AI won't fix venture's bias. It'll launder it.
june 2026
The machine learns who got the yes, not who was right.
./writing
substack
Five venture takes I'm willing to be wrong about
may 2026
On the companies that shouldn't raise venture, the path into VC nobody recommends, and the prestige flow reshaping both.
You shouldn't learn how to be a VC yet
april 2026
Notes from an accidental internship, and what it taught me about venture capital
selected research
who gets funded: gender and follow-on capital in indian startups
2026 · 249 hand-collected startups
Four specifications on a dataset built by hand, testing whether the funding gap survives sector and state fixed effects. It does not, within firms that already raised.
navigating the margins: street food vendors in new delhi
2024 · laidlaw fellowship
Forty vendors surveyed on foot, on licensing, bribery, and informal credit. The barriers to formalization are structural rather than individual.
capabilities at term sheet time: rethinking venture as development
2025 · thomson reuters ventures
Amartya Sen read against a venture fund's deal flow. Venture decides which problems get treated as solvable, which is a development question whether or not anyone files it as one.