GPT-5.5 prompting guide
Simon Willison's Weblog299 字 (约 2 分钟)
90
OpenAI 发布了 GPT-5.5 的提示工程指南，强调新模型需重新调优而非直接复用旧提示。
入选理由：GPT-5.5 需要从零开始优化提示，而非简单复用旧版本的配置。
FeaturedArticle#GPT-5.5#OpenAI#Prompt Engineering#AI英文
traeai topic radar
Prompt Engineering、提示词模式与上下文工程
追踪 prompt engineering、context engineering、system prompt、few-shot、结构化输出、工具调用提示与提示词评测。
What searchers are trying to solve
想找到可复用的提示词模式，理解提示词和上下文组织如何影响模型输出。
Why this is worth tracking
即使模型能力提升，提示词和上下文仍是把模型变成可靠产品能力的关键设计层。
Prompt Engineering提示词工程context engineeringsystem promptfew-shot结构化输出prompt eval
长尾组合
这个主题可以沿着工具、实践、对比等搜索意图持续扩展，不靠空壳换词，而是用真实材料更新。
Prompt Engineering 工具Prompt Engineering 实践Prompt Engineering 对比提示词工程 工具提示词工程 实践提示词工程 对比context engineering 工具context engineering 实践
可自动化内容模块
精选材料
持续抓取与 Prompt Engineering 相关的高分文章、播客、视频和推文。
趋势判断
把最近变化、反复出现的观点和争议点整理成稳定摘要。
实体关联
自动连接相关公司、模型、产品、人物和概念，形成可继续深挖的入口。
Featured content
Filtered by relevance, score, and recency.
#543. Why 2026 is the Year of Harness? Deep Dive by IBM Expert
跨国串门儿计划1189 字 (约 5 分钟)
88
2026 will be the year of AI Harness. Using engineering methods like guardrails, validation, and automation processors, unreliable AI Agents can be transformed into stable, controllable systems without modifying Prompts, marking key infrastructure for AGI.
入选理由：AI Harness consists of five core components: tool registration, context compress
FeaturedPodcast#AI Agent#Harness#IBM#Prompt Engineering#RAG中文
Prompt Engineering for Agentic AI
Machine Learning Mastery4533 字 (约 19 分钟)
87
In agentic AI systems, prompt engineering has evolved into context engineering: reliable behavior requires deliberate design of four components—system prompt, tools, examples, and context state management—otherwise, context rot causes behavioral drift across multi-step reasoning, as identified by Anthropic's research.
入选理由：Agentic prompts require four components: system prompt, tool definitions, exampl
FeaturedArticle#Agentic AI#Prompt Engineering#Context Engineering#Anthropic#LLM英文
🔬Doing Vibe Physics — Alex Lupsasca, OpenAI
Latent Space1542 字 (约 7 分钟)
87
Alex Lupsasca of OpenAI demonstrates that GPT-5 series models have achieved breakthrough scientific reasoning, reproducing his months-long theoretical physics paper in just 11 minutes — signaling AI’s transformation of fundamental scientific discovery.
入选理由：GPT-5 can reproduce a theoretical physicist’s months-long paper in 11 minutes —
FeaturedArticle#AI for Science#GPT-5#Theoretical Physics#Prompt Engineering#OpenAI英文
Changes in the system prompt between Claude Opus 4.6 and 4.7
Simon Willison's Weblog1305 字 (约 6 分钟)
87
Anthropic 在 Claude Opus 4.7 系统提示中新增儿童安全标签、工具搜索机制，并优化交互逻辑以减少冗长和侵扰性行为。
入选理由：新增 <critical_child_safety_instructions> 标签，强化儿童安全策略
FeaturedArticle#Claude#AI系统提示#大模型#Anthropic#Agent设计英文
The End of Prompting: Why the Future of AI Experience Design Is Constraint-First
UX Magazine1916 字 (约 8 分钟)
87
AI体验设计正从提示工程转向约束优先架构，以解决大模型幻觉与可信度问题。
入选理由：提示仅能引导风格，无法保证事实准确性或防止幻觉。
FeaturedArticle#AI体验设计#大语言模型#人机交互#约束优先#Prompt Engineering英文
Pydantic + OpenAI: The Cleanest Way to Get Structured Outputs from LLMs
Towards Data Science2355 字 (约 10 分钟)
85
Pydantic与OpenAI结构化输出结合能高效处理LLM输出，提升Python应用可靠性。
入选理由：Pydantic通过类型注解自动验证数据结构，减少手动解析JSON的错误
FeaturedArticle#Pydantic#OpenAI#LLM#Python#数据验证英文