Blog
Blog and studies
Records about applied AI, web development, tools, behind-the-scenes work, and technical decisions that appear during practice.
blog
How to use an AI agent to reason about frontend performance
Learn how to guide an AI agent to measure and improve React or Vue frontend performance with SSR, caching, lazy loading, and before-and-after validation.
blog
Practical guide to caching AI responses and searches
Understand prompt caching, exact-match caching, semantic caching, and caching in RAG to reduce cost and latency without reusing incorrect or stale answers.
blog
Is Claude Academy free? How to learn AI with better judgment
Claude Academy offers free courses from AI fundamentals to Claude Code, the API, and MCP. Learn how it works, what its badges mean, and why the four Ds are its most useful idea.
blog
If GitHub goes down, does your application stay up?
The GitHub incident shows that an application can stay online while losing the ability to deploy, scale, or roll back safely.
blog
Everything Claude Code: the story of ECC and what is worth learning from it
Learn the story behind Everything Claude Code and how ECC organizes skills, agents, rules, hooks, memory, and security for coding agents.
blog
When AI tries to cheat on the test: OpenAI, GPT-5.6 Sol, and Hugging Face
Understand the incident where OpenAI models searched for benchmark answers on Hugging Face, and what it teaches about agents, sandboxes, permissions, and evaluation.
blog
Long-horizon agents: why approving one action at a time is not enough
Why agents with shell, browser, and files need trajectory monitoring, checkpoints, and rollback, not only command-by-command approval.
blog
Practical guide to security and sensitive data in LLM applications
Learn practical safeguards for handling sensitive data in LLM apps: prompts, logs, RAG, tools, permissions, human review, and context limits.
blog
GPT-5.6 in ChatGPT: judgment before hype
A practical read on GPT-5.6 in ChatGPT, what to check before switching models, and how to choose between speed, reasoning, cost, and human review.
blog
Practical guide to observability, cost, and fallback in AI apps
Learn which logs, metrics, cost limits, retries, and fallbacks help operate AI features beyond manual testing.
blog
A practical guide to OpenClaw, Hermes Agent, and personal agents
Understand what OpenClaw and Hermes Agent are, which benefits AI personal agents can bring, and which technical safeguards should come before autonomy.
blog
A practical guide to agents with tools
Learn when agents with tools make sense in AI apps and which limits, permissions, and validations help use this pattern with better judgment.
blog
A practical guide to harnesses for testing AI prompts, models, and workflows
Learn how to build a simple harness to test AI prompts, models, and workflows with cases, repeatable execution, logs, and version comparison.
blog
A practical guide to evals for AI features
Learn how to use evals to assess AI prompts, models, and workflows with simple criteria, test cases, version comparisons, and regression checks.
blog
A practical guide to structured LLM outputs
Learn how to use structured LLM outputs to generate more predictable data, validate responses, handle parsing errors, and integrate AI into applications.
blog
A practical guide to chunking, embeddings, and retrieval
Understand how chunking, embeddings, and retrieval affect the quality of RAG applications and why the problem is not always the model.
blog
A practical RAG guide for developers
Understand RAG in practical terms: when to use it, when to avoid it, how the basic flow works, and which common mistakes hurt LLM applications.
blog
Tools I have been testing to study and build faster
Practical criteria for testing tools for studying, AI-assisted development, documentation, prototyping, and validation without turning every new app into part of the workflow.
blog
Trump reversed the Claude block: what changes for developers?
The reversal around Claude Fable 5 and Mythos 5 shows why developers need to consider access, regulation, fallbacks, and vendor dependency when using AI.
blog
What I learned from organizing an AI First project
Practical lessons from organizing a project around context, documentation, lightweight specs, AI agents, and technical validation.
blog
From requirement to pull request: my flow for building Full Stack features with Codex and Claude
A practical workflow for turning requirements into specs, guiding coding agents, reviewing implementations, and reaching a pull request with stronger technical judgment.
blog
How to integrate AI into a Full Stack application without building an improvised chatbot
How to structure an AI feature considering backend, frontend, queue, worker, validation, security, costs, and user experience.
blog
How to design a multi-LLM architecture with Claude, Gemini, and OpenAI
How to create a multi-LLM architecture with an AI Gateway, adapters, routing, fallback, cost control, and observability without over-abstracting.
blog
How to turn a rough idea into a navigable prototype
A practical process for turning a rough idea into a clear problem, focused scope, main flow, and navigable prototype with thoughtful use of AI.
blog
Trump, Claude 5, and AI restrictions: what changes for developers
The Fable 5 and Mythos 5 case shows how regulation, access, and vendor dependency can affect AI development workflows.
blog
Claude Fable 5, Mythos 5, and Opus 4.8: what actually changes?
Understand the differences between Claude Fable 5, Mythos 5, and Opus 4.8 across capabilities, access, safeguards, costs, and practical use for developers.
blog
What vibe coding means when technical responsibility still matters
Vibe coding can help developers explore ideas and create prototypes faster, but fast code generation only becomes useful when it goes through context, review, validation, and technical responsibility.
blog
Behind the scenes of building brunopaim.tech
A behind-the-scenes look at how I organized brunopaim.tech with Nuxt, Tailwind, a blog, AI First documentation, and criteria for evolving the project.
blog
Claude Opus 4.8: what dynamic workflows actually change
A practical look at Claude Opus 4.8, dynamic workflows in Claude Code, and what they may change for real AI-assisted development work.
blog
A simple checklist for reviewing AI-generated code
A practical checklist for reviewing AI-generated code with context, technical judgment, tests, maintainability, and validation before using it in a project.
blog
How I use documentation to work better with AI
A practical reflection on how context, specs, constraints, and criteria make AI more useful in development.
blog
How I am using AI without giving up technical judgment
A practical reflection on using AI in development with context, review, validation, and technical responsibility.
blog
Why I created brunopaim.tech
The idea behind having my own space to organize studies, links, and projects.
blog
How I am using AI to study and create projects
Notes on the practical use of AI for learning, creating interfaces, and organizing ideas.
blog
Tools I want to test in my day-to-day as a developer
An initial list of digital tools to try in real development flows.
blog
What is vibe coding in practice?
An objective view of creating with AI support without leaving review, tests, and technical intent behind.