I'm Parth, a quantitative researcher at Orbis Investments. The rest of the time I write data-driven economics. The instinct is the same either way: take a messy question, point data at it, and see what actually holds up.
Most of my thinking ends up on my Substack, Parth's Datastack. A few pieces I'm proud of:
Is AI Writing American Law? (Effort, with Brian Chau): we scanned every bill in Congress since 2023 and found AI-written text in findings and preambles tripled to about 6% by mid-2026, with ~15% of Congressional Record Extensions of Remarks.
Slop Detective: I ran a few thousand Substack posts through an AI detector to see who's publishing machine-written work, and whether readers can tell. I also discussed it on YouTube with University of Hong Kong and Sheffield researchers.
UK AI Economic Indicators: a live dashboard tracking AI's economic footprint in the UK, who's adopting it, whether it's showing up in growth, and how it's reshaping who gets hired. A UK build of Stanford's framework, on public data.