Cláudio Araújo https://claudiodearaujo.dev.br/pt Engenharia de software, arquitetura, inteligência artificial e liderança técnica. pt-BR Why Keep Incremental Hydration After It Missed Its Target? https://claudiodearaujo.dev.br/pt/engineering/decisions/measuring-incremental-hydration https://claudiodearaujo.dev.br/pt/engineering/decisions/measuring-incremental-hydration Sat, 19 Sep 2026 00:00:00 GMT Cortar 65% do trabalho de hydration comprou 6% de tempo. A distância entre os dois números era o resultado. Architecture Engineering Evidence Why This Site Intercepts Its Own Fragment Links? https://claudiodearaujo.dev.br/pt/engineering/decisions/intercepting-fragment-links https://claudiodearaujo.dev.br/pt/engineering/decisions/intercepting-fragment-links Fri, 18 Sep 2026 00:00:00 GMT Uma tag que o framework exige quebrou o link mais importante da página para quem navega por teclado. Architecture Engineering Accessibility Why Human Control Belongs in the Architecture? https://claudiodearaujo.dev.br/pt/engineering/decisions/human-control-in-architecture https://claudiodearaujo.dev.br/pt/engineering/decisions/human-control-in-architecture Thu, 10 Sep 2026 00:00:00 GMT Participação humana deve ser um boundary explícito do sistema. Governance Human-in-the-loop Architecture Evidence Before Autonomy https://claudiodearaujo.dev.br/pt/writing/evidence-before-autonomy https://claudiodearaujo.dev.br/pt/writing/evidence-before-autonomy Tue, 08 Sep 2026 00:00:00 GMT Por que sistemas inteligentes deveriam conquistar autoridade através de evidência. AI Evidence Autonomy Evaluating Agentic Systems https://claudiodearaujo.dev.br/pt/labs/evaluating-agentic-systems https://claudiodearaujo.dev.br/pt/labs/evaluating-agentic-systems Sat, 05 Sep 2026 00:00:00 GMT Como avaliar comportamento de agentes completos além da resposta final. AI Agents Evaluation From Automation to Autonomy https://claudiodearaujo.dev.br/pt/writing/from-automation-to-autonomy https://claudiodearaujo.dev.br/pt/writing/from-automation-to-autonomy Tue, 01 Sep 2026 00:00:00 GMT Automatizar uma tarefa e delegar uma decisão são problemas diferentes. AI Autonomy Governance Why Shadow Mode Before Autonomy? https://claudiodearaujo.dev.br/pt/engineering/decisions/shadow-mode-before-autonomy https://claudiodearaujo.dev.br/pt/engineering/decisions/shadow-mode-before-autonomy Fri, 28 Aug 2026 00:00:00 GMT Como construir evidência prospectiva antes de permitir autonomia operacional. Autonomy Evidence Research AI Agents Need Architecture, Not Just Prompts https://claudiodearaujo.dev.br/pt/writing/ai-agents-need-architecture https://claudiodearaujo.dev.br/pt/writing/ai-agents-need-architecture Tue, 25 Aug 2026 00:00:00 GMT Por que agentes confiáveis exigem muito mais do que bons prompts. AI Agents Architecture Long-Term Memory for Agents https://claudiodearaujo.dev.br/pt/labs/long-term-memory-for-agents https://claudiodearaujo.dev.br/pt/labs/long-term-memory-for-agents Thu, 20 Aug 2026 00:00:00 GMT Como construir memória útil sem transformar contexto acumulado em ruído. AI Agents Memory Why MCP-first? https://claudiodearaujo.dev.br/pt/engineering/decisions/why-mcp-first https://claudiodearaujo.dev.br/pt/engineering/decisions/why-mcp-first Sat, 15 Aug 2026 00:00:00 GMT Agentes devem depender de capacidades, não de implementações específicas. Architecture MCP Agents