Pricing Perspective
No returns: Mega-IPOs and the promise of AI as economic measurement
Mega-IPOs, the convergence of public and private markets, and the case for a new approach to financial measurement.
Writing
Working notes on research, software, and evidence
The decisions behind quantitative work: choosing models, evaluating experiments, reproducing results and understanding the opportunities and uncertainty introduced by AI.
Pricing Perspective
What are we actually studying when we study AI in finance?
AI in finance plays three distinct roles — a tool for financial practice, a source of economic change, and a means of measurement — and each demands different evidence.
Pricing Perspective
Do language models in finance need a Markowitz moment?
Language models read individual companies remarkably well; the harder question, as Markowitz taught finance, is what that information means together inside a portfolio.
Pricing Perspective
Reproducibility is a Graph
A reproducible result is not a folder of code. It is a dependency structure connecting data, assumptions, transformations, software, environments, and outputs.
Pricing Perspective
The Tidal Wave of AI Research in Finance
AI has expanded what can be measured in finance. It has also expanded the space of plausible-looking mistakes: a note on scale, validation, and methodological discipline.
Pricing Perspective
Vintage Reproducible ET Tax Models
Analytical results change when data, policy rules, assumptions, and software change. A practical note on preserving vintages so historical results stay reconstructable.
Hugging Face
Models are Markup, Tokens are Features
A practical account of treating model structure as markup and token-specific adaptation as a feature-selection problem.
Solenya
Beyond Average Revenue
An experiment-analysis perspective on commercial impact beyond a single average treatment effect.
Solenya
Architecting Solenya API V1
Design choices behind an applied-AI API and the infrastructure around it.
Solenya
What is Zero-shot Discovery?
An introduction to finding useful structure without task-specific labelled examples.