Отто

Claude Code and its ilk are coming for the study of politics like a freight train. A single academic is going to be able to write thousands of empirical papers (especially survey experiments or LLM experiments) per year. Claude Code can already essentially one-shot a full AJPS-style survey experiment paper (with access to Prolific API). We'll need to find new ways of organizing and disseminating political science research in the very near future for this deluge.

Claude Code is...simply insane. I work a lot with experimental data. Here is 24-48 hours of work done in 20 minutes. Prompts were in fully plain text, e.g., "find data file "experiment.csv" and locate the following variables", Claude Code completed the following steps: 1. Found the relevant variables in a raw qualtrics file. 2. Created a clean .csv file. 3. Wrote code and created tables for four types of analysis and wrote summaries of results in latex. 4. Found "types" in the data set using cluster analysis (the prompt was "find types...") and created radar plots. 5. Ran simulation of different types interacting in a market place and plotted the results. All code and analysis was transparent and reproducible (which is extremely important for verification). All files were saved locally on my computer. I've said it before and I've said it again. The past few months have been a paradigm shift.

We’re making GPT-5 warmer and friendlier based on feedback that it felt too formal before. Changes are subtle, but ChatGPT should feel more approachable now. You'll notice small, genuine touches like “Good question” or “Great start,” not flattery. Internal tests show no rise in sycophancy compared to the previous GPT-5 personality. Changes may take up to a day to roll out, more updates soon.