Scott Weberg

How are folks solving the rapidly evolving situation of non-tech or semi-tech folks using things like ChatGPT to automate their toil and doing so in ways that are to at least some extent anti-patterns? So I'm asking: do you make these folks commit to github? do you run static analysis on that code? do you run something like endpoint security management with special features now turned on? do you have an overview committee? etc etc. #cantcontrolpeoplereally

4

3 Comments

Payment rails are infrastructure, but infrastructure can be a product. The rail determines settlement speed, capital efficiency, compliance architecture, and operational overhead. Those aren't just technical details, they're the actual product.

4

Staff Consultant Brian Hughes breaks down the myth that software tests slow down developers. If you're doing it right, software tests can actually speed you up.

7

1 Comment

If you were a front-end engineer in 2013, you remember the chaos. Everyone was on jQuery. Life was simple. Then overnight you had to choose between Knockout, Backbone, Angular, Ember, and this weird little library from Facebook called React. Every week a new framework dropped. Every conference talk started with "so I rewrote our app in…" Your tech stack had a half-life of about six months. That's exactly where we are with AI tooling right now. ChatGPT. Claude. Copilot. Cursor. Codex. Gemini. Perplexity. Lovable. V0. Base44. I could keep going. Every Monday there's a new model, a new wrapper, a new "this changes everything" demo on X. If you're trying to pick the right tool to build with, it feels like 2013 all over again. But here's where the analogy breaks down — and this is the part people aren't talking about enough. The front-end wars eventually stabilized. React emerged as the dominant winner, the ecosystem consolidated, and we all moved on. That took roughly a decade. I don't think AI is going to play out the same way. Because AI is now building itself. The models are writing the code. The code is improving the models. The feedback loop is compressing what used to take years into months, and months into weeks. That's why this revolution feels different — the thing accelerating isn't just the tooling, it's the rate of acceleration itself. So what do you actually do in an environment like this? You stop trying to pick the "winner" and start building. Right now. With the best tools available today. Personally, I'm all in on what Anthropic is building. They invented MCP (Model Context Protocol) — the open standard that lets AI models actually connect to your systems and tools. Claude Desktop with its connectors, skills, and Cowork is genuinely powerful. Claude Code is a different animal entirely. And the rest of the industry is following their lead — OpenAI's Codex now has a CLI, Gemini is heading the same direction. If you're an engineer and you haven't picked up an agentic coding tool yet, here's my honest advice: grab one that has access to your actual system. Claude Code. Codex CLI. Even OpenClaw. Something that runs from your terminal, reads your codebase, executes commands, and operates alongside you — not in a sandbox. That's where the real leverage is. Not in chat windows. In your CLI. The front-end framework wars taught us something: the winners weren't the people who picked the right framework first. They were the ones who kept building while everyone else was debating. Same rules apply. The tools will keep changing. Build anyway.

11

1 Comment

Excellent review of the history of Frontend Architecture. Like a lot of the decisions we make, there are tradeoffs. https://lnkd.in/gb6qDzBg

7

🚀 NgRx 21 — 10 Years of State Management, Now Signals-First NgRx 21 celebrates a decade of NgRx while taking a big step toward Angular’s Signals-based future with the introduction of @ngrx/signals/events. 🔍 What’s technically significant in NgRx 21? 🧩 @ngrx/signals/events A new primitive focused on event-driven architecture using Angular Signals: • Explicit event modeling instead of action overload • Clear separation of events → side effects → state updates • Reduced boilerplate compared to traditional Actions + Effects • Fine-grained reactivity with Signals (no unnecessary stream emissions) This makes complex flows easier to reason about while keeping NgRx’s core principles intact. ⸻ ⚡ Signals-First Reactive Model • Moves away from RxJS-heavy patterns where they aren’t required • Enables synchronous, predictable state transitions • Better performance via fine-grained dependency tracking • Ideal for modern Angular apps using signal(), computed(), and effect() ⸻ 🧠 Improved Architecture & DX • Cleaner APIs across NgRx packages • More composable building blocks • Easier onboarding for teams new to NgRx • Stronger alignment with Angular’s recommended direction ⸻ 🏗️ Why this matters for real-world apps For large-scale Angular applications: • Less cognitive overhead in complex workflows • More maintainable state and side-effect management • Better testability (events and state updates are easier to isolate) • Smooth migration path from classic NgRx to Signals-based patterns NgRx 21 proves that enterprise-grade state management doesn’t have to be verbose — and that Signals can coexist cleanly with proven NgRx patterns.

10

I have heard a lot about the decline of junior roles, but here at Equals Money Railsr we disagree. What is declining is the need for junior developers who wait to be trained, work on small isolated tasks, and add value slowly. What is growing is the demand for associate developers who can amplify themselves with AI and learn at speed. One strong associate developer, working effectively with AI, can now deliver what used to require a small team. That means getting comfortable with AI coding agents like Cursor, Claude Code, Gemini CLI, and similar tools. Use them to build bigger features faster, but do not treat them as a black box. The real differentiator is focusing on skills AI cannot easily replace. Clear communication. Breaking complex problems into well-defined pieces. Strong domain knowledge. These are the skills that turn AI-generated code into reliable, valuable software. Junior roles are not disappearing. The bar is simply moving. The opportunity is for associate developers who are immediately useful, AI-native, and able to learn fast.

48

4 Comments

If your dev team isn’t documenting, they’re not building software... they’re building dependency! Lack of documentation doesn’t just slow you down. It locks you in, kills scalability, and makes every handover a nightmare. ✅ Want maintainable code? ✅ Want to scale your team? ✅ Want freedom to grow without bottlenecks? Start with documentation. It’s not optional—it’s foundational. Check out the full article here: https://lnkd.in/ghkmjHh2 #softwaredevelopment #startuplife #techleadership #foundertips #scaleup #productmanagement #devculture #maintainablecode #iS2Digital #iS2

6

Token-Oriented Object Notation(TOON) It is a new, lightweight data serialization format designed specifically to optimize data exchange with Large Language Models (LLMs). It is often referred to as "JSON for LLMs" because its core purpose is to represent the same structured data as JSON while achieving massive savings on token on usage. Few key points about toon: * Token Optimization: TOON is a data serialization format designed to significantly reduce the token count of structured data passed to and generated by Large Language Models (LLMs * Token Savings: It typically achieves 30-60% fewer tokens compared to formatted JSON, directly lowering inference costs and improving speed. * Minimalist Syntax: It minimizes the use of repetitive punctuation (quotes, braces, commas) that consume tokens in formats like JSON. * Tabular Arrays: Its key innovation is handling uniform arrays of objects by declaring keys once (like a CSV header) and listing data in comma-separated rows, eliminating the need to repeat key names for every object. * Indentation-Based Nesting: It uses indentation (similar to YAML) to represent nested object structures. * Primary Use Case: Optimal for tasks where token efficiency is critical, such as in Retrieval-Augmented Generation (RAG) context injection and instructing LLMs to output clean, structured data. https://lnkd.in/gNRMgkp6

10

📯With 𝗦𝗽𝗿𝗶𝗻𝗴 𝗙𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸 𝟳.𝟬 set to release by end of 2025, one of the major excepted feature is improved native image support for 𝗚𝗿𝗮𝗮𝗹𝗩𝗠. While Spring started supporting native images from version 6.x via Spring AOT hints, version 7.0 pushes this further by deepening support for:  • Spring AOT contributions, making native image builds more robust and reliable.  • Removing deprecated methods and aligning more cleanly with GraalVM expectations. The key to choosing 𝗚𝗿𝗮𝗮𝗹𝗩𝗠 𝗼𝘃𝗲𝗿 𝗢𝗽𝗲𝗻𝗝𝗗𝗞 however depends on understanding 𝘄𝗼𝗿𝗸𝗹𝗼𝗮𝗱, 𝘂𝘀𝗲 𝗰𝗮𝘀𝗲 𝗮𝗻𝗱 𝘁𝗿𝗮𝗱𝗲-𝗼𝗳𝗳𝘀 - which I have documented in this post to best of my knowledge. ⚖️ Key Trade-offs ✅ Choose GraalVM Native if:  • Cold start speed is business-critical (serverless, on-demand scaling).  • Memory footprint must be minimized (edge devices, containers with hard limits).  • You’re building batch/CLI tools that should run instantly.  • You can tolerate longer build times in CI/CD. ✅ Choose JVM (HotSpot/OpenJDK) if:  • App runs 24/7 and throughput > startup time (e.g., APIs, Kafka, banking apps).  • Heavy runtime reflection, proxies, or dynamic class loading.  • Developer productivity (fast builds, hot reload, live debugging) is a priority.  • You need mature monitoring, profiling, and ops tooling. Hope this helps! Which are the use cases where you think GraalVM would prove better over OpenJDK? #OpenJDK #GraalVM #AOT #SpringFramework #Java #SoftwareDevelopment #JavaEngineering

5