A working ledger of AI, economics, and human agency
I am Derek Ozkal. For more than a decade I have funded the research, convened the researchers, and written the field notes. This site is the ledger: the work, the writing, and working prototypes of the new kinds of work frontier models make possible.
The pattern across all of it: point the best researchers at the right questions, then get the answers into the hands of people who decide things.
Kauffman Foundation's flagship research competition on entrepreneurship and the economy. I helped design it and ran it: $12 million deployed across more than 60 projects, selected from a pool of over 1,100 submissions.
Funded work by Brynjolfsson, Bloom, Haltiwanger, McElheran, and Chatterji.
A working relationship with the top applied economists on productivity, labor, and firms. Not a contact list. A feedback loop that runs both directions.
Acknowledged in NBER's Entrepreneurship and Innovation Policy and the Economy, Vols. 1 and 2 (2022, 2023), and by James Poterba. Example acknowledgement: Nick Bloom on NBER Working Paper 33384 (2025).
Participant in the 2021 visioning workshop later credited as foundational to NSF's Regional Innovation Engines program, now a cornerstone of federal place-based innovation policy.
Documented in the program's origin record.
More than 90 convenings organized, chaired, or moderated. The job in each room was the same: translate between the people who produce knowledge and the people who need it.
Presentation and convening partners include OECD, Brookings, NSF, KC Federal Reserve Bank, Upjohn Institute, and NBER.
Current focus: mapping where AI substitutes for tasks, where it complements them, and where it creates work that did not exist before. County-level analysis using occupational exposure data and the Anthropic Economic Index, with special attention to the American heartland.
Working prototypes in the Lab.
Agents and Agency translates research into observations, stories, and a few policy ideas. One question runs through all of it: how do we keep human dignity, judgment, and meaning at the heart of an AI-driven world?
Featured essay
AI capability is compounding. Human cognitive dependency is compounding faster. Institutions are prepared for neither. The essay that named the problem this site keeps working on.
AGENTS AND AGENCY · SUBSTACK
New essays land on Substack first. Subscribe to get them as they publish. No noise between issues.
The Lab
Automation substitutes. Augmentation assists. The third column of the ledger is genesis: tasks, tools, and whole occupations that could not exist before frontier models. The Lab holds working prototypes of that column.
An interactive model of the Liquidity Crisis of Expertise. Set the pace of AI adoption, apprenticeship investment, and encoding fidelity. Watch a profession's knowledge balance sheet over 25 years.
PROTOTYPE 02A living index of jobs that did not exist before frontier models, each entry tagged with the capability that unlocked it and why it was impossible before.
I grew up in Nevada, Missouri. The heartland is not a research subject to me. It is home.
For thirteen years I have worked at the Ewing Marion Kauffman Foundation in Kansas City, where my portfolio spans entrepreneurship research, economic mobility, and the economics of AI. I have been funding research on AI and the economy since 2015, before it was fashionable and long before it was urgent.
The through line is translation. Researchers and decision makers rarely speak the same language. My job has been to sit in the middle and make the exchange work: fund the right questions, convene the right rooms, and write the findings in plain English.
Agents and Agency is where the writing lives now. The Lab is where the arguments become working software.