返回
想在 Claude、ChatGPT、Codex 等工具里更省 Token,最有效的做法不是讲文言文,而是减少重复上下文、控制输出长度,并把简单任务分流给更便宜的模型。
资源指南
ai工具
skills推荐
token
prompt engineering
复盘一门沟通课程后,我更确定:职场沟通最常见的问题,不是表达技巧不够,而是没有从人、目标与情境出发。
经验谈
沟通
管理
教练对话
向上管理
会议沟通
Most overseas campaign failures are misdiagnosed. The problem is rarely content or channel — it is the information structure inside the building.
案例复盘
b2b营销
前瞻思考
Our TikTok channel broke out by month three — 400 inquiries/month. We still did not convert. Why traffic was never the real problem.
内容营销
Six months, 2–3 people, one niche B2B product. How multilingual content work earned a Google AI Overview result and converted an inquiry from India.
用户增长
How I went from translating psychology videos to running a 50-person volunteer team on Bilibili — and what it taught me about growth.
个人成长
社群运营
No CS degree. Here are 6 internal tools I built with Claude Code — a CRM, scrapers, a cost estimator — and what the process taught me.
vibe coding
工作流
Markdown is a lightweight markup language.
markdown
Markdown 是一种轻量级的「标记语言」。
没有符合条件的文章。