Écrits
Des textes courts sur les agents en terminal, le choix des modèles et la maîtrise des dépenses en tokens.
- 从 JSON 示例生成 TypeScript 类型：分步实战指南
- 如何为 Python 代码生成单元测试：从行为定义到回归验证
- 如何用语言模型起草技术文档：从资料整理到发布检查
- 如何可靠地总结会议记录：从原始笔记到可核查纪要
- 真正有用的少样本示例：从选择到验证的六步实践
- 让系统提示词保持简短：从整理规则到回归测试
- 一个下午，为现有产品加上第一个 AI 功能
- 从厂商 SDK 迁移到兼容端点：分步操作指南
- 为每个环境配置独立 API 密钥：分步操作指南
- Embeddings Endpoints: When and How to Use Them
- 逐步读懂 API 响应中的用量字段
- Logging Model Requests Responsibly, Step by Step
- How to Rotate API Keys Without Downtime
- Keeping API Keys Safe Inside a Team
- Concurrency Limits and Request Queues: A Walkthrough
- Timeouts for Long Streaming Requests: A Walkthrough
- How to Call a Language Model API with curl
- Setting baseURL in the OpenAI Node SDK
- Counting Tokens Before You Send: A Practical Guide
- Function Calling Basics: A Step-by-Step Guide
- Retries With Backoff That Never Double-Charge
- What HTTP 402 Means on a Prepaid API
- What a 401 Error Is Telling You
- How to Handle 429 Rate Limit Responses Gracefully
- Gemini generateContent Explained, Step by Step
- Server-Sent Events: Stream Responses Step by Step
- Anthropic Messages API: A Beginner’s Walkthrough
- Prepaid Credit, Balances and Refunds Explained
- When a Bigger Model Is Actually Cheaper
- Cheap Drafts First, Expensive Polish Last
- Cost per Merged Pull Request: A Practical Guide
- Budget Planning for a Small Team Using AI
- The Hidden Cost of Very Long AI Conversations
- Streaming and Cost: What Changes and What Does Not
- Batching API Requests to Cut Overhead
- A Spending Ceiling for Every Project Key
- Quota Passes: The Official-Price Ratio, Step by Step
- Stop Runaway API Loops Before They Run Up a Bill
- Cheaper Models for Background Jobs: A Walkthrough
- Trim Prompts Without Losing Quality: A Practical Guide
- Keep Your Coding Agent Within a Daily Budget
- Claude Cache Reads vs Writes: A Practical Walkthrough
- Pay-as-You-Go vs. a Monthly Pass: A Practical Guide
- When a Quota Pass Makes Sense
- Estimate Monthly API Spend from One Week of Usage
- 提示词缓存如何降低账单：从请求布局到成本验证
- 切换模型之前，先完成一次可回退的测试
- 如何读懂按令牌计费的价格表：逐步核算输入、输出与缓存成本
- 在同一产品中混用 Claude 和 GPT：从接入到任务路由
- 推理令牌是什么：从用量、费用到参数设置的分步指南
- 模型别名与模型标识为什么会变：六步管理调用与迁移
- 翻译任务该用哪个模型：从样本测试到稳定交付的六步指南
- 上下文窗口有多大，究竟能帮你做什么
- Opus and Sonnet on the Same Refactor
- 如何为代码审查评论选择模型：从样本评测到上线
- 小模型还是大模型：分类任务的逐步选型指南
- One Model per Task: A Practical Routing Walkthrough
- MiniMax M3 for Lightweight Assistants: A Walkthrough
- Qwen Models for Multilingual Products: A Walkthrough
- Grok 4.7 for Open-Ended Questions: A Practical Guide