Writing on AI, labor markets, and the future of work.
A plain-language tour of recent work on huggingfaceR, foundryR, onet2r, and cmapr, with guidance on what each package is for and where to get it.
A finished three-study essay at The Commonplace. A broader 2026 review has no announced publication date.
O*NET is one of the best public sources for what occupations involve, and easy to misuse as a time series. A new R package for reading archived releases, bridging taxonomies, and weighting by employment without fooling yourself.
On craft, curiosity, and what the machines can't replicate. A letter to my son about following excitement, building situated knowledge, and paying attention.
Everyone asks how many jobs AI will destroy. The better question is what happens to the demand for work when the cost of building drops to near zero.
Most real work looks more like playing in a band than reading a book. The next wave of AI needs to learn by living inside situations, sharing the world with us.
If agents can build and do almost anything, why do apps still matter? Exploring the relationship between traversal and commitment in an agent-driven world.