Aaron Levie's observation reframes AI product development around a fundamental tension: models need rich context to be useful, but excessive context degrades performance. This shifts the primary product design challenge from "what can the model do?" to "what should the model see?" The selection, prioritization, and presentation of context becomes the differentiating factor between AI products built on the same underlying models. Context engineering involves managing three types of information: user history (past interactions and preferences), workflow data (current task context and state), and domain data (relevant documents, code, or information). The challenge is determining which pieces of this potentially massive context window actually improve model output versus which create noise. This requires product designers to deeply understand both user workflows and model behavior—a combination of traditional UX design and emerging AI system design. The constraint of context window limits (even as they expand to millions of tokens) forces prioritization decisions that define the product. A code completion tool might prioritize recent edits and current file structure over full codebase history. A customer service AI might prioritize recent conversation history and specific customer data over general knowledge base articles. These architectural decisions about context selection become the product's competitive advantage. The framing also implies that raw model capability matters less than thoughtful context provisioning. A less capable model with excellent context engineering can outperform a more advanced model with poor context management. This democratizes AI product development—success comes from understanding user workflows and designing context pipelines, not from having access to the best models. ## Key Insight In AI product development, context engineering—deciding what information the model sees and in what order—is the primary product design challenge, more critical than the choice of underlying model. ## Cross-Topic Connections - [[AI-Assisted Development]] — Context selection determines code completion quality and relevance - [[Knowledge Management]] — Parallels to information architecture and retrieval system design - [[Product Design]] — Context engineering as new fundamental product design discipline