Leo Cavalcante

An agent creates an endpoint and gets validation wrong.

Some models impress you with their first answer.

A new Go release usually arrives without making much noise.

The conversation about AI productivity is stuck on a number. 55.8%. 26%. 19%.

When someone asks what the next trend in AI and software development will be, the easiest answer is to point to another model, another editor, or another promise of code generated from a sentence.

An AI agent gets a request such as: “renew the customer contract, but do not exceed the approved budget.”

Whenever we talk about excessive agency in AI agents, the conversation quickly becomes a security conversation.

There is an optimistic reading of the current phase of software development, and part of it is true: more people coding, more repositories, more pull requests, more code reaching production.

If the Agent Harness is the system around the model, the Meta Harness is the system around harnesses.

There is a comfortable way to build an AI agent: take a good model, write a large system prompt, connect half a dozen tools, add a few MCP servers, drop in Markdown skills, inject conversation history, and call it a platform.

Your agent nailed the demo and everyone loved it.

Introduction: the API bill shock For any developer who has integrated autonomous agents into real workflows, the initial excitement around “zero-touch” automation usually lasts until the first API invoice closes.

The pace of software development has undergone unprecedented compression.

In Go, context.Context is the standard mechanism for propagating cancellation, deadlines, and values across Goroutines and function calls.

You have probably experienced this: in one moment, GPT or Claude solves a complex coding problem in seconds; in the next, the same AI forgets basic context or invents nonexistent information.

If you follow the GitHub Copilot ecosystem, you have probably heard of *.agent.md files.

The paradox of modern speed We live in the paradox of technical abundance: in the age of Artificial Intelligence, we generate code in minutes, but organizations have never struggled so much to convert that volume into real value.

The scale paradox There is an efficiency gulf between biological architecture and silicon.

The productivity paradox Engineering leaders and developers are constantly searching for more productivity.

If you are starting to venture into the world of AI Engineering and want to go beyond the basics, you need to deeply understand what Retrieval-Augmented Generation (RAG) is.

You know when you are trying to solve a complex problem with AI and it feels like a single model cannot handle the job?

Exactly one year ago, a commit with the message “Golanged” marked the beginning of a new phase in my career as a developer.

In the world of software development, prompt engineering is emerging as a crucial discipline.

Go 1.25 was officially released in August 2025, bringing a series of significant improvements for developers.

Maps in Go have always been one of the language’s most widely used and optimized data structures.

Go 1.24 introduced a new package called weak, bringing support for weak pointers.

Version 1.24 of the Go language was released in February 2025, bringing a series of improvements and new features that promise to increase developer productivity and optimize application performance.

The Go language was criticized for its simplicity and design philosophy in a recent article called “Why Go Should Sometimes Be a No-Go”, where the author raises some negative points about the language, arguing that its minimalist approach can be a limitat…

Functional programming (FP) has gained popularity in recent years due to its ability to create more expressive, modular code with fewer side effects.

Google’s style, decisions, and best practices guide for the Go language is a set of recommendations intended to promote clarity, simplicity, and efficiency in code development.

The Go Developer Survey is an essential tool for understanding the evolution and challenges faced by developers who use the Go language.

Since its release in 2009, the Go programming language has won over developers around the world, especially in distributed systems, cloud computing, and DevOps.

I think one of my biggest adventures with Go has been working with tests.

Semantic versioning, also called SemVer, is a standard used by developers to number software versions clearly and predictably.

In today’s fast-paced, technology-driven world, the success of software-based organizations depends heavily on the performance of their developers.

In the ongoing search for efficient and concurrent programming languages, the Go language has stood out for its simplicity, performance, and powerful features.

At PicPay, to improve the scalability of our PHP applications, we use Swoole as a high-performance runtime, making applications asynchronous and non-blocking.

Swoole is about to deliver something very, very cool: its own CLI.

Swoole will be shipping something really-really cool that is it’s own CLI.

Originally posted by @Reasno at guxi.me/posts/cloudnative-hyperf Hyperf officially provides container images and the configuration options are very open.

Inspired by the article about Hyperf by the great @Leonardo do Carmo (and also by his tip about Hashnode hehe), I decided to stop by and write about Nano, the Swoole micro-framework from the Hyperf folks.

I’m assuming you already know what is GraphQL and Swoole, so what about getting started right straight to the code, shall we?

When I’m not in front of a “soup of colored letters” (people say this when they see my screen on VS Code) or “jumpy green text” (yeah, this one is for the terminal); I like to play and study music!

I have just a couple years in functional programming, just a couple weeks in F# and no background or experience with .NET, so don’t take this as a reference from an expert, it’s just me documenting what I’ve learned so far.

One the of the hardest things in development - IMHO - after naming, is handling State, at least handling it reactively without overs (over-building, over-rendering, over-painting and overhead!).

It is common for people to argue that hashing is not cryptography; they do it so vehemently that they ridicule things like “I encrypted it with MD5.”

I used to love everything about object oriented programming, ignoring everything else, when I finally got it, I thought that every piece of software should be developed using OO and if it is not, it is wrong.