{"id":37980,"date":"2026-03-15T18:18:13","date_gmt":"2026-03-15T18:18:13","guid":{"rendered":"https:\/\/www.duck9.com\/blog\/?p=37980"},"modified":"2026-03-15T14:18:49","modified_gmt":"2026-03-15T18:18:49","slug":"sxsf-markdown-qualifies-as-code","status":"publish","type":"post","link":"https:\/\/www.duck9.com\/blog\/sxsf-markdown-qualifies-as-code\/","title":{"rendered":"SXSF Markdown Qualifies as Code"},"content":{"rendered":"<div class=\"postie-post\">\n<div>\n<div dir=\"ltr\"><a href=\"https:\/\/x.com\/larrychiang\/status\/2033245506772643919?s=43\">https:\/\/x.com\/larrychiang\/status\/2033245506772643919?s=43<\/a><\/div>\n<p><br id=\"lineBreakAtBeginningOfSignature\"><\/p>\n<div dir=\"ltr\">\n<div>\n<div>&#8211; Garry Tan, Y Combinator CEO, argues Markdown qualifies as code by encoding AI workflows, prompts, and processes in files that direct LLMs, as exemplified in his gstack tool for Claude Code setups.<\/div>\n<div><img decoding=\"async\" alt=\"image0.png\" src=\"https:\/\/www.duck9.com\/wp-content\/uploads\/2026\/03\/image0-8.png\"><img decoding=\"async\" alt=\"image1.png\" src=\"https:\/\/www.duck9.com\/wp-content\/uploads\/2026\/03\/image1-3-scaled.png\"><img decoding=\"async\" alt=\"image2.jpeg\" src=\"https:\/\/www.duck9.com\/wp-content\/uploads\/2026\/03\/image2-5.jpeg\"><img decoding=\"async\" alt=\"image3.png\" src=\"https:\/\/www.duck9.com\/wp-content\/uploads\/2026\/03\/image3-2.png\"><\/div>\n<div>&#8211; The post&#8217;s image critiques GitHub&#8217;s prompt-based assumptions versus traditional engineering, highlighting LLM-driven evaluation infrastructure as a scalable alternative to stack traces and unit tests.<img decoding=\"async\" alt=\"image4.jpeg\" src=\"https:\/\/www.duck9.com\/wp-content\/uploads\/2026\/03\/image4-6.jpeg\"><\/div>\n<div>&#8211; Replies underscore practical shifts, like storing product specs and agent instructions in Git for PR reviews and CI validation, enabling teams to iterate 2x faster as AI and documentation co-evolve.<\/div>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n","protected":false},"excerpt":{"rendered":"<p>https:\/\/x.com\/larrychiang\/status\/2033245506772643919?s=43 &#8211; Garry Tan, Y Combinator CEO, argues Markdown qualifies as code by encoding AI workflows, prompts, and processes in files that direct LLMs, as exemplified in his gstack tool for Claude Code setups. &#8211; The post&#8217;s image critiques GitHub&#8217;s prompt-based assumptions versus traditional engineering, highlighting LLM-driven evaluation infrastructure as a scalable alternative to stack [&hellip;]<\/p>\n","protected":false},"author":1,"featured_media":37981,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[1],"tags":[],"class_list":["post-37980","post","type-post","status-publish","format-standard","hentry","category-uncategorized"],"post_mailing_queue_ids":[],"_links":{"self":[{"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/posts\/37980","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/comments?post=37980"}],"version-history":[{"count":0,"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/posts\/37980\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/media\/37981"}],"wp:attachment":[{"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/media?parent=37980"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/categories?post=37980"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.duck9.com\/blog\/wp-json\/wp\/v2\/tags?post=37980"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}