Tag
The author found that constraining an AI agent's output to a fixed, boring format improved reliability more than using better prompts or models, offering a cost-effective technique for developers.
This paper demonstrates that output formats confound data quality metrics and model capability assessments in instruction tuning, causing significant accuracy shifts and rendering current practices ineffective without interface-aware adjustments.
The article argues that HTML is a superior output format for AI agents compared to Markdown due to richer information density, visual clarity, ease of sharing, and two-way interaction, and shares why the author and others at Claude Code prefer HTML.