philschmid：系统提示词是持续成本，指令应尽量精简
DeepMind 的 philschmid 分享了一个很反直觉的经验：他把系统提示词和脚手架删掉约 90%，效果反而更好。核心是按需加载技能，而不是把所有指令常驻上下文。正在做 Agent 和提示词工程的人值得看这场分享。
Google DeepMind 的 @_philschmid 提出把系统提示词视为一种反复产生的成本：每次请求都会重复消耗 token，因此指令应尽量简短。他的做法是把详细说明从系统提示词中移出，改为在需要时再加载对应的技能说明，而不是一开始就全部塞进上下文。
这一思路的核心是“按需加载”：常驻上下文里只保留最少量、最高频的指令，其余内容拆成可延迟载入的部分，从而降低每次调用的开销，也减少长提示词对模型遵循指令能力的干扰。他将在 AIE CODE 大会上分享题为“I deleted 90% of my harness (And It Got Better)”的实践，说明删掉约 90% 的脚手架代码后效果反而更好。
Your system prompt is a recurring cost. @_philschmid of @GoogleDeepMind on keeping instructions short and loading skills when needed. He’s back at AIE CODE tomorrow: “I deleted 90% of my harness (And It Got Better).” ai.engineer/code/schedule
This report is categorized under AITOP's "How-to" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include philschmid, Google DeepMind, 提示词工程, 智能体.
Source attribution: curated from AI Engineer (@aiDotEngineer).
