If You’re a Software Engineer, You’re Probably Using AI

Apr 05, 2026 Features / Columnists, News

---

By Dr. Karen Abrams, MBA, AA

(Kaieteur News) – If you are a software engineer in 2026, you are probably using artificial intelligence and using it a lot. Deeply, daily, and increasingly as part of the real work of building and shipping software. The 2025 Stack Overflow Developer Survey found that 84% of respondents were already using or planning to use AI tools in their development process, and among professional developers, just over half said they used them daily. At the same time, 66% cited their biggest frustration as AI solutions being “almost right, but not quite,” while 45% said debugging AI-generated code could be more time-consuming than writing it themselves.

That was 2025. Since then, Anthropic, OpenAI, Lovable, Replit, and others have significantly improved their platforms. Tools such as Codex and Claude Code have accelerated the pace of development dramatically, and while debugging and quality control still matter, many of the earlier pain points have been reduced as the tools have become more capable, more context-aware, and more reliable.

Those improvements matter enormously for Guyana.

We are living through a period of rapid oil-fueled development, aggressive digitization, and rising demand for ‘technical talent’. Companies, public agencies, schools, clinics, banks, and utilities are all under pressure to modernize. New systems are being procured, old systems are being patched, and digital promises are being made to the public with increasing frequency. In that environment, AI is already part of the toolkit and in many cases, it is becoming the toolkit.

A striking consensus has emerged among influential voices in technology. Anthropic co-founder, Dario Amodei has predicted that AI will write the overwhelming majority of code. Former Google CEO, Eric Schmidt has argued that the most valuable programmers are increasingly those who write specifications and evaluation criteria rather than every line by hand. Moonshot’s Peter Diamandis and Stability AI former CEO, Emad Mostaque have each warned more bluntly that the number of human coders needed may fall sharply, while Silicon Valley venture capitalist, Ben Horowitz, more grounded in the realities of building companies, has focused on what this means for execution, judgment, and competitive advantage. They differ on pace and tone; some celebratory, others cautionary, but the underlying message is the same, the software profession is being reorganized in real time.

Crucially, Amodei also made the point that the human still has to define conditions, overall design, integration with existing code, and whether a system is secure or insecure. (Council on Foreign Relations) The software engineer of today is becoming, in part, a spec writer, reviewer, debugger, evaluator, orchestrator, and decision-maker. That distinction is the heart of the matter, and too often it gets lost in public conversation.

My own view, shaped by more than two decades of building and managing development teams, is that there will still be an enormous amount of work for skilled engineers for quite some time. Every serious organization I have known had a roadmap full of features, fixes, integrations, and overdue strategic upgrades extending far into the future. AI does not eliminate that backlog. Innovative leaders will grow more ambitious; those who see AI merely as a cost-cutting tool will lose to the innovators. When the cost of producing software falls, demand for software tends to rise.

But that does not mean all software jobs are equally safe, or that all software teams are equally prepared.

A country like Guyana should be especially careful not to confuse faster coding with better systems. The risk is not merely that some engineer somewhere may become less relevant. The real risk is that organizations, intoxicated by speed and dazzled by demos, will underinvest in seasoned oversight. They will let inexperienced teams build critical systems too quickly. They will accept software that appears polished on the surface but has weak architecture underneath. They will deploy systems without sufficient testing, validation, backup planning, user training, cybersecurity review, or long-term maintenance discipline. Then the predictable will happen; systems will crash, records will go missing, customers will be locked out, payments will fail, staff will improvise with paper and WhatsApp, and public confidence in digitization itself will be damaged.  This does not need to happen.

Amodei has since gone beyond code, warning that AI could wipe out half of all entry-level white-collar jobs within five years, including roles across technology, finance, law, and consulting. The warning should force us to confront a serious question; which parts of professional work are being automated, and which parts become more valuable in an AI-rich world?  Guyana should be asking that question now, not after the fact. The so-called untouchable professions are not untouchable. Law will be affected. Medicine will be affected. Engineering will be affected. Accounting will be affected.  That is not a reason for despair. It is a reason for preparation.

We should be preparing children and young adults not just to acquire credentials, but to develop durable advantage. That means mastering foundational literacy and numeracy. It means reading deeply and often. It means becoming genuinely strong in a domain, whether engineering, health, law, education, finance, agriculture, energy, logistics, or design, and then, on top of that domain knowledge, mastering AI in a serious way. Not merely knowing how to prompt ChatGPT for a quick answer. It means becoming a life-long learner.  The future will belong to people who combine domain expertise, judgment, curiosity, communication, and real AI fluency.

My advice to Guyana is straightforward. Use it to speed up development, reduce repetitive work, prototype ideas, and empower local teams. But do not confuse acceleration with maturity. Put seasoned professionals over critical systems and keep humans in the loop. Demand architecture. Demand testing. Demand documentation. Demand accountability.

An engineer using AI is already more powerful than one who is not. The deeper question is whether our institutions are wise enough to use that power well.

---

• AI, Engineer, Karen Abrams, Software Engineer

---