半小时入门The Half-Hour Primer
约 30 分钟about 30 minutes
先读十个常用概念，再做三道自测。目标是分清模型、产品与能力边界，不需要记住全部名词。Read ten common concepts, then try three questions. Aim to distinguish models, products and limits without memorizing the whole glossary.
- 向外行解释：大模型是怎么「学会」的，为什么按 Token 计费
- 说清「AI 产品」和「AI 模型」的区别，各举一个例子
- 说出大模型至少一件明确做不到的事
- Explain to a newcomer how a large model learned, and why billing counts tokens
- Tell apart an AI product from an AI model, with one example each
- Name at least one thing a large model plainly cannot do
先试着回答，再展开参考答案。答不上来的地方，可从上方步骤返回对应词条。参考回答是核对要点，不要求逐字复述。Try answering before opening the reference. Revisit the relevant step above if needed. These answers highlight key ideas; you do not need to repeat the wording.
第 1 题 · 参考回答Question 1 · Reference answer
训练用样本调整模型参数，推理时依据输入生成输出。Token 是处理文本的单位；服务商通常分别统计输入与输出的用量，具体收费规则看产品说明。Training adjusts model parameters from examples; inference generates output from an input. Tokens are text-processing units. Providers commonly count input and output usage separately; the product defines billing.
第 2 题 · 参考回答Question 2 · Reference answer
模型是生成或处理内容的核心；产品还包括界面、检索、存储和权限。例如，一个聊天网站是产品，它接入的语言模型是其中一个部件。A model generates or processes content. A product also has an interface, retrieval, storage and permissions. A chat website is a product; its language model is one component.
第 3 题 · 参考回答Question 3 · Reference answer
没有可靠资料或相应工具时，模型不能保证给出最新事实，也不能仅凭流畅的回答证明内容正确。可举一个需要核对原始出处的例子。Without reliable sources or suitable tools, a model cannot guarantee current facts. Fluent wording does not prove correctness. Give one example that needs an original-source check.
用在一个具体任务里Apply this to a concrete task
试着写一份有资料、有边界的提示Try writing a source-grounded, scoped prompt →
一个下午看懂原理One Afternoon Under the Hood · 一周路线The Full Week · 内容工作者路线The Writer's Route · 动手开发者路线The Builder's Route · 选型决策者路线The Decision-Maker's Route · 研究学习者路线The Researcher's Route · 运维上线路线The Operator's Route · 普通消费者路线The Everyday User's Route · 图像 / 视频 / 音频创作者路线The Image / Video / Audio Creator's Route