Paul Sabou

🚀 Join us for an exciting live session on 𝑴𝒐𝒃𝒊𝒍𝒆 𝑨𝒑𝒑𝒍𝒊𝒄𝒂𝒕𝒊𝒐𝒏 𝑴𝒐𝒅𝒖𝒍𝒂𝒓𝒊𝒛𝒂𝒕𝒊𝒐𝒏 with 👾 Tjeerd In 't Veen https://lnkd.in/eg2hcCkG Are you looking to deepen your understanding of modularization and its critical role in modern app development? Don’t miss out on this insightful discussion where we'll cover: 🔍 𝐔𝐧𝐝𝐞𝐫𝐬𝐭𝐚𝐧𝐝𝐢𝐧𝐠 𝐌𝐨𝐝𝐮𝐥𝐚𝐫𝐢𝐳𝐚𝐭𝐢𝐨𝐧: What does it mean to modularize a mobile app, and why is it crucial for today’s development landscape? 🔍 𝐁𝐞𝐬𝐭 𝐏𝐫𝐚𝐜𝐭𝐢𝐜𝐞𝐬 𝐟𝐨𝐫 𝐂𝐨𝐝𝐞 𝐑𝐞𝐯𝐢𝐞𝐰: Discover effective strategies for conducting code reviews in feature modules, ensuring high-quality code and smoother collaboration. 🔍 𝐁𝐚𝐥𝐚𝐧𝐜𝐢𝐧𝐠 𝐋𝐞𝐠𝐚𝐜𝐲 𝐚𝐧𝐝 𝐍𝐞𝐰 𝐌𝐨𝐝𝐮𝐥𝐞𝐬: Learn how to progressively transition from legacy systems to a modular architecture while continuing to deliver new features. 🔍 𝐌𝐞𝐫𝐠𝐞 𝐂𝐨𝐧𝐟𝐥𝐢𝐜𝐭𝐬 & 𝐕𝐞𝐫𝐬𝐢𝐨𝐧 𝐂𝐨𝐧𝐭𝐫𝐨𝐥: Get practical tips on managing merge conflicts, version control, and maintaining consistency across different modules. Whether you're a seasoned developer or just starting out, this session is packed with valuable insights that can take your skills to the next level. 🕗 Time: 1 PM (Paris, France) 🇫🇷 / 1 PM (Amsterdam, Netherlands)🇳🇱 📍 Location: Only on Swift Academy 🇹🇳 See you there, and may the code be with you! ----------------------------------------------- 🚀 Support my work if you'd like! ☕️: Buy Me a Coffee https://lnkd.in/ei7cVejt 📨 Don't forget to subscribe to the newsletter of our site Swift Academy - Modern Mobile Development. https://lnkd.in/eUvvBk46 🎥You can also subscribe to my Youtube channel https://lnkd.in/eHKCft-E #MobileDevelopment #Modularization #CodeReview #VersionControl #AppDevelopment #TechTalk

36

I published my first blog post about statically deobfuscating the Lazy Importer. Please check it out and don't forget to let me know your thoughts. Blog Link: https://lnkd.in/eW-9QV_2 .

4

🚀 Exciting News for Developers! Join us on Friday, November 15th for an in-depth session on the upcoming new major version of the Fluent UI Blazor library. This release brings groundbreaking Web Components and major architectural innovations to take your web development to the next level. With Vincent Baaij https://www.dotnetconf.net 🔍 What to Expect: Overview of New Features: Explore the latest enhancements, including cutting-edge Web Components. Architectural Changes: Understand the performance, scalability, and maintainability improvements in the new architecture. Migration Guide: Yes, there will be breaking changes! But we’ll guide you with practical tips and best practices to smoothly transition your projects. Hands-On Demos: Watch the new features in action with live coding demonstrations. 📅 Don’t miss out! Mark your calendar and get ready for this transformative session. #Blazor #FluentUI #WebDevelopment #WebComponents #Architecture #Migration #LiveCoding #TechTalk #WebInnovation

38

2 Comments

Are you someone who wants to make a real impact and look back one day and say, "𝐈 𝐰𝐚𝐬 𝐩𝐚𝐫𝐭 𝐨𝐟 𝐭𝐡𝐚𝐭"? Then you (🫵🏽) might just be the developer who can help bring Starflow’s vision to life! Let’s get one thing straight: working at Starflow isn’t just about writing code. It’s about creating something new from the ground up, breaking boundaries, and tackling challenges head-on. But rest assured – the founders promised me there’ll also be plenty of fun along the way! Curious if you and Starflow could be the right match? Check out the job ad in the comments! 👇🏽 Jonas Helmikstøl Nikolai Konopelko Ola Stengel

27

2 Comments

Alright, it's a little behind schedule, but v4.15.0 of Nebula Logger is now available! This release brings several new features & bugfixes to the managed package (as well as few small enhancements to the unlocked package). Some of the highlights for anyone upgrading from v4.14.0 to v4.15.0: Key Features ✅ OmniStudio Logging: Built-in support to easily log from OmniStudio, a much-requested feature! ✅ Loosely-Coupled Dependencies: ISVs and consultants can now seamlessly leverage Nebula Logger without requiring it as a dependency. ✅ Enhanced JavaScript Logging: A new getLogger() function makes JavaScript logging even easier & provides improved support for logging in LWC lifecycle events Other Enhancements ℹ️ Dynamic highlights and related list components for flexipages, leveraging Salesforce's Winter '25 features. ℹ️ New list views and formula fields on logging objects, improving usability and visibility. ℹ️ Apex updates for better HttpRequest logging, async context handling, and more flexible LogEntryEventBuilder methods. ℹ️ Optimized internal performance and streamlined sf CLI commands to align with Salesforce’s latest standards. Bugfixes 🐞 Enhanced regular expressions for data masking sensitive info (e.g., SSNs). 🐞 Fixes to Apex context handling, long code snippets, and visibility rules on console tabs. More details are available in the release notes: https://lnkd.in/eBKVkfZr #Observability #Logging #Salesforce #SalesforceAdmins #SalesforceDevelopers #SalesforceArchitects #OpenSource #Apex #OmniStudio #ISV #LWC

99

7 Comments

My new open source project is called "erwtensoep" (Dutch for "Pea soup") and it contains a helpful subcommand that yields my favorite recipe: ``` ❯ make && bin/erwtensoep recipe go build -o bin/erwtensoep cmd/*.go Ingredients: 500 gram of split peas (dried are fine) 2 decent sized boneless pork chops 200 gram of panchetta 2 liter of water 2 medium sized onions 2 leeks 900 gram celeriac 180 gram carrots 7.5 gram of fresh thyme 2 dryed bay leaves 2 beef Kielbasa sausages All of this can be obtained at the Albert Hein super market in the Netherlands or at the Wholefoods in the US. Step 1) Bring the peas to a boil along with the pork shoulder chops, pork belly strips, and water. Let it simmer on low heat for 30 minutes. Regularly skim the foam from the soup and stir well. Frequently scrape the bottom of the pan to prevent the soup from burning. Step 2) Meanwhile, finely chop the onion. Wash the leek and cut it into thin rings. Peel and dice the celery root into 1 cm cubes. Peel the carrot, cut it lengthwise into quarters, and then into 1 cm cubes. Strip the thyme leaves from the stems. Step 3) Add the onion, leek, celeriac, carrot, bay leaf, and thyme to the soup, and let it simmer gently for 30 minutes with the lid on the pan. Season with pepper and, if desired, salt. Step 4) Remove the pork shoulder chops and pork belly strips from the pan and cut the meat into small cubes. Slice the smoked sausage and stir it into the soup along with the meat. Let the soup simmer gently for another 15 minutes. Divide the soup into bowls and serve. Recipe courtsey of https://lnkd.in/eExnX_sC. Translation by ChatGPT 4o. ```

6

2 Comments

That's Pieter Levels. The guy who created a bunch of lucrative start-ups that are generating him $3M/year. Did he use Angular, React or Vue? No, just good old PHP, HTML, CSS and jQuery. No fancy design patterns or architecture either, just inline code. Did he use some fancy IDE? No, wrote all using Sublime text, recently switched to VS Code. Is he paying $20K / month for hosting with containers and K8S? No, just $400 for a single VPS with 16 cores and 64GB of RAM. And that's for all of his apps combined. Simplicity is the ultimate sophistication. #programming #webdevelopment

1

♦ Wrong mindset as a developer: “Why do I need to help unblock others? They should be able to take care of it on their own. I will focus on my own work.” ♦ Right mindset as a developer: “My team is blocked on this particular issue. I read this helpful blog yesterday. I will forward this with a summary. I think it will help a great deal, at the end of the day, we need to ace this project” Great software is always a collaborative effort. The sooner you understand that, the better you will do as a developer. 

127

5 Comments

Update: CodeSee has been acquired by GitKraken! --- I am very sad to announce - officially - that CodeSee will be closing its doors and shutting down commercial operations on Thursday, Feb 22nd. All paying customers and free users can continue using CodeSee until Feb 22nd. We have a team that will maintain the service during this period and continue to support you. Thank you to everyone Thank you to all the customers who partnered with us. You were invaluable for helping us build CodeSee. Thank you to the open source community who were our closest collaborators from the very beginning. Thanks to all the employees who joined CodeSee. We’re proud of both the team and the product we built. Thanks for taking a bet on this startup journey. Lastly, thank you to all our investors who bet and supported us on our journey. Why we made the decision I can only apologize to everyone for this difficult turn. The company had more than doubled the number of new product logos in 2023 and grew by many thousands of free users. However our sales growth was inconsistent, and we needed to make revenue faster and more consistently. Given the nature of trying to solve code understanding, this meant being able to deeply cover more complex codebases, more IDEs, more tech stacks, and more programming languages which would have taken time that we didn’t have. And Gen AI has made this space even more complex, both as a tool to help and a source of ever more code to understand. As any startup founder, I had to make some tough calls on where to focus our efforts, but in the end those choices did not lead to the consistent revenue we needed for us to continue. Although we did receive a term sheet to continue going, we ultimately decided that it wasn’t enough to cover this mandate and deliver a truly valuable product. Here's our goal with CodeSee right now... Plan A: - Complete the process of finding a home for the tech and team. We're already in talks with several companies now and want to find the best mutual fit asap. Plan B: - If we don't find a good mutual fit, the team will (sadly) go our separate ways. Code Understanding is still awesome After this journey, we have no less enthusiasm for developer productivity and developer experience. We still think that when developer’s lives are easier, it’s pure magic. After speaking with thousands of companies, it’s clear that code understanding is still a vital problem to solve. It’s been an absolute pleasure to build tools to bring new and impactful ways of understanding code to the developer community. Even though CodeSee’s journey may be coming to an end, our passion to serve this amazing community is as strong as ever.

747

126 Comments

Being in this profession for almost 28 years, I like to think I've seen the evolution that developers go through. In fact, I've experienced it firsthand and made my fair share of mistakes. I don't think terms like junior, medior, and senior make a lot of sense anymore, though. I've seen aspiring developers whose technical skills and breadth of knowledge far exceeded those of developers with 15+ years of experience. That’s why it’s so important not to judge a book by its cover. Nonetheless, experience is something to take into account when working with developers on your team.   In general, developers fresh out of school have a lot to learn. The gap between what they are trained for, and the reality of our profession is simply too big. They'll be introduced to many new techniques, so giving them overly strict guardrails isn't prudent. Even those who are self-educated need some guidance. Over time, you'll notice that they start to question those guardrails and offer suggestions that show they’re becoming more aware of the rationale behind them.   This is the point where you need to remove the guardrails to give them more room to explore and fail safely. This is also the fun period. You'll have great discussions about why certain patterns or principles make sense in particular scenarios—and when they don’t. Ironically, I've learned that developers with 15-20 years of experience are often the most dogmatic and the hardest to work with. And obviously, those with 25 years or more know exactly what they're doing, right… Or not? 🫣 #lessonslearned #consultancy #avivasolutions #carreerpath #softwareengineering

16

In 2021, I tried to understand how tech compensation really works - at least in the Netherlands. I published "the trimodal nature of software engineering salaries" article. Time for an update: Back then, based on ~100 data points (and my own experience as a hiring manager,) I suspected that there are three, distinctively different compensation bands for software engineers: 1. Local companies (Tier 1) Most local businesses that benchmark to the numbers on most salary comparison sites. They usually offer close to – and sometimes slightly above – what government positions advertise for senior engineering roles. 2. Ambitious local companies (Tier 2) Startups, scaleups, and more ambitious local companies that want to hire and retain the best people, locally. 3. Companies benchmarking across a region (Tier 3). Big Tech companies, hedge funds and scaleups that hire across regions, frequently hire from abroad, and often compete with fellow Big Tech giants for talent. Curiously, the third tier (Tier 3) seemed to be missing from most public compensation data sources. Since publishing the model, I've both received lots of feedback suggesting it applies to the US, Canada, UK, Europe and many other countries. I've since then also collected far more data points, manually tagged companies by tier, and can finally confirm that yes, this model holds up when compared with actual data. Here is the longer analysis (and the stats behind this graph below) in yesterday's The Pragmatic Engineer issue. Read it here: https://lnkd.in/e7FV4gBB

708

43 Comments

Resharing this as endorsement - tech compensation is *at least* bimodal, and likely trimodal (depending on your location). This article is a great read for all engineers, to help explain the market and know their worth. I saw this firsthand in my job search last year. There were local companies in the Chicago area that capped out at ~120-160k for senior roles, and then larger firms (especially in finance and tech) that reached ~180-200k. But it wasn’t until I got into interviews with large public tech companies that I saw base salaries of 200k+ advertised, and total compensations reaching from 250k to 500k and up! Gergely Orosz's articles on this topic have been a huge help to me. If I had stuck to what I saw in the local market, I would've drastically limited my options and short-changed myself. The surge in remote work has made it more possible than ever to work for some of the best companies in the world, and even with lots of return-to-office happening, there are still plenty of great and well-paying companies out there. So if you're on the market (or thinking about it), keep this in mind! 🚀

6

I'm currently working on a web application that is 100% built/coded by generative AI. It will soon be available in beta to anyone. So I thought, why not redesign and code my personal website 100% by AI. Here is the result: https://lnkd.in/evFJQwaG I let Claude 3.5 Sonnet do the design part into a HTML mockup and transformed it with ChatGPT 4o into a Wordpress theme. I absolutely did not anything other than copy paste code. What do you think? I used the starting prompt: "Act as a visual designer. I need a design for a personal website. It must be visually appealing, modern, and responsive. I am a digital strategist and digital transformation lead. My focus includes consultancy, advisory work, customer experience, generative AI, and guiding clients in their digital ambitions, CX via digital channels, composable architecture, and more. Create the design using vanilla HTML, CSS, and JavaScript in the artifact window." From there I prompted for an hour or so in refining, changing and improving the design, animations, etc. #generativeai #artificialintelligence #ai #claude #chatgpt

16

3 Comments

Got inspired by a dutch podcast that creates stories from wikipedia pages. I used gpt-4o and their text-to-speech model to build something that sounded human. Of course it's not as great as the original podcast but this tech works way better than expected! Try it yourself 👇 Huggingface Space: https://lnkd.in/eKdT3XBY How did I do it? - Use <break time="1s"/> tag on strategic places to create breaks. - The TTS model is smart enough to understand these tags! - Prompt to use colloquialism english. - Use filler words like "um", "well", ... Bonus: I also generated an intro song using facebooks "MusicGen" and tweaked it a bit with with online tool called "AudioMass" Tools I used: AudioMass: https://audiomass.co/ MusicGen: https://lnkd.in/eAfjHZzA If your dutch speaking; the original podcast is called "Nooit Geweten", I can highly recommend: https://lnkd.in/ex66X6_a #gpt-4o #text-to-speech #tts #huggingface #spaces #podcast

14

2 Comments

My talk at #FrontendNation is available in full length on youtube. Let me know what you think.👇🏼 https://lnkd.in/e_GBNXTv I will post a link to an automation guide of my talk as comment. In this blog post you can also find the slides. ✅ #a11y #FrontendDevelopment #wcag #WebAccessibility

32

2 Comments

OpenAI didn't have a great week, losing out twice: first to Runway's Gen-3 and then to the Moshi voice AI model. Here’s what else you missed in AI this week: 🧠 𝗜𝗺𝗮𝗴𝗲𝘀 𝗳𝗿𝗼𝗺 𝗯𝗿𝗮𝗶𝗻 𝗮𝗰𝘁𝗶𝘃𝗶𝘁𝘆 👉 Radboud University developed AI to reconstruct images from brain activity recordings accurately. 👉 The AI uses fMRI and electrode recordings, significantly improving reconstruction accuracy. ​ 👉 This tech could help stroke victims and recreate dreams, despite dataset limitations. 💬 𝗠𝗼𝘀𝗵𝗶 𝘃𝗼𝗶𝗰𝗲 𝗔𝗜 👉 Kyutai's Moshi, a real-time AI voice assistant, offers 70 emotions and speaking styles. 👉 Moshi features 160ms latency, surpassing OpenAI, and will be open-sourced soon. 👉 Developed by 8 researchers in four months. 🍏 𝗔𝗜 𝗼𝗻 𝗔𝗽𝗽𝗹𝗲’𝘀 𝗩𝗶𝘀𝗶𝗼𝗻 𝗣𝗿𝗼 👉 Apple is adapting its new AI features for the Vision Pro headset. 👉 Features include improved Siri, writing tools, and an OpenAI-powered chatbot. 👉 High price tag might deter users, despite advanced AI capabilities. 🎥 𝗥𝘂𝗻𝘄𝗮𝘆’𝘀 𝗚𝗲𝗻-𝟯 𝗔𝗹𝗽𝗵𝗮 👉 Runway's AI video generator Gen-3 Alpha is now available to all users. 👉 Gen-3 Alpha improves character consistency, camera motion, and scene transitions. 👉 Limited generation time for standard plans may challenge power users despite impressive features. 🏃‍♀️ LLM routing framework 👉 RouteLLM is an open-source framework for cost-effective routing between large language models. 👉 It routes queries between expensive and cheaper LLMs, optimizing cost and performance. 👉 RouteLLM achieved cost savings while maintaining 95% of GPT-4's performance in tests. Did you like this post? Please share ♻️ it with your network! Let me know about your favorite AI news this week in the comments 👇 --------- ➕ Follow Sebastian Illing for the latest AI news. ♻️ Repost to keep your network up to date.

39

18 Comments

🚀  Llama 3.1 is a game-changer! I had to blink my eyes today when I ran some tool-calling experiments (in Dutch): the 70B version performed perfectly on tough cases, just like GPT-4o. And Officially Llama 3.1 does not even speak Dutch. I ran it on a double RTX 4090 GPU for 80ct/h (20 euros a day)! ‘Tool calling’ is the glue between human speech and the system executing the spoken request. So far self-hosted LLMs were not strong enough to do this reliably even after fine-tuning, so an external LLM like OpenAI GPT-4 was always needed. No more! Now banks, insurance companies, and governments who are afraid of privacy and security issues with external LLMs can run a high-quality model on 4000 euros of graphics card! The limit is gone! Last but not least, the 8B version (which you can host for 15ct an hour) is also really good at tool-calling. It has minor glitches, but if your use case is not too demanding it is a blazingly fast-responding little beast.

26

6 Comments

The 𝘂𝗻𝗶𝘁 𝗼𝗳 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴 𝗼𝗿 𝗿𝗲𝗮𝘀𝗼𝗻𝗶𝗻𝗴 could be a combination of time, energy, cognitive complexity, and neural activity, but for practical purposes (like comparing to AI), 𝘁𝗶𝗺𝗲 𝗮𝗻𝗱 𝗲𝗻𝗲𝗿𝗴𝘆 𝗰𝗼𝗻𝘀𝘂𝗺𝗽𝘁𝗶𝗼𝗻 are often used as proxies. So, to normalize the relationship between "human thinking time" and "ChatGPT query" suggests that 𝟳 𝘁𝗼 𝟭𝟰 𝗺𝗶𝗻𝘂𝘁𝗲𝘀 𝗼𝗳 𝗵𝘂𝗺𝗮𝗻 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴 is roughly equivalent to one ChatGPT query in terms of energy consumption or carbon emissions. This means that if a task can be completed in 𝗹𝗲𝘀𝘀 𝘁𝗵𝗮𝗻 𝟳 𝘁𝗼 𝟭𝟰 𝗺𝗶𝗻𝘂𝘁𝗲𝘀 𝗼𝗳 𝘁𝗵𝗶𝗻𝗸𝗶𝗻𝗴, it’s more efficient to use your brain. However, this is a simplified comparison, and there are many other factors to consider. The human brain operates with immense complexity—billions of neurons firing in parallel, processing information through axons and synapses at remarkable speeds, potentially even at a quantum level. While AI can handle specific tasks quickly, the brain's intricate, concurrent parallel processing makes it incredibly efficient for diverse types of problem-solving that 𝗔𝗜 𝗰𝗮𝗻'𝘁 𝘆𝗲𝘁 𝗲𝗮𝘀𝗶𝗹𝘆 𝗺𝗮𝘁𝗰𝗵 𝗶𝗻 𝘁𝗵𝗲 𝗯𝗶𝗻𝗮𝗿𝘆 𝘅𝗣𝗨 (𝗖𝗣𝗨/𝗚𝗣𝗨) 𝗲𝗿𝗮. 𝗠𝗮𝘆𝗯𝗲 𝗶𝗻 𝘁𝗵𝗲 𝗤𝘂𝗮𝗻𝘁𝘂𝗺 𝗲𝗿𝗮. In short, while AI might be better for certain tasks, the brain’s unparalleled complexity and efficiency can't be overlooked.

2

About the difference between OpenSpace Beta and the so-called "Agile scaling frameworks" (some of which claim not to be frameworks): https://lnkd.in/ePygVfen

1