READING LIST

Editorial digest

A small reading list for each edition. Dates describe editorial publication, not the age of every resource.

Subscribe with RSS ↗

2026-W44 · 2026-10-29

Choose, inspect and evaluate research AI

Six starting points for screening, transcription, Chinese text analysis and research agents. The linked sources have been read; tools have not been tested by this site. Method readings explain what to check rather than offering a performance ranking.

ASReview

Open-source active learning for systematic-review screening; the model re-ranks records as you label them.

Screen papers · LiteratureSource checked

noScribe: local interview transcription

Free desktop app by a sociologist that transcribes interviews entirely on your computer with Whisper, identifies speakers, and includes an editor for corrections; about 60 languages.

Interviews & codingSource checked

cntext (Chinese text analysis for social science)

Python library from the Chinese text-as-data community (textdata.cn) with Chinese sentiment dictionaries, word embeddings, dictionary expansion, semantic projection and an LLM annotation helper.

Text & data analysisSource checked

Evaluating research agents

A research report organizes agent evaluation around methods, real-world settings and the people doing the evaluation.

Research agents · Text & data analysisSource checked

Choosing AI tools for real tasks

Ethan Mollick distinguishes chat, hosted agents and desktop agents through examples of knowledge work.

Research agents · Writing & referencesSource checked

Designing for human and AI collaboration

Gabriel Unger and Arvind Narayanan outline a research agenda for AI that supports human work.

Research agents · Surveys & experimentsSource checked