Posts

Building AI agents with LangChain, or any agentic framework, starts innocently enough. You define a tool, give the agent access, perhaps wire in some prompts, and watch as it begins to reason and act. But then the complexity creeps in. Quickly, one realizes this is not software development in the classical sense. The metaphors of functions, classes, and control flow, which served faithfully for decades, begin to feel like the wrong map for this terrain. ...

Since Lovable Launched its agent, in my opinion, is the turning point towards consistent value creation by Agentic AI. For prototyping ideas and merging product design and software in one entity is where it shines, but what about brown field scenarios in software development? When people talk about agentic AI, they tend to imagine a kind of digital colleague—capable, tireless, and independent—capable of running through the backlog while the rest of the team enjoys a leisurely coffee. ...

We’ve all been there. The quarterly board meeting where “digital transformation” gets thrown around like confetti, followed by whispers of massive IT overhauls and million-dollar modernization projects. But here’s the thing—application modernization doesn’t have to be a moonshot. Start with Business Pain, Not Technology Hype Before you even think about containers, microservices, or whatever’s trending on Hacker News this week, ask yourself: what’s actually hurting the business? Technology exists to serve your organization, not the other way around. Yet too often, I see teams diving headfirst into architectural rabbit holes without understanding the business problems they’re solving. Your CTO might be excited about Kubernetes, but if your real issue is time-to-market for new features, maybe the answer isn’t orchestrating a thousand containers. ...

What effectively changes with software development in the AI era, is that you are going to program with natural language instead of machine language. We made machine language somewhat easy to understand by humans through time (sixties until now-1) and is mandatory for people to understand and use for mundane tasks and software implementations. The understanding requirement part of mundane stuff will disappear, so you spend more time on the creative process instead of the technical process. This will make the (global) economy scale. ...

In the agile universe, we constantly translate business needs into actionable work. Traditionally, this process unfolds through a clear hierarchy: Epics break down into Spikes, Spikes further unfold into Stories and (Sub-)Tasks. Each step refines abstraction into tangible work packages, providing engineering teams clarity and structure. Yet, solely relying on Epics and Stories to articulate business intent carries inherent limitations, often leaving gaps between technical teams and stakeholders. Epics and Stories naturally embody a detailed narrative—a roadmap defined primarily from a “ground-up” perspective, detailing the “what” and “how.” While thorough, this bottom-up approach can obscure the broader purpose, causing stakeholders to feel disconnected amidst technical detail, uncertain about the ultimate business value. ...

The AI revolution, at the time of writing, didn’t arrive like an asteroid crept in like weather. Gradual, persistent, and deeply altering. We woke up one morning and it was here, everywhere: auto completing emails, filtering resumes, writing poetry, grading code, recommending products, forecasting churn. In boardrooms and bedrooms, AI became the silent co-pilot. The buzz was about intelligence. Artificial, general, narrow, super. Debates flared about sentience, risk, regulation. But underneath the hype and hand wringing, a quieter question lingered: What actual value has AI brought us? ...

We’ve been sold a lot of reinvention in the past decade. Cloud-native architectures, Kubernetes-first everything, throw away the old and start clean. But reality is stubborn: enterprises have decades of investment in legacy systems, sprawling VM fleets, bespoke configurations—and yet they’re under pressure to “modernize” yesterday. The problem isn’t just technical. It’s economic. You can’t simply deprecate your previous investments, nor should you. What you can do is rethink your build artifacts. ...

Most companies don’t aspire to build perfect DevOps or cloud engineering teams. They just want the problems solved, quickly, safely, and without hiring dozens of specialists. That’s where Internal Developer Platforms (IDPs) come in: a productized way to give developers self-service deployments, secure environments, and golden paths to production without reinventing the wheel. But IDPs aren’t for everyone. If you’re Netflix, Shopify, or Airbnb, companies where platform engineering is a strategic differentiator, you should architect and manage your own stack. But if you’re a growing SaaS, a retailer modernizing online experiences, or a financial institution moving to the cloud, platform engineering isn’t your business. Shipping features is. ...

Most teams treat sovereign cloud like a logistics problem. Spin up a local region. Wrap it in compliance. Copy the infra. Move the data. Done. But this isn’t about copying and pasting your stack into a new jurisdiction. It’s about accepting a new reality: the cloud is no longer neutral ground. A new law, a new sanction, or a new interpretation of “access” can ripple through your architecture overnight. Sovereignty means, besides control, knowing where your data lives, who can touch it, and what happens if the situation changes tomorrow. ...

There’s a metaphor I keep coming back to when I think about modern cloud development: the vending machine. Think about it. When you walk up to a vending machine, you don’t wonder how it was stocked or who maintains it. You just press a button and get what you need. Fast, predictable, and (usually ;-)) reliable. That’s how the cloud should work. But for most teams, building in the cloud still feels like assembling the vending machine from parts every time you want a snack. You’re setting up infrastructure, configuring pipelines, stitching together services, and navigating endless layers of DevOps complexity. Even with great platforms like AWS, Azure, or GCP, the work behind the scenes can be overwhelming—especially when velocity matters. ...