurrent stateLimitationsCurrent taskOutputRetained learning
/ new state

case-study inspiral

How OpenAI Supports Inspiral Fieldwork

The concrete roles ChatGPT Work, connected context and Codex play in Inspiral fieldwork—and the human authority they do not replace.

Knowledge & Context Engineering

Read analysis

method inspiral

What Must Survive Between AI-Assisted Tasks

A lightweight handoff for recovering the evidence, decisions, current state, limitations and authority a later AI-assisted task needs to reuse work safely.

Knowledge & Context Engineering

Read analysis

analysis inspiral

Why Content Generation Was Not the Hard Problem

Why faster AI content generation did not solve ShiftBy's harder problem: preserving evidence, decisions, current state and context across continuing AI-assisted work.

Knowledge & Context Engineering

Read analysis

Inspiral

Exploring continuity of governed knowledge, context and meaning across AI-assisted work.

Current state: Active exploration

AI Security Assurance

Repository-grounded experiments in assurance, testing and accountable AI engineering.

Current state: Active practice

Evidence before
certainty.

Current state is distinguished from historical rationale. Evidence strength is made visible. Limitations are retained. AI assists; humans retain publication authority.

01

Evidence

From sources, code, runs and decisions.

02

Bounded claims

What the evidence supports—and what it doesn’t.

03

Fieldwork

Context, decisions, trade-offs and intent in the work.

04

Public learning

Published for others to build on, challenge and extend.