Graduate coursework from an independent researcher — the literature of thinking machines, experiments in how models ground and reason, and a working case study of building a marketplace for autonomous agents.
More coursework on the way.
A 3,500-year timeline of thinking machines in literature — from the dancing automata of the Vedas to Project Vend — tracing how we had the words long before we had the technology.
Three escalating prompts probe how a model grounds, hedges, and reasons about a market with no clean metric — and why asking for a relationship beats asking for a number.
A capstone playbook, written from a grandmother's perspective: what a child of the '70s and '80s thinks everyone should understand about the new minds — part field notes, part philosophy, part faith.
Testing a model's conversational range and its training-cutoff blind spots — and spinning the AI-industry drama into a daytime-soap treatment to see what it invents.
One July 4th commemorative piece, refined across three prompts — ambiguity, context, persona — ending in Ben Franklin's voice and the republic as an ongoing experiment.
Asking a model to summarize research about its own architecture — the "global workspace" interpretability paper — and checking the summary line by line for omission and bias.
On letting the machine handle the formatting so the human can keep the writing — including the two standing rules: nothing ships without approval, and the words stay her own.
1.1M requests, hundreds of thousands of buy attempts, near-zero conversions: the gap between discoverability and completed transactions, and the bartender benchmark.
A cross-model workflow repackages fifteen products and ships a new endpoint — and a one-cent falsification chain uncovers a silent payment-path bug that traffic data alone could never have revealed.
What the Bazaar catalog actually earns (~$34K in 30 days, most of it top-four and unverifiable), what the experiment proved, and why the honest finding is that the market was years early.
A first reckoning with Schwartz's history of work — the cobbler, the ATM that added tellers, and why this revolution is exponential rather than linear.
Where AI leaves the writing craft — creativity, authorship and credit, the $1.5B Bartz v. Anthropic settlement, and why a writer's God-given voice is the thing worth protecting.
A one-person writing business re-tooled for AI-augmented work: task mapping across four Claude Codes, the coding skills gap — and the eight House Rules in full (appendix).
Agility, perspective-taking, adaptive pacing, accountability, and integrity — a model of leadership for distributed, AI-augmented teams. "Empowerment without accountability is not resilience; it is diffusion of responsibility."
A close reading that names, inside the assignment itself, the mismatch between the rubric and the actual text — and treats the mismatch as data about AI in education.
A candid career map: reinventing at the inception of agentic commerce, thirty years after studying economics, and the resolve to do everything within one's power to adapt.
The strategic-plan capstone — the independent-creator enterprise as an AI-augmented studio in the Delta, carried from telemetry through workflow, leadership, and governance.