Brief

Andrej Karpathy proposed a shift from "prompt engineering" to "context engineering," arguing that the former term inaccurately represents the complex task of building production AI applications. He clarified that real-world AI systems rely on dynamically assembling a context window from various sources like system instructions, retrieved documents, and conversation history, rather than just simple user prompts. This assembly process, which Karpathy deems the true engineering challenge, is crucial for optimal model performance and is distinct from merely crafting task descriptions. AI

LLM-based reasoning agents are improving their ability to interpret ambiguous human language, making software specifications a more dynamic source of truth. However, the stochastic nature of LLMs necessitates constraints, leading to the rise of Context Engineering. This discipline focuses on providing clear intent and instructions to AI models through structured artifacts like skills, rules, scripts, feedback loops, and evaluation metrics. AI

A software developer expresses concern about AI agents and context engineering potentially diminishing coding skills, likening it to telling a painter to only describe their work. Despite this, the developer acknowledges the positive impact of AI on productivity and the democratization of coding, noting their own increased output over the past two years. AI