TL;DR
The software industry has always been defined by waves of change. The lessons from the dot-com era—adaptability, sound engineering, and a relentless focus on business value—are as relevant now as ever. AI is the latest wave, not the last. If we engineer for change and keep our eyes on outcomes, we will always have a place at the table, no matter how the technology shifts.
Introduction: The Rhythm of Change
Software engineering has always meant reinvention. I remember when “web developer” was not a job title at all. Then, almost overnight, it became one of the most sought-after roles in technology – I know, I was one The dot-com boom brought a wave of opportunity, plenty of confusion, and ultimately a hard reset for the whole industry.
Now, with AI, the tune may be different but the rhythm is unmistakably familiar. There is the same excitement, some apprehension, and the usual talk about jobs disappearing. Yet the one thing that has not changed is the need to deliver business value. We do not engineer for its own sake. Businesses pay for technology because it delivers results, not because it is novel. If we ever get too comfortable, we risk stagnation, and that is something no organisation can afford.
Familiar Patterns, Higher Stakes
In 1996, building an online eCommerce system meant figuring things out as you went. Sometimes it felt like magic. Sometimes it was just smoke and mirrors. The rush for market share brought breakthroughs, but also spectacular failures.
Every wave of technology reshapes the roles around it. The dot-com era did not just introduce new tools, it created entirely new jobs. Web developer, UX designer, and later new disciplines like DevOps and cloud computing, none of these existed before the web. At the same time, roles such as assembly programmer, COBOL programmer, and even the whole notion of the analyst programmer, have largely faded away. We do not mourn their loss. We have simply evolved to something new.
Each wave brings its own rush of investment and a scramble for skills. There was a huge demand for technology roles and plenty of money followed in the late 90s and early 2000s. Some realignment is always needed, but that does not mean all software engineering roles become obsolete. Remember when Visual Basic was going to allow anyone to build applications without developers? It did not eliminate the need for software engineers. AI will be no different. The engineers who continue to adapt and deliver business value will always find a place, even if the specifics change.
Looking at today’s landscape, the AI wave feels familiar. Hype cycles are full of companies promising the world but shipping prototypes. The difference this time around is that the cost of making things, and the cost of writing code, has collapsed. But still, what has not changed is the value of sound engineering – building the right thing right.
AI’s Real Shift: Recomposition, Not Replacement
AI in software engineering is not about replacing developers. It is about shifting where leverage happens. AI can write boilerplate, parse documentation, and surface patterns. It cannot replace the judgment needed to decide what to build, or why.
Every system, every project, must be justified by the value it brings. That was true when we rendered our first credit card form, and it is true now as we integrate AI.
Just as the early web democratised access but not meaning, AI now democratises output but not architecture. Engineering judgment, trade-off analysis, and long-term design still matter. In fact, AI increases the need for clarity and coherence.
If our platforms and practices are rigid, we break. If we expect change, design for it, and stay focused on outcomes, we do not just survive—we help our organisations lead.
Patterns and Risks: The Tools That Change the Game
Every wave brings new enablers. In the dot-com era, it was LAMP stacks and virtual hosting. Today, it is LLMs, orchestration layers, and new data primitives. Speed alone is never enough.
The risk is the illusion of mastery. Just as drag-and-drop tools could not create great user experiences, AI tools cannot create engineering excellence. They can accelerate both solutions and mistakes.
Culture Under Pressure: The Real Test
AI is not a replacement for engineers. It is a pressure test for engineering culture.
- Teams with disciplined review and strong practices will be augmented, not undermined.
- Teams built on speed over substance will see technical debt and chaos grow faster.
- Architectures that favour modularity, observability, and testability will only increase their advantage when using AI.
Regulation, Risk, and Responsible Scale
Regulation is catching up. In the late 90s, compliance lagged years behind. Today, responsible AI demands traceability, explainability, and governance from day one. Regulated industries like finance, health, and government cannot afford shortcuts.
From a global perspective, regulation is not uniform. The EU’s AI Act enforces strict transparency and risk management, while US regulation remains fragmented. Engineering leaders should design for the strictest environment they operate in.
Modern engineering leaders must build for auditability and trust. That means pipelines with built-in lineage, versioned models, and features that protect user privacy and address systemic bias.
From Hype to Practice: What Actually Works
Three rules for the current wave:
- Adopt with Intention
Use AI to remove toil, such as scaffolding, test generation, and migrations, but keep human oversight and understanding central. - Double Down on Fundamentals
AI is only as good as the codebase and architecture it inherits. Clean boundaries, explicit design, and enforced patterns are essential. - Measure Real Impact
Productivity is not lines of code. Use DORA (lead time, deploy frequency, failure rate, MTTR), SPACE (satisfaction, efficiency), and business KPIs (customer adoption, cost-to-serve).
If you are not measuring, you are not engineering.
Commercial Realities and the Talent Market
This wave is not driven by speculative capital, but by incumbents with deep pockets and established markets. The winners will not be the fastest builders, but those who can adapt their platforms, govern their data, and scale talent alongside their systems.
AI is changing the talent profile. Tomorrow’s engineers must combine development skill with architectural thinking and prompt literacy.
Effective teams combine prompt literacy, architectural thinking, and hands – on experience with new tooling. Leadership should treat AI capability as a core development track – run regular workshops, review usage patterns, and pair engineers on AI-augmented projects.
It is not just engineers who must adapt. Product managers, designers, and compliance teams are learning to work with AI, shifting focus to integration, governance, and user trust. The entire business evolves together.
Leadership must invest in upskilling, mentorship, and capability uplift, not just tool adoption.
How We Measure What Matters
The real test of AI in engineering is not output, but value.
- DORA: lead time, deployment frequency, change failure rate, MTTR
- SPACE: satisfaction, performance, activity, communication, efficiency
- Business KPIs: customer adoption, churn, cost to serve
Continuous measurement, not just output, shows whether teams are building the right things, not just more things.
The Long Arc: Building for What Lasts
For example, the need for adaptability and business value never goes away, but the disruptions and their nature are always changing:
| Need | Dot-Com Disruptions | AI Disruptions |
|---|---|---|
| Adaptability | Delivery channels, LAMP stack, open source, the rise of web as a platform | Platformisation, automation, AI-augmented workflows, compliance demands |
| Focus on business value | E-commerce, digital marketing, first generation SaaS, global reach | AI literacy, prompt engineering, new ways of measuring value, data as product |
| Engineering discipline | Surviving rapid growth and crashes, emergence of DevOps and cloud | New risks from automation, continuous learning, tighter governance and measurement |
After the dot-com crash, the survivors rebuilt with better systems, stronger habits, and a clearer sense of purpose. AI offers a similar invitation, but only for teams willing to measure, adapt, and invest in their own foundations.
Let’s not chase acceleration for its own sake. Build for resilience, transparency, and impact, because the hardest problems in software are not technical. They are human, and they always have been.
We are lucky to work in an industry that is constantly evolving, with access to the latest software and hardware. The landscape may look different with every wave, but our real job does not change: deliver business value, adapt to what comes next, and help our teams do the same. We have done it before. We are doing it now. And, with the right mindset, we will do it again.
If you have not yet audited your team’s AI readiness, start now. Run an AI readiness audit quarterly or after major releases, perhaps as part of your DevEx survey. Use the results to prioritise investment and track improvement. Adaptation is not a one-off exercise; it is now part of the engineering discipline. Map out current workflows, review measurement practices, and invest in both technical and leadership upskilling. The next wave of value will be captured by those who measure, adapt, and build business value.
Leave a Reply