Using AI in software development: A glimpse into 2024

Joonas Palomäki | 8 September, 2023

As the summer  is nearing it’s end soon here I’m in to my  crystal ball to see what will the year 2024 bring us? AI here, AI there – news tell us of the advancements, hopes and fears related to the artificial intelligence. How will the industries take that? Software development, along with it’s integral parts of specifying and managing needs and ensuring quality, will exprience changes. While some may be quick to think that our future software will miraculously evolve overnight, reality might be different. The lords of Kobols still haunt the server rooms of major banks, but they represent the perseverance of time-tested systems. Yet, it’s undeniable that change is brewing in the air.

In a world characterized by an ever-increasing interplay between humans and technology, one might wonder: Where does artificial intelligence fit into the software development landscape of 2024? Amidst the cacophony of voices—some waiting for AI to solve most if not all our problems and others prophesizing an impending AI winter—the truth lies somewhere in between. For businesses and tech experts charting the future, the challenge isn’t about hopping onto the AI bandwagon; it’s about discerning fact from fiction and concentrating on what matters.

So, what can we realistically expect from AI in software development? Let’s dive a bit deeper.

The current AI landscape: An appraisal of today’s software development

The tech world has never been a stranger to change and the past few years have been a testament to this. The introduction of ChatGPT and the meteoric rise of large language models (LLMs) have etched artificial intelligence onto the world’s canvas. While many shout about the benefits these LLMs bring, they represent just the tip of the iceberg.

The rise of GPT and the AI popularity surge

Since ChatGPT burst onto the scene, the media’s spotlight has been firmly fixed on AI, triggering widespread interest and curiosity. Beyond its newfound fame, the real significance of ChatGPT and similar tools lies in their practical applications. Many professionals and enterprises have experienced firsthand the transformative power of LLMs. But let’s not be swayed by sporadic tales of success. A broader look reveals that while many sectors have begun embracing LLMs, they remain a novelty for many enterprises. Their potential, though vast, remains largely untapped.

The advent of AI-infused tools

For developers, tools like GitHub’s Copilot have transformed the coding landscape. These AI-assisted platforms streamline the coding process, offering code suggestions, detecting bugs (and adding some sometimes!), and improving overall productivity. But these advancements represent mere ripples in the vast ocean of software development.

Currently, the overarching narrative isn’t about a seismic industry shift due to AI, but rather the early tremors signaling an imminent revolution. There’s an undeniable undercurrent of change, with more companies beginning to experiment with AI tools to streamline tasks, improve efficiency, and innovate in ways previously considered unfeasible. Many companies see that this is the way to go, but do not have the exact means.

The road ahead: envisioning AI’s evolution in software development

In the realm of technology, if the past is our tutor, the future is our guidepost. To fully appreciate where AI in software development might lead us, we need to understand where it is good and where it is not.

Co-working with AI: A fundamental shift

AI’s integration into software tools isn’t merely a fleeting trend; it represents a paradigm shift. While various software products have boasted “AI capabilities” for years, the unveiling of powerful LLMs were the thing that started a new era. Now, instead of AI being a mere sophisticated algorithm, it’s evolving into a foundational layer — an “Intelligent Layer.”

Historically, every leap in programming languages or frameworks led to marginal efficiency gains. Modern developers, equipped with contemporary tools, certainly outpace their 1980s counterparts coding in C. Yet, AI’s influence is going to dwarf these previous technological shifts. For the first time, machines can tackle challenges previously reserved for human cognition.

The emergence of hybrid systems

However, the notion that AI will singularly dominate the software ecosystem is a misconception. The true strength of AI emerges in collaboration with human oversight. An autonomous AI might stumble when faced with intricate tasks, but in tandem with human expertise, it becomes a force multiplier.

The future, thus, lies in hybrid systems. Here, AI handles data-heavy tasks, pattern recognition, and repetitive functions, while humans guide, interpret, and provide the essential “human touch” to solutions. This symbiotic relationship will redefine productivity, allowing organizations to either accomplish more with the same resources or achieve similar outcomes with reduced workforce overhead.

The renaissance of written specifications

As we navigate the burgeoning realm of AI-powered software development, an unlikely hero emerges from the shadows: the written word. Its resurgence is not merely a nod to tradition but an absolute necessity in this evolving digital landscape. But why, in a world increasingly leaning towards visual and auditory interfaces, is the written specification regaining such prominence?

  1. AI, namely large language models (LLMs), primarily operates using language. Although these systems have the capacity to interpret images and other data forms, language remains their most potent medium when interfacing with humans, especially when defining the behavior of IT systems. This linguistic prowess holds the key to unlocking the full potential of AI in software development.
  2. Complex IT projects are no longer solely the domain of human developers liaising with clients. The dynamics have shifted, introducing a new, intelligent collaborator: AI. This three-way collaboration between humans that need the software, other ones creating it, and AI assisting throughout, necessitates a shared language. Written specifications, comprehensive and up-to-date, become the bedrock of this tripartite relationship.
  3. The intricate nature of modern software projects amplifies the importance of these specifications. As projects swell in complexity, clear, concise, and current documentation isn’t just desirable; it’s imperative. AI excels in maintaining and refining these documents, ensuring that they are not only detailed enough but are also written in language usable by Business and Development alike. Organizations that adeptly leverage this AI capability can anticipate more streamlined project workflows and more consistent deliverables.
  4. Such clarity in documentation enables profound shifts in software development processes. The old challenges, the time-consuming bottlenecks, begin to diminish. Development cycles are shortened, feedback loops are tightened, and the end products are more aligned with operational requirements – requirements that are finally meaningful. The human role changes from being suffocated in detail to one of oversight and strategic direction. Without a lucid view of the specifications, however, this transformation is not going to happen.
  5. Yet, not all organizations are going to ride this wave. Many, tied to legacy systems and fragmented documentation practices, face an uphill battle. Scattered, unorganized specifications become a millstone, hampering their transition into this new AI era. While AI can assist in consolidating and organizing this fragmented information, the initial step of transitioning from chaos to clarity remains a daunting task. Only after this hurdle is surmounted can these organizations truly harness the power of AI in their software development endeavors.

Navigating the Transition: The Challenges Ahead

Adapting to an AI-centric model is not without its share of obstacles. Traditional organizations, with their disjointed documentation habits, confront steep learning curves. While AI offers the promise of streamlining and decluttering, the transition requires patience, effort, and a commitment to adapt. The rewards, however, are worth the journey: a future where AI and humans harmoniously co-create, guided by the clarity of impeccably written specifications.

The inevitable transformation

Regardless of the challenges, AI’s incorporation into software development is inevitable. Organizations that proactively adapt and evolve will enjoy competitive advantages, while those resistant to change find themselves struggling to keep pace. The productivity boost of AI is just too significant to dismiss.

In summation, the horizon of 2024 and beyond paints a picture of coalescence, where AI and human expertise meld to usher in an era of unparalleled software development prowess. Isn’t it just exciting?


(Pictures by Midjourney with the help of Dall-E, our cool AI artist friends)