Rogerio Rodrigues da Silva

E a polêmica da vez é Linus e o vibe coding. Em 2008 no primeiro capítulo do clean code , Uncle Bob já escrevia o subtítulo “There will be code”. “…We will never be rid of code, because code represents the details of the requirements. At some level those details cannot be ignored or abstracted; they have to be specified. And specifying requirements in such detail that a machine can execute them is programming. Such a specification is code.”

13

Projeto Novo em Golang nascendo https://lnkd.in/dHQvsqZT Go-Rundeck is a web-based runbook automation and task orchestration platform inspired by Rundeck, built natively in Go. The goal is to provide a lightweight, highly performant, and modern alternative to the original Rundeck (Java), maintaining a familiar layout and workflow but with a much leaner and more efficient tech stack.

7

🚨 The Inside Job That Shook Brazil's Financial System 🇧🇷💥 One of the most devastating cyberattacks in Brazil’s history hit the Brazil's Payment System (SPB) and it wasn’t just technical. It was INSIDER-LED. A trusted employee at C&M Software, a key provider in the banking infrastructure, sold access to attackers. With stolen credentials and digital certificates, they executed valid transactions that drained hundreds of millions — instantly and silently. These fraudulent transfers were cryptographically sound and cleared through the Central Bank as if nothing was wrong. This wasn’t just an attack on one company. It was a systemic shock to the financial backbone of Brazil and at least 6 institutions were compromised. Funds were laundered through mule accounts and converted to Bitcoin/USDT gone in seconds. Learnings: 🔸 Lack of behavioral monitoring 🔸 Poor privilege control 🔸 Insecure certificate governance 🔸 No just-in-time access 🔸 No isolation between client secrets In 2025, Identity and Access are not just security concerns they’re the first attack vector. Dive deep into how Identity Threat Labs by Segura investigated it, how this happened and what must change NOW: 👉 https://lnkd.in/df6vtTqw 💥 Let’s start a conversation: What’s your biggest concern when it comes to insider threats? Drop your thoughts below 👇 #segura #identity #identitythreatlabs #Cybersecurity #InsiderThreats #SPB #RedTeam #CISO #Governance #PSTI #DigitalTrust #FintechSecurity #ZeroTrust #CryptoLaundering #BrazilCyber #SPI #PrivilegeAbuse #SecurityByDesign #IdentityIsTheNewPerimeter #SeguraSecurity #ThreatIntelligence #CyberAttack2025 #CodeRed

43

10 comentários

Tip: Want to study troubleshooting in Linux environments? SadServer helps with that! It creates an environment with a "sad" server 😢 (with some error) and you need to fix it. For those who like this type of challenge, it's great, I find it very fun (should I be worried?) But that's not all, there are options to elevate privileges, do things from scratch, etc.

2

LM Gateway Design: Rate Limiting, Caching e Fallback para Múltiplos Providers Chamar OpenAI diretamente acopla seu app a um vendor. LLM Gateway abstrai providers (OpenAI, Anthropic, local models), adiciona caching, rate limiting e fallback. This article covers essential points every developer should know about this topic. Key highlights: Architecture and design patterns applied Real-world implementation trade-offs Production best practices When to use and when to avoid If you work with modern software engineering, it's worth reading. Read the full article, links on comments. #SoftwareEngineering #Development #Architecture #BestPractices

6

1 comentário

In software development, documentation should evolve with your code—not trail behind it. Discover how to integrate this practice into your daily routines, the benefits of collaborative and AI-supported approaches, and more: https://lnkd.in/ddD-wVnj ​ ​ ​ Article by Compasser Rodrigo Cezario ​

7

Speculation Rules API, Optimistic UI, Streaming SSR, Priority Hints, Virtual Scrolling, Web Workers, SWR e INP — o que Google, Shopify, Netflix, Meta e Figma fazem de diferente pra ter apps que parecem instantâneos. https://lnkd.in/dSJeziDa

9

1 comentário

💡 Workflow with tasks in GLPI 🤔 Have you ever thought about having a work instruction that guides analysts through every step they need to take, ensuring they follow the correct procedure for the requested service? 🤔What if, when a ticket is escalated or transferred, the instructions automatically update? As soon as a new user is assigned, a fresh set of guidelines appears based on the service and the new group. 😁 Well, this is entirely possible without any external integrations, custom coding, or database tweaks—using nothing but task templates and business rules. If you don’t believe it, check out the video below. #GLPI #ITSM #ESM #TaskWorkFlow

12

SE INSCREVA NO CANAL CEOTVPODCAST NO YOUTUBE The complexity of Brazil's tax system, with its overlapping federal (IPI), state (ICMS), and municipal (ISS) taxes, creates significant conflicts and challenges for businesses. Historically, outdated definitions, like those for software downloads, were only resolved decades later, long after technology evolved. This prolonged uncertainty burdened taxpayers. The recent tax reform, by consolidating multiple taxes into a single one (IVA), represents a crucial step towards simplification and clarity. #ReformaTributária #Brasil #Tributos #Negócios #IVA

1

AI that just writes code isn’t enough. Planning + Execution + Review is where the magic happens. A perfect example from Israel Evangelista showing how Traycer helps teams move fast without breaking things.

23

Restarting my journey as a developer in a new environment taught me something simple but powerful: Great products don’t come from having thousands of ideas, they come from listening. Having direct access to users changed the way I build. Instead of guessing, I focus on constant feedback, working closely with customers, and refining each feature with them. The real value isn’t in how much you build, but in how well you understand the people using it.

11

We've been building a tool called spec-sync at CorvidLabs and figured it was worth sharing. The problem: we have a codebase with 200+ modules where AI agents write a lot of the code. Specs (structured markdown files) are our source of truth for what each module should do. But keeping those specs accurate as the code evolves? That was breaking down fast. So we built spec-sync together. It's a Rust CLI that checks your code against your spec files — do the exports match what's documented? Are the file paths real? Are the dependencies valid? If something drifts, the build fails. We run it in CI on every PR. It's simple, and it catches the thing that always gets you — code that changed but nobody updated the docs. The part that actually matters for us: when an AI agent reads a spec to understand a module, the spec is guaranteed to be accurate. No hallucinated interfaces, no stale docs leading to wrong assumptions. It's open source. Still early and mostly built for our own workflow, but if you're dealing with similar problems — especially with AI agents writing code against documentation — it might be worth a look. https://lnkd.in/gGPRKUU2

4

I used AI to help me build a Mobile Test Automation Framework Lately, I’ve been really curious about how AI can actually support us in real QA workflows and not just for small tasks, but in building something E2E. So I decided to test it in practice! I built a Mobile Test Automation Framework using Appium + WebdriverIO, and throughout the process, I intentionally used AI as part of my development workflow and it made a big difference. ⚙️ Tech stack • Appium • WebdriverIO • Node.js • Android Emulator • Jenkins (CI/CD integration) • Allure Report 💡 Where AI actually helped • Structuring the framework from scratch • Speeding up test development • Troubleshooting environment/setup issues • Exploring better approaches during implementation 🧠 AI tools I used • AI assistant integrated with Visual Studio Code • ChatGPT • Cursor AI 🚀 What the project includes ✔ Automated Android UI tests ✔ Integration with Jenkins for pipeline execution ✔ Detailed reports with Allure (logs + screenshots) ✔ A scalable and organized test structure 📊 With CI + reporting + AI working together, the result is: • Faster feedback • Easier debugging • A more efficient development flow AI doesn’t replace the engineer, but it definitely helps you move faster and think better. 🔗 GitHub repository: https://lnkd.in/d2sSMTa4 Still improving this project and exploring new ways to integrate AI into QA. Please, feel free to contribute to the project if you’d like :)

17

Aqui vai a versão em inglês com hashtags fortes e estratégicas 👇 ⸻ Architecture is not a detail. It’s maturity. Understanding the difference between DTO and ViewModel separates developers who just “make it work” from those who build scalable systems. 🔹 A DTO is not a disguised domain model. 🔹 A ViewModel is not a reused DTO. 🔹 Every layer must have a clear responsibility. When you mix them, coupling begins. And technical debt follows. Professional software demands separation of concerns, well-defined contracts, and respect for architecture boundaries. Real engineers think in layers. Ramon Fullstack 🚀 #SoftwareArchitecture #CleanCode #DotNet #BackendDevelopment #SystemDesign

6

1 comentário

Half a century ago, I wrote my first program in COBOL. It was the beginning of my computing journey, and I never imagined that, in 2025, this language created in the 1950s would still be the backbone of the digital economy. While Python, JavaScript, and Go dominate the headlines, COBOL continues to process trillions of dollars a day. Quietly, with the same reliability as ever. In the 1950s, programming was chaotic. Each manufacturer had its own language, and changing machines required rewriting everything from scratch. That's when the US Department of Defense convened industry and academia to create a common, readable, and portable language. COBOL (Common Business Oriented Language) was born, with the brilliant Grace Hopper championing the idea that code should look like English, not math. The result? A language any business analyst could read, which quickly became the universal language of corporate mainframes. From the 1960s to the 1980s, COBOL reigned supreme in banks, governments, and large corporations. At its peak, 80% of the world's business systems were written in it. Today, in the midst of the AI ​​era, the numbers are still impressive, with estimates indicating that between 220 and 850 billion lines of COBOL remain active. $3 trillion is processed daily using COBOL code, and 95% of ATMs and 80% of card transactions depend on it. In 2022, in the US alone, banks spent $36.7 billion maintaining legacy systems, a figure expected to exceed $57 billion by 2028! What explains this longevity? Several factors, such as stability and performance: COBOL systems are robust and virtually immune to failure. The cost of replacement, since rewriting millions of lines is risky and extremely expensive, is high. And the embedded knowledge, when much of the business rules reside solely in these systems. Far from being a museum piece, COBOL has evolved. Recent versions support object orientation, JSON, XML, and REST APIs. Today, it runs on hybrid clouds, communicates with Java, Python, and C#, and can be accessed via microservices. The trend isn't to erase COBOL, but to encapsulate, refactor, and integrate it, preserving decades of business logic with new digital layers. COBOL doesn't grace the stages of tech conferences, but it holds the stage. And the fact that COBOL is still here, 65 years later, is perhaps the greatest testament to software engineering the world has ever seen.

1

Cybersecurity it’s the first in your systems of data center, protección infrastructure, security,politics, rules, and technology advanced administration, vision in any way you can stay vulnerable. All time stay safe data move passwords and systems inside in your network

2

Texto excelente com o momento atual e quando havia muitos "runtime" de containers como Mesos, Docker, RKT, etc. Vale uma leitura. "From Docker to LangGraph: Why the agent framework shakeout is the container wars all over again. Learn to bet on protocols, not frameworks, in this new era." https://lnkd.in/dSdr6ytE

18

📝 Novo artigo do nosso Founder no AWS Builder Center! 🤖 Automating AWS Community Onboarding with Lambda, GitHub Actions, and Cognito Como automatizamos o onboarding de membros em menos de 60 segundos — custo praticamente zero. Toda a infra da comunidade roda por ~$8/mês. Do formulário ao card no site: Lambda cria PR, GitHub Actions deploya, Cognito libera acesso. Tudo serverless, tudo automatizado. 📝 https://lnkd.in/e7fdrSKM Ricardo Gulias Founder & Global Lead — Golden Jackets #AWS #Serverless #Lambda #GitHubActions #Cognito #GoldenJacketsBrazil #AWSCommunity #BuildOnAWS ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━ 📝 New article from our Founder on AWS Builder Center! 🤖 Automating AWS Community Onboarding with Lambda, GitHub Actions, and Cognito How we automated member onboarding in under 60 seconds — at virtually zero cost. The entire community infra runs for ~$8/month. From form submission to live member card: Lambda creates PR, GitHub Actions deploys, Cognito grants access. Fully serverless, fully automated. 📝 https://lnkd.in/e7fdrSKM Ricardo Gulias Founder & Global Lead — Golden Jackets #AWS #Serverless #Lambda #GitHubActions #Cognito #GoldenJacketsBrazil #AWSCommunity #BuildOnAWS

12