Jeremy Howard is an Australian data scientist and educator, co-founder of the research lab Answer.AI and of fast.ai. In 2024 he proposed llms.txt, a convention for a curated, machine-readable manifest that tells a language model what to read at inference time.
Significance
Howard’s proposal extends the agent-readable web standards lineage — robots.txt, sitemap.xml — from crawling to inference, inheriting the Sitemaps pattern of publishing an operator-curated manifest. It is part of the protocol surface of the Independent Internet, and a current example of how operators reclaim control over how machines read their sites.