AI写得顺，怎么判断它说得对？AI寫得順，怎麼判斷它說得對？A fluent AI answer: is it correct?
把生成、查证和判断分开，找到最需要自己核对的地方。把生成、查證和判斷分開，找到最需要自己核對的地方。Separate generation, verification and judgment to find what needs checking.
继续阅读 →繼續閱讀 →Read more →KNOWLEDGE & JUDGMENT
把一个问题讲清，
让判断多一层依据。把一個問題講清，
讓判斷多一層依據。Understand the question.
Build a reasoned judgment.
从AI知识、认知与工程行业的具体场景开始。事实、假设和观点分别说明。從AI知識、認知與工程行業的具體場景開始。事實、假設和觀點分別說明。Explore AI knowledge, judgment and engineering contexts, distinguishing facts, assumptions and views.
START READING
五个问题，五个理解入口。五個問題，五個理解入口。Five questions to begin with.
这里的文章是观点与方法说明，不是课程成果、客户实绩或收入证明。這裡的文章是觀點與方法說明，不是課程成果、客戶實績或收入證明。These articles explain views and methods; they are not evidence of course outcomes, client results or income.
工程资料，哪些可以交给AI？工程資料，哪些可以交給AI？Which engineering information belongs in AI?
先辨认敏感信息、处理范围和使用目的，再决定输入什么。先辨認敏感資訊、處理範圍和使用目的，再決定輸入什麼。Consider sensitivity, processing scope and purpose before choosing inputs.
继续阅读 →繼續閱讀 →Read more →演示能跑，为什么还不能直接交付？演示能跑，為什麼還不能直接交付？A demo works. Is it ready to deliver?
看正常结果，也看错误、复核、版本和责任怎样衔接。看正常結果，也看錯誤、複核、版本和責任怎樣銜接。Look beyond a successful run to errors, review, versions and responsibility.
继续阅读 →繼續閱讀 →Read more →工程经验，怎样帮助理解AI？工程經驗，怎樣幫助理解AI？How does engineering experience help with AI?
专业经验带来条件、尺度和复核路径，而不只是更多背景信息。專業經驗帶來條件、尺度和複核路徑，而不只是更多背景資訊。Experience contributes conditions, standards and ways to review an answer.
继续阅读 →繼續閱讀 →Read more →一人公司，先看业务怎样组织。一人公司，先看業務怎樣組織。A one-person business: start with how work is organized.
把工具、责任与客户需求放在同一张图里，理解这一方向的条件。把工具、責任與客戶需求放在同一張圖裡，理解這一方向的條件。Consider tools, responsibility and customer needs together to understand this direction.
继续阅读 →繼續閱讀 →Read more →