David Camp

🚀 Alright, buckle up, because .NET 10 just dropped some serious performance magic and, wow, it’s honestly pretty wild if you’re into this stuff. First off, the JIT (that just-in-time thing that makes your code actually run) has gotten a glow-up. We’re talking legit support for AVX10.2 instructions—yeah, that’s nerdy, but it basically means it chews through data way faster. Toss in struct stack allocations and slicker method inlining, and you’re looking at a runtime that’s moving like it’s had three espressos. And get this: memory management is actually smarter now. Fewer heap allocations, so your apps aren’t tripping over themselves or gobbling RAM like a starved Pac-Man. That means less crashing, better stability, and your code doesn’t feel like it’s dragging a boulder uphill. So, if you’re building stuff that actually has to scale and not just survive on your laptop—.NET 10 is kind of a game-changer. Faster apps, happier users. 👏 #runtime #optimization #softwaredevelopment #dotnet10 #notyouraverageupdate #yassineennajem

7

I've been documenting software for years, but last night’s discovery about developer blind spots genuinely surprised me... The wild part: sometimes the devs don’t fully understand their own product. I just ran into it again. There’s a disconnect between three things: what devs think the feature is, what the product actually shows on screen, and what users understand when they try to use it. That gap isn’t small. I spent two hours decoding inconsistencies between the dev docs and the UI. Same feature. Different words. Different meanings. No wonder people get stuck. This is why a documentation expert does more than “write docs.” We create clarity where there isn’t any. We line up the language so the product, the docs, and the user are speaking the same thing. If your team feels confident but your users feel confused, check the words. Where do your docs and your UI disagree?

10

I’ve seen too many junior devs burn out after their 12th tutorial in a row. Here's something that works better: 1️⃣ Do the tutorial once 2️⃣ Rebuild it without looking 3️⃣ Change it. Make it weird. Break it. Fix it. That’s where learning happens. The engineers I’ve mentored who leveled up fastest weren’t the ones who had the most certificates. They were the ones who got stuck, debugged the issue, and took the time to understand what broke.

62

15 Comments

Experienced developers: Stop saying you're "just a developer." You're not. You're a problem solver who happens to use code. I see 10+ year veterans underselling themselves every day: "I'm not ready to lead a team" "I'm never gonna be better than the 10x developers" "I'm not senior enough for that role" Meanwhile, you've: - Created products that made your boss rich - Prevented system outages that would cost millions - Built features used by thousands of users daily - Mentored junior developers who are now thriving But you focus on what you DON'T know instead of what you DO. Here's the truth: Imposter syndrome isn't about lacking skills. It's about lacking perspective. You compare your behind-the-scenes to everyone else's highlight reel. Stop it. Your experience IS valuable. Your solutions ARE needed. You DO belong in those senior roles. The companies that don't see your worth aren't worthy of your talent. Find the ones that are. ♻️ Repost if you needed to hear this today. What's one thing you accomplished this year that you're proud of? Share it below 👇

22

9 Comments

An unexpected repercussion of vibe coding that I've noticed: it really magnifies the quality of your developer experience. A lot of folks have compared coding AIs to something like an over-eager intern: energetic and prolific, though frequently lacking in experience and judgment. If you want to get the most out of it, ask yourself: how good is the day-1 experience of interns on your codebase? Because vibe coding forces you to replay that day-1 experience over and over. The quality of a coding agent's output improves dramatically if you give it a good fitness function to optimize for. It's one thing to ask, "just spit out code for this." It's quite another to say, "fix this bug, add tests to this file, make sure lint passes and the code compiles and all the tests pass before you talk to me again." It's vastly more successful with the guard rails of typecheckers and compilers and tests telling it whether it's achieved a minimum acceptable quality bar. If your typical testing flow is "fire up a local server and click around a few pages," then suddenly your high-velocity intern is going to struggle. Most agents won't even be able to *do* that; if they can, there's no clear pass/fail signal, so it has to exercise judgment in deciding whether its code works or not, and recall that this particular intern has terrible judgment. Maybe your tests are flaky or don't run locally. That's gonna ruin your intern's day. Maybe you sometimes need to start a docker container before you run your tests. That's an extra step your intern needs to figure out and will sometimes forget. If you want to be able to leverage background agents, then you need to have a simple and fast path to bring a blank new machine into a state where it can run your tests. If any of the steps involve "just figure it out," your agent will definitely fail. And of course, absolutely none of this is new. This has been true of carbon-based interns before silicon-based ones. These matters of codebase quality have always had an impact on your company's productivity. If you don't invest in the developer experience, you just pay a perpetual tax on the labor of your high-value employees. Predictability breeds velocity. And always has.

134

17 Comments

I've been noticing two prompting patterns that haven't gotten much attention. Programmers should recognize these patterns, but others may not. What are they? Commands and events. 🔹 Command prompts drive action. 🔹 Event prompts trigger reasoning. Together, they mirror how real systems scale. Prompting isn’t just about asking a question to an LLM. It’s also about architecture. Command prompting is like those in CQRS. You tell the model what to do. 👉 “Summarize this.” It’s direct, explicit, and non-negotiable. You get speed, determinism, and token efficiency. Event prompting is like event-driven development. You signal when and why to act. 👉 “When the user uploads a file, validate it.” This isn't just context; it's conditional logic. It enables Chain of Thought. Event prompts guide reasoning. They help the model think before acting. Like middleware for LLMs. Used together, these patterns form intelligent, scalable systems. ⚡ Commands define action. ⚡ Events trigger judgment. It’s not about choosing one. 👉 It’s about combining both just like we do in modern architectures. Next time you design a prompt, ask: 🔹 Am I giving an instruction, or creating a decision point? Both matter. Together, they make AI systems smarter, faster, and more aligned. How are you structuring your prompts? #PromptEngineering #CQRS #EventDrivenDesign #ChainOfThought #GenerativeAI #CustomGPT

19

12 Comments

AI Definitely Replaces Engineers I had some Python code in a personal project that prepares data for an Astro/Node.js website and decided to use Claude to improve it. Rather than refactor my existing 20-line script, Claude decided to completely rewrite it as a 200-line JavaScript file with raw string parsing for quotes and commas. When I suggested using a proper CSV library, it dutifully implemented my feedback. Then, like many engineers I've worked with, it got frustrated when the library didn't work on the first try and quietly reverted back to the manual parsing approach—while confidently telling me what an excellent job it had done. Technically, it did follow my instructions since it used the library... for about 10 minutes. When I finally pointed it to a specific, well-documented library, Claude implemented it correctly and once again praised its own brilliant solution. Fun Fact: My original Python file was only 20 lines because it used pandas to parse a gzipped CSV, and this would have taken 30 seconds for me to add a column. So yes, this AI session was exactly like working with past engineering teams. The only thing missing is Claude boasting to the business how much smarted it is than the rest of the team. #AI #SoftwareEngineering #TechHumor #Claude #Development

104

21 Comments

I guess I’m not in touch with the current generation of developers, but when I see someone post skill issue in a response to someone’s post it seems like a dismissive and useless comment. My guess is that anyone who says skill issue has never worked with the person they are posting the comment about. Everyone in this industry believe it or not is growing their skills which means that skill issue is a reductive and a meaningless point. No one has reached the precipice of knowledge in tech and never will, so get over yourself and join everyone who is learning everyday and getting better at the craft of building software to solve human problems.

103

17 Comments

Coding agents are often like over-eager interns. They start typing before they’ve finished listening. The problem isn't the code. It's the lack of a plan. I’ve been looking into a framework called Superpowers. It’s not just another tool for generating snippets. It’s a methodology for the AI. It forces the agent to stop. To ask questions. To brainstorm before the first line is written. No more "guessing" what you want. No more code without tests. No more hallucinations born from rushing. It turns the LLM into a disciplined engineer. One that uses test-driven development. One that breaks big problems into tiny, verifiable tasks. One that actually follows through. We’ve spent a year asking AI to work faster. Maybe it’s time we asked it to work better. If you’re using Claude Code or Cursor, this is how you move from "prompting" to "engineering."

33

1 Comment

One thing I've heard a lot in junior dev interviews a lot lately: "I just want a mentor to show me how to design a system the right way." Unfortunately, there is no "right way." Much like intelligence, it's not about knowing the answer. It's about understanding why that's the answer. Teaching someone that a certain system design is "correct" is actually a disservice. Why? Because every system is a unique snowflake of requirements and constraints. A "perfect" architecture for a high-frequency trading platform is a disastrous architecture for a mobile video chat which would be different than an enterprise SaaS app which in turn... System design isn't about following a manual; it’s about asking questions and navigating tradeoffs: * Latency vs. Throughput: Do you need it fast or do you need a lot of it? * Consistency vs. Availability: Can the data be slightly old, or must it be 100% accurate 100% of the time? * Cost vs. Performance: Do you have a blank check for AWS, or are we running this on a shoestring budget? * Build vs. Buy: Is this core to our business, or should we just pay for a SaaS? Stop giving them the answers and start telling them about the disasters (best way to learn is falling on your face, but you can prevent some of that). Start asking them the "Why." Instead of saying "Use microservices," ask them: "What happens if the network fails here?" The goal shouldn't be to teach them how to build this system perfectly. It's to teach them how to ask the right questions, avoid the common pitfalls, and build every system better than the last.

18

10 Comments

I don't write dedicated "unit" tests for every public method or every function. Some might interpret that as not needing high test coverage. Just to be clear, my test coverage is usually 100% or as near as dammit. There's no code that isn't tested.

31

2 Comments

Had an interesting conversation with a junior developer today that made me reflect on how much our field has changed. When I started working in software development, senior developers would tell stories about having to solve problems by flipping through physical books and manuals. By the time I hit my stride, Stack Overflow had become my go-to resource. Today, I watched a younger programmer tackle the exact same issue I had struggled with years ago. But instead of spending hours scouring Stack Overflow threads, they turned to ChatGPT and had quickly developed a working solution. What impressed me most wasn't the speed - it was their approach. They didn't just copy and paste the AI's response. They researched the solution, validated it was correct, and thoughtfully integrated it into their codebase. Each generation of developers has had better tools, but the fundamentals remain the same: understand the problem, verify your solution, and apply it thoughtfully. The tools evolve. The critical thinking never goes out of style. #Programming #AI #SoftwareDevelopment #ProblemSolving

20

You don’t need a title to mentor someone. You don’t need 10+ years of experience. - If you’re a mid-level dev who’s survived your first production incident - If you’ve refactored something that was painful the first time - If you’ve learned to ask better questions in PRs You’ve got something valuable to share. Odds are that if you're reading this, you DO have something valuable to share. Mentorship isn’t about having all the answers. It's about sharing your experience and offering guidance. Sometimes the best mentors are only a few "steps ahead". And that’s enough. That’s how the next generation levels up. Who has been a helpful mentor to YOU? ---- 📨 Sign up for my email newsletter! 🗣️ Share with your network!

33

9 Comments

Your cache probably isn’t broken because of LRU, it’s because you missed adding a TTL. I learned this the hard way while debugging a caching layer that seemed perfectly fine... until it wasn’t. LRU (or LFU) handles memory pressure, it kicks old data only when the cache is full. But if your keys never expire, stale data stays forever. On the other hand, TTL ensures freshness but can lead to a sudden spike in traffic when everything expires at once. Think of your cache like your fridge: 🧊 Eviction keeps things from overflowing. 🥛 TTL ensures you’re not drinking expired milk. ✨ Together, they keep your cache healthy, both clean and correct. Curious, which one caught you first: stale values or memory overload? #Caching #DistributedSystems #BackendEngineering #Performance #Scalability #LearnWithNayan Credits: https://lnkd.in/gZ7n7_NN

8

🤔 if you say trunk based development won’t work because it requires excellent engineers… 😢What are you saying about your current team? I love trunk based development because it’s a GREAT way to train newbie devs. The more jr, the better - I am not kidding. You want to gel a new team fast? Push. To. Main. Daily. Aim for 100% build success in CI from main every day from multiple devs, on main. Does it require other practices and principles? Yes, absolutely, but this idea that it requires engineering excellence… well that’s driven from leaders, full stop. If you’re a leader saying trunk based dev won’t work for your team because they’re not good enough or the other processes aren’t in place - ask yourself why you wouldn’t set that as the goal, strive for excellence, and start working towards it today? It is your technical and process debt - start paying it down🤷‍♂️ What rationale is there as a professional manager and leader of engineers, building valuable products, in NOT investigating and pursuing ideas from DORA or other research-based organizations? #meatbasedengineer

114

122 Comments

Thinking of staying in software development? I wrote up a few thoughts, observations, and opinions on the topic for anybody wanting to get into or stay in software development in the coming years. The changes coming aren't going to be easy, they're likely to be somewhat brutal, for a number of reasons. This particular post I've kept the politics and the initial history of why we're here out of the post and focused on what to do. But I've got more coming and for now would love your thoughts, observations, hot takes, and the like.

8

9 Comments

Your GitHub link on your resume probably doesn't matter. Same with the link to your portfolio site (unless you're a designer). Very rarely will anybody take the time to go sifting through your repos of code to figure out if they should interview you. Now this isn't to say that you shouldn't contribute to open source, or work on personal projects, but those aren't going to help you get an interview unless you can tell a compelling story about it on your resume. The goal of those used to be to show off coding ability, but now coding ability is starting to matter less. What will matter more is did you build a business of some sort? Were you able to impact users in some way, or make even a small amount of money? If yes, that shows your ability to use code as a tool to solve a problem, and that's the story that matters regardless of how the tools change and evolve with AI.

188

48 Comments