<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>汀禾新 · 研究与现场笔记</title><link>https://tinghexin.com/research</link><description>真实研发、创作与实验记录</description><language>zh-CN</language><atom:link href="https://tinghexin.com/research/rss.xml" rel="self" type="application/rss+xml"/><item><title>AI视频开工前，先把故事、参考和镜头材料整理好</title><description>先写清作品要表达什么，再给参考材料和镜头各自的位置。一份开工清单，帮助后续生成、比较和修改有据可循。</description><link>https://tinghexin.com/research/prepare-an-ai-video-project</link><guid isPermaLink="true">https://tinghexin.com/research/prepare-an-ai-video-project</guid><pubDate>Sat, 24 Oct 2026 00:00:00 GMT</pubDate></item><item><title>一首歌准备交付时，如何整理音频版本、封面、信息与反馈</title><description>给每次交付留下明确的音频、材料和办理记录，收到反馈后继续同一件事。来自音乐工作台实践的整理方法。</description><link>https://tinghexin.com/research/organize-music-versions-and-feedback</link><guid isPermaLink="true">https://tinghexin.com/research/organize-music-versions-and-feedback</guid><pubDate>Sat, 24 Oct 2026 00:00:00 GMT</pubDate></item><item><title>互动原型先做多大？用一个完整操作循环验证想法</title><description>把进入、操作、反馈、失败和再来一次做完整，再观察玩家卡在哪里。以电池托盘实验说明如何缩小原型范围。</description><link>https://tinghexin.com/research/test-an-interactive-prototype-loop</link><guid isPermaLink="true">https://tinghexin.com/research/test-an-interactive-prototype-loop</guid><pubDate>Sat, 24 Oct 2026 00:00:00 GMT</pubDate></item><item><title>为什么 AI导演工作台没有采用传统节点式工作流</title><description>创作材料之间的关系，是否应该天然等同于执行顺序？</description><link>https://tinghexin.com/research/canvas-not-workflow</link><guid isPermaLink="true">https://tinghexin.com/research/canvas-not-workflow</guid><pubDate>Thu, 24 Sep 2026 00:00:00 GMT</pubDate></item><item><title>当 AI 扩大执行力，小团队真正稀缺的是什么？</title><description>当“做出来”越来越容易，真正稀缺的也许不再只是执行能力，而是选择方向、验证判断和持续负责的注意力。</description><link>https://tinghexin.com/research/judgment-in-ai-native-teams</link><guid isPermaLink="true">https://tinghexin.com/research/judgment-in-ai-native-teams</guid><pubDate>Thu, 24 Sep 2026 00:00:00 GMT</pubDate></item></channel></rss>