Change is the Only Constant, and I've Never Seen It Happen in Such a Short Span
David Balzan·February 28, 2026·3 min read
About 30 years ago, I dove headfirst into software development. It started innocently enough with QBasic on an old DOS machine, then Pascal for structured thinking, C++ for object-oriented power, and even dipping into assembly and machine language to understand what was really happening under the hood. Back then, the Software Development Life Cycle (SDLC) was dominated by Waterfall—linear, predictable, and often painfully slow. Over time, I watched (and lived through) the rise of Scrum and agile practices, which promised faster iteration and closer collaboration.
I've worn a lot of hats since: developer, architect, DevOps engineer, implementor, DBA, QA tester. I've worked across industries, from startups to enterprises, and thrived in small, capable, multi-talented teams where people could switch roles fluidly and deliver real value without layers of bureaucracy. Those teams felt magical—focused, product-obsessed, and efficient.
Fast-forward to today, and we're in the thick of the AI revolution. When these tools first appeared a few years back, I was fascinated—curious, even skeptical. Could something really understand code the way a seasoned developer does? I wanted to try it for myself. My initial attempts were frustrating: suggestions that missed context, hallucinations in logic, outputs that required heavy editing just to run. It felt like babysitting a talented but unreliable junior. But I kept at it. Subsequent attempts yielded better results as the models improved and I learned to guide them more effectively—better prompts, clearer context, iterative refinement. Nowadays, I can confidently roll out a mid-complexity website in just a few days: from requirements to design, implementation, testing, and deployment, with AI handling large swaths of the boilerplate, refactoring, debugging, and even architectural decisions.
The progress over the last year has been astonishing. What started as helpful code-completion assistants evolved into true agentic systems—AI that doesn't just suggest snippets but plans multi-step tasks, navigates entire codebases, executes commands, iterates on failures, runs tests, and commits changes autonomously. Tools moved from chat-based helpers to production-ready agents capable of shipping real features with minimal oversight. Reliability improved dramatically: fewer wild errors, better long-horizon reasoning, stronger handling of complex projects. Inference got faster and cheaper, multimodal understanding deepened, and the gap between experimental toys and everyday power tools closed almost overnight.
This acceleration has triggered visible shifts in tech companies. We've seen major layoffs explicitly tied to AI's potential—firms cutting significant portions of staff because smaller, highly talented teams armed with these tools can achieve far more. Leaders talk openly about intelligence tools reshaping what it means to build and run a company: flatter structures, less coordination overhead, a return to lean operations. It feels like the industry is rediscovering the power of small, agile, product-focused teams—the exact kind I grew up with and loved.
Those teams of a handful of multi-talented people, heavily invested in the product, seem ideally sized for flourishing with today's AI tools. Instead of large groups bogged down in handoffs and meetings, a tight crew can leverage agents for the heavy lifting—staying laser-focused on strategy, user needs, innovation, and the human elements that AI can't replicate.
Things are improving so fast that in just a few months, these tools could become dramatically more reliable—longer autonomous runs, even fewer hallucinations, seamless integration across workflows. What does that mean for our tech jobs? We have to change. We have to adapt and transform ourselves. The developers who thrive won't be the ones churning out lines of code fastest; they'll be the ones who orchestrate AI agents, engineer effective contexts and prompts, design robust systems, ensure alignment on ethics and security, and keep the product vision sharp.
The recipe for success that made those small agile teams shine—focus, versatility, product obsession, rapid feedback—feels more relevant than ever. In this new era, it's not about replacing humans; it's about amplifying the best of what we do. The constant is change, and right now, it's accelerating at a rate I've never witnessed. The question isn't whether we'll adapt—it's how boldly we embrace it to build even better things.
What about you? How has AI shifted your own work over the past year? I'd love to hear your reflections in the comments.