Samuel Urbanowicz

Another Kotlin Meetup is happening this week! 🎉 On 𝗧𝗵𝘂𝗿𝘀𝗱𝗮𝘆, 𝗝𝘂𝗻𝗲 𝟭𝟮𝘁𝗵, there will be two great talks from my fellow colleagues at GoodData: First, Jiří Mikulášek will present "𝗗𝗜 𝗼𝗿 𝗗𝗶𝗲? 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗥𝗲𝗮𝗹𝗶𝘁𝘆 𝗖𝗵𝗲𝗰𝗸 𝗳𝗼𝗿 𝗞𝗼𝘁𝗹𝗶𝗻 𝗕𝗮𝗰𝗸𝗲𝗻𝗱," exploring the real-world performance impact of popular dependency injection frameworks and helping you decide when to use them for optimal performance. Then, Peter Plocháň will present "𝗧𝗮𝗻𝗴𝗹𝗲𝗱 𝗶𝗻 𝗖𝗼𝗿𝗼𝘂𝘁𝗶𝗻𝗲𝘀: 𝗙𝗶𝗴𝗵𝘁𝗶𝗻𝗴 𝗕𝘂𝗴𝘀 𝗶𝗻 𝗢𝗦𝗦 𝗟𝗶𝗯𝗿𝗮𝗿𝗶𝗲𝘀 𝗮𝗻𝗱 𝗛𝗼𝘄 𝗡𝗼𝘁 𝘁𝗼 𝗨𝘀𝗲 𝗞𝗼𝘁𝗹𝗶𝗻 𝗖𝗼𝗿𝗼𝘂𝘁𝗶𝗻𝗲𝘀." He'll reveal common coroutine pitfalls he's encountered, from broken cancellation to context confusion, and show you how to avoid them in your own async code. Don't miss out! Come join the discussion, bring a friend, and learn something new 🙂

4

1 Comment

Yesterday I spoke at Devoxx Poland in #Krakow about #GarbageCollection and it was fun 😁 As always before a session I‘ve modified my slides and I was able to make it through 188 slides in 50min…that was tight and I directly reduced the number of slides again after my session. It’s really not that easy to adapt the number of slides to the available time, esp. if you will find sessions with lengths of 20, 25, 30, 40, 45, 50 and 60 minutes. I wish conferences could stick to the 45min format 🤷🏼‍♂️ The #TrashTalk tries to cover memory management, esp. garbage collection in the #Java #Virtual #Machine (#JVM). It’s not a deep dive on this topic but more an intro and overview session that should provide everything you need to know to get a better understanding about garbage collection. Once you understand the principles and know how the collectors in #OpenJDK work, you should be able to make a more educated decision when it comes to selecting the right collector for your application. Collectors are programs that have their use case and requirements. E.g. a fully concurrent collector runs (as the name implies) concurrently to your application which means it will require CPU and memory resources that won’t be available for your application. Meaning to say, you will probably see a drop in application performance when switching from a non-concurrent to a concurrent collector on the same machine (depending on the available CPU and memory). That also means, systems and their software are different and it needs monitoring + tuning to find and configure a garbage collector for your application. So my advice is…do not simply use something because someone else is using it 😁

68

2 Comments

🎉 My first PR to OCA just got merged today — and I couldn't be happier. It took 8 months, but not because the change was hard. OCA maintains hundreds of modules used by thousands of companies worldwide, and the bottleneck is always the same: not enough contributors. These are volunteer developers and reviewers keeping one of the most important open source ERP ecosystems alive. If you've ever built something on top of OCA modules — and chances are you have — you've benefited from their work. Now it's our turn. If you're an Odoo developer, I want to challenge you: open your first PR to OCA. It can be a small fix, a missing test, a documentation improvement. The barrier to entry is low. The impact on the community is real. Let's make the review queues shorter. Let's make the merge times faster. Let's make OCA stronger. 👉 https://lnkd.in/dA-utxpm #Odoo #OCA #OpenSource #GiveBack #Python #BackendDevelopment

20

I can’t even remember who first introduced me to NetBeans IDE about 14 years ago. At the time, I was coding with Notepad++, and honestly… if I had stuck with that, I might have quit coding altogether 😂. That experience reminds me of how I view AI coding tools today. They’re not here to replace us, but to supercharge us. I don’t lean on AI much when working with languages I’ve mastered over the years, but when I do, the productivity boost is undeniable. For me, AI coding feels like the next evolution of IntelliSense that not just suggest snippets, but accelerating problem-solving, learning, and creativity. The way we code is evolving. The real question is: are we evolving with it?

11

Thanks to a few people around me (looking at you, Lars Marowsky-Brée, Anna Maresova, Michal Vyskocil, Petr Baudis), I have a better understanding of the current #AI iteration¹ and I have finally started looking at potential use cases. Yes, I know I'm late. But it's a tough search nevertheless. You need something that: • takes longer to produce than to verify, • does not cause too much harm if wrong, • is economical. IIUC this so-called #GenAI is a powerful pattern-matching engine (like traditional AI/ML but with way more parameters) whose output is generated by a transformer, pre-trained on a large corpus of text. Quite importantly, this pattern matching is still rather superficial. Since I'm not in a position to train my own specialized model, the input should resemble what commercial #LLM engines were trained on (plus some limited in-context learning), or I'm very likely to get pure garbage (colloquially called “hallucination”). Unfortunately, the bulk of my job (~80%) requires deep understanding and real intelligence (i.e. infer from observed behaviour how a system works internally). AFAICT #LLM cannot help with that. So, I'm now looking for potential applications in the remaining ~20% of my job, but at this point I'm afraid statistics suggests that #AI won't boost my productivity by more than 20%. Well… ¹ The term was coined long before I was born and has been used for different things. Go figure out.

25

5 Comments

As technical founders, it’s tempting to obsess over clean architecture and elegant code. But remember: users don’t buy code, they buy solutions. Build for value first, then polish. In that order… #saas #tech #product

2

🚀 After wanting to do this for MONTHS... I finally hit publish Just published an article giving a clear, practical overview of performance optimizations ⚡️ Not overly academic. Not unnecessarily deep. But the kind of foundation every frontend engineer should understand. I talk about: ⚡️Why apps feel slow 🧠 What actually causes unnecessary re-renders 🔍Where performance bottlenecks usually hide 🛠 Simple optimizations that make a real difference Performance has always been something I care deeply about. There's something satisfying about turning a laggy experience into something that feels instant. This is just the start. 1'1l be sharing more as I continue learning and building. If you're building with React/ React Native, this might be useful👀 #reactjs #reactnative #webdevelopment #mobiledevelopment #developer #poojachawla #cross_platform

2