Blog – Benjamin T. Vincent

Shipping a major feature in `pathmc` in a day: a case study in agentic engineering

software

ai

engineering

A case study in using coordinated AI agents to design, implement, review, document, and release a second causal estimator in a single day.

Benjamin Vincent

You Don’t Understand the ‘Chart of the Century’

Baumol’s Cost Disease and the Arithmetic of Diverging Costs

economics

simulation

interactive

Why healthcare and education get more expensive while televisions get cheaper, and why this is nobody’s fault.

Why you need more than one lift test per channel

marketing

bayes

experimentation

Imagine you ran a clean lift test on one of your channels. The experimental estimate is precise, the MMM calibration works, and the obvious stakeholder question arrives:

Benjamin Vincent

Goodness of fit is not the objective in causal-first projects

Maximizing for decision outcomes, not just predictive performance

causality

marketing

decision-making

Imagine an online shop deciding whether to spend more money showing ads to people who have already visited its website. That practice is often called retargeting (also…

Benjamin Vincent

Bayesian models + grammar of graphics + Python = ❤️

python

bayes

data-viz

Bayesian modelling in Python is pretty solid: We’ve got PyMC for model building and inference, ArviZ covers diagnostics and plotting, and we’ve got helpful pacakges like Bambi…

Benjamin Vincent

Causal Inference: have you been doing science wrong all this time?

python

bayes

causality

If you are a researcher in academia or industry, depending where you are in your causal journey, this post may blow your mind! Its primary purpose is to convey that if you…

Benjamin Vincent