Diogo de Oliveira Biazus

Pull Requests shouldn’t feel like this. We’ve all opened a large PR review and thought: “Well... this is going to hurt.” Sure, small PRs are great, but sometimes you need the full picture. That’s where stacked PRs help (been using Graphite for that, a tool for managing PRs and code reviews). Stacked PRs treat your feature development as a chain of commits / PRs layered on top of each other. Think of it like a Jenga tower: each block (PR) builds on top of the previous one. pr1 → main pr2 → pr1 pr3 → pr2 Now it’s getting even better. I just got early access to Graphite Chat 👇 - I can literally ask it “what’s going on in this PR?” and it gives me the full story, codebase history, past PRs, even outside context. - I can just highlight it and ask “explain this”, instant breakdown. - Flags risks before reviewers even show up. - When someone leaves feedback, it helps me make sense of it and suggests fixes. - And the best part? When CI blows up, I can just ask “why did this break?” and actually get an answer. I’ll even be able to pop open the Graphite Editor right inside Chat to view and edit the codebase. Reviews feel lighter, more collaborative. Exactly how they should.

1,570

141 Comments

I can’t stress this enough: To become a Senior engineer you don’t need to complicate things. What you need is to focus your efforts on being pragmatic, business-oriented, and solution-oriented. I’ve seen too many talented mid-level devs shooting themselves in the foot by: 👉🏻Adding unnecessary abstraction layers “for future flexibility” 🤦‍♂️ 👉🏻Implementing design patterns that solve problems that don’t exist 👉🏻Building “elegant” solutions that take 3x longer to ship. Meanwhile, senior devs are out there asking the right questions: 👉🏻”Whats the simplest solution that works?” 👉🏻”What are we actually optimizing for?” 👉🏻”Will this help the business or just my ego?” Your promotion isn’t blocked by lack of complexity, actually it’s blocked by too much of it. Save your creative energy for problems that actually need complex solutions.

677

46 Comments

A pattern I see emerging is a kind of outside-in development cycle. You start from bare requirements, iterate quickly to a working model, where the internal structure is cared about primarily to keep the agent on track (and so that you can comunciate effectively about what is happening - this is DDD at its most real, honestly.) Then you start diving in to the details of the implementation, and doing things like security, structure, etc in successive waves of the plan/execute/review loop.

15

6 Comments

Unpopular opinion (apparently): .NET devs who genuinely value OSS, simplicity, and minimal corporate nonsense are a minority in the .NET ecosystem. I say this as someone who’s been around since 2004. I understand the legacy. I lived through it. I even survived it. But at some point… can we move on? Too much modern .NET code still feels like it was designed in 2009, just with newer keywords and a fancier DI container. Endless abstractions, framework worship, ceremony over clarity. More patterns than problems. What honestly makes me sad is this: if you want sane defaults, low-BS development, and tools that don’t assume an enterprise committee per line of code, the most realistic option often seems to be… switching ecosystems entirely. Go. Rust. Pick your flavor of “less nonsense.” Not because .NET can’t do it. But because too many .NET devs won’t let it. The platform evolved. The language evolved. The runtime evolved. Some of the mindset didn’t. And no, adding one more layer, interface, or “enterprise-grade” framework is not progress. This is my last shout-out of the year. Next one will be next year, something more creative, less ranty… probably still inconvenient for someone. 🔥🎅

178

58 Comments

James Montemagno didn’t plan to build a full analytics platform in 5 minutes. He just wanted to explore podcast data from Merge Conflict, 492 episodes deep, trapped in a CSV file. So he opened VS Code, fired up GitHub Copilot with Claude Sonnet 4.5, and started vibe coding. No specs. No frameworks to debate. Just vibes and a clear outcome: 💬 “Build me a site that visualizes podcast metrics beautifully.” Five minutes later, the first version was live. Fifteen minutes later, it was deployed to GitHub Pages. That’s the power of trusting AI to handle the how so you can focus on the what. Copilot (with Sonnet 4.5 reasoning) made every architectural call: 1️⃣ React + TypeScript 2️⃣ Tailwind CSS 3️⃣ Recharts 4️⃣ Automated GitHub Actions deployment The result? A polished, data-driven dashboard that James didn’t “code” in the traditional sense... he collaborated it into existence. He calls it flow-state development. No docs. No overthinking. Just outcome-driven creativity. If you’ve ever wondered what it feels like to have an AI co-architect your app, this post is your blueprint. 👉 Read the full breakdown here: https://lnkd.in/gbpWgyVv #Developers #AI #GitHubCopilot #VibeCoding #VSCode

34

1 Comment

ES2025 is official! The 16th edition of ECMAScript (the standard that defines JavaScript) has been officially approved by Ecma International. This update brings a host of powerful new features: 0️⃣ Iterator Helpers: We can now call methods like .map(), .filter(), .take(), and .toArray() directly on any iterable (not just arrays!). This allows for more memory-efficient data processing by avoiding the creation of intermediate arrays. 1️⃣ Native Set Operations: ES2025 introduces native methods for Set objects, including union(), intersection(), difference(), symmetricDifference(), isSubsetOf(), isSupersetOf(), and isDisjointFrom(). 2️⃣ Native JSON Modules: Importing JSON files has standardized syntax - import config from './config.json' with { type: 'json' };. This simplifies a common workflow by eliminating the need for fetch or file system reads followed by JSON.parse(). 3️⃣ RegExp.escape() & Enhancements: a native way to escape strings for use in regular expressions, helping to prevent security vulnerabilities like ReDoS attacks. This release also includes inline modifiers for changing regex flags within a pattern and better support for named capture groups. 4️⃣ Promise.try(): This new static method on Promise provides a clean way to start a promise chain with a function that might throw a synchronous error. It wraps the execution in a promise, catching any exceptions and turning them into a rejected promise, which can simplify our async error handling. 5️⃣ Float16Array: A new TypedArray for half-precision 16-bit floating-point numbers. This is a significant addition for performance-critical applications like graphics programming, scientific computing, and machine learning, where we can trade some precision for a smaller memory footprint and faster calculations. 🔥 This is a powerful release focused on reducing boilerplate and improving the developer experience. 🔥 We can start experimenting with most of these features today in the latest versions of Node.js, Deno, and modern browsers, or by using Babel. For a deep dive into all the new features, check out the official specification: 🔗 ECMA-262 16th edition - ECMAScript 2025 (https://lnkd.in/dxBghb2Z) JavaScript continues its evolution, becoming a more robust and elegant language with each update. #JavaScript #ES2025 #WebDev #ECMAScript #DeveloperTools #Programming #TechTrends

12

Flaky tests feel random. They’re not. The patterns are clear: * Timing issues (async workflows, race conditions) * Shared state (bad test isolation, order dependencies) * Environment drift (things that work locally but fail in CI) By definition, though, they’re not obvious to spot. A missing await, an unseeded random number, a fragile assertion. These problems tend to show up after the developer who wrote them has moved on to the next thing, reducing the likelihood that anyone will take time to fix them. AI is going to multiply these issues. It’s *also* going to help us catch them faster. The challenge for leaders: don’t accept flaky tests as inevitable. They’re just tech debt we’ve been too slow to pay down.

43

Developers don’t read your docs the way you think they do. They’re not sitting down with a coffee and reading cover to cover. They’re: → Googling your product + error → Scanning examples → Copy-pasting whatever gets them unblocked fastest And yet so many teams write docs like textbooks. Linear. Dense. Outdated. Here’s the fix: Structure around how developers search. Lead with examples. Surface the top 10 “how do I...” questions and answer them clearly. Your docs should work like a great search result, not a PDF.

73

9 Comments

Today I tried to deploy my new Elixir app to a PAAS platform. 2 hours in? I'm still trying to figure out why the deploy is failing. No coherent errors. In fact, at one point, no errors at all. Just "hey, the error has been dumped to a file in disk that you can't now access". AI was the one saving grace here. Cursor figured out why it might be failing. But then I got a huge wall of an error with no clear reason what's going wrong. And this is very common. In fact I'd go so far as to say it's the norm. When things work, great. When things fail, the error message is usually obscure. So what can we do? Well, think about the failure cases. Think about what happens when things go wrong. What information will you want your clients to know? Think about your clients. Because one day that client might be you.

9

2 Comments

❓A lot of developers ask me, particularly in workshops, ‘what happens if I do X?’ or ‘what happens if I load Z before Y?’-style questions. The thing is, if I don’t know the answer, I have the same tools as everyone else in order to work it out. Build a demo and see! To help, here are the two tools I lean on most frequently when I’m making small proofs of concept to demonstrate resource loading: 1. https://slowfil.es – a tool that Ryan and I built many moons ago which provides slow-loading resources on demand. 2. The smallest viable HTML skeleton page: <!doctype html> <html lang=en-gb>  <​meta charset=utf-8>  <​meta name=viewport content="width=device-width, minimum-scale=1.0">  <title></title> Armed with those two tiny utilities, I can build and model many different resource loading scenarios within minutes. Enjoy!

48

7 Comments

I've been reviewing a tons of DDD examples from various articles lime rhis over decade now and they all have these same issues: - Excessive network usage back and forth with database engine. - Over selecting too much data. - Riddled with race conditions problems. - They all look the same no matter which style they claim: clean architecture, hexagonal, whatever. - They all think that modern database engine is mere IO device tyat must be abstracted with repository which tells me they don't have a clue. https://lnkd.in/ebyuMA4J

21

17 Comments