One body of work, built around a single question.
How do you tell a genuine AI investment process from a repackaged one? Answering that produced a protocol, the evidence behind it, an index, a quarterly report, a set of practitioner tools, and the company it became. They connect, and they are all here.
Frameworks & Protocols
A falsifiable due-diligence standard and the evidence map behind it.
SPEC
A four-axis, architecture-agnostic, falsifiable due-diligence protocol for allocators evaluating AI-driven managers. Includes the paper, a 68-question diagnostic library, the scoring sheet, and replication pseudocode.
OpenSPEC Research Catalog
Everything that breaks an AI investment manager, mapped to the research and organized by SPEC’s four axes. Every row is anchored to its source.
OpenIndex
An original ranking that did not exist before.
Report
Recurring research for CIOs and investment committees.
Tools
The frameworks turned into things allocators can use in a meeting.
The SPEC Test
Four questions to ask a quant manager in your next meeting, with guidance on how to read the answers.
OpenSPEC Process Diagnostic
11 questions, 5 minutes. An AI-powered read of your evaluation process across all four SPEC dimensions.
OpenSPEC Manager Assessment
Tailored SPEC questions before the meeting, and a structural analysis of the transcript after it.
OpenWriting
Where the thinking gets published.
The Specification
My Substack translating frontier AI research into practical tools for institutional allocators.
VisitCFA Institute, Enterprising Investor
A four-part series on quantitative due diligence for the CFA Institute research blog, published between March and September 2026.
OpenMaking allocation decisions where AI is a factor?
I work with a small number of institutional clients on manager evaluation, AI governance, and the frameworks above.
Work With MeStay ahead of AI in investing, in two hours a month.