Skip to Content

2026

135 posts
Previous Posts

Deep Research on a Loop: Using AI Agents to Construct Economic Datasets

2026ToolingWorking
The AI agent receives a research instrument and a unit assignment, then autonomously decides what to search for, which sources to consult, how to evaluate conflicting information, and how to code the result. The agent performs discovery, finding relevant sources alongside extraction and coding. The input is a research question. The output is an auditable research record with full citations.

NBER paper on using AI agents to construct economic datasets. Same methodology I used to develop the Breakouts dataset for On_Discourse.

How Go Players Disempower Themselves to AI — LessWrong

2026Reading
My view of AI use (especially cheating) in Go originally manifested as disgust for its practitioners. I switched eventually to an attitude of compassion and pragmatism towards a habit that was clearly much more vulgar and weak than it was evil. Over time, I have progressed to feeling deep sadness for a group that surrenders much of what it claims to value. The thing I want to impress with this article is the consistency with which we as a species underestimate our own willingness to give up our culture, economy and autonomy to AI, even without monetary incentives. For this to happen, AI does not even need to be superhuman. Indeed, Go AI automates human players’ role in culture as shallow simulacra. All an AI needs to do is be passably good at a task and that may well be enough for people to volunteer their own replacement.

LessWrong post on how Go players limit themselves when utilizing AI for post-game review. Compare to a previous paper in 2023.

Load More You've reached the end of the list
Free

Subscribe to Newsletter — Stay up to date with email notifications of new posts.