AI… The speed of things to come

So… I tried Copilot, and guess what? I liked it.

I also tried Gemini, and liked it too.

In this particular context, I was playing around with some projects in Python, a language I only “quasi-knew,” but was certainly comfortable enough to navigate around.

Then, all of a sudden, a new project came along: to take a branch of a solution built on the Java stack and implement some new features before checking everything in for a production release.

Having a background in mostly C# .NET, I could certainly grasp Java, but I was far from up to date with the latest libraries or components. Normally, this would have made me quite slow at being productive in Java. What surprised me most was that, with a little (okay, a lot) of support from my friend Copilot, I was able to create a branch, implement all the new features, and push a merge request — all within 4 days.

Upon completing this task, I was surprised… Only 4 days to do a sizable chunk of work and have it ready for production? Copilot accelerated my ability to navigate through the libraries, and I simply had to review the code and unit tests. I was able to speed through everything. I suspect that in the pre-Copilot days, a similar task would have taken me a solid 3–4 weeks, mainly because I would have been immersed in learning the Java libraries at the same time.

So what does this all mean? AI can help accomplish tasks much faster…

Faster IT, means organisations can be more responsive

In my very first programming job, releasing a product — or even a feature — involved writing a full spec (usually in the range of 20–100 pages long), getting approval, securing budgets, etc., and then moving to implementation. Back then, releasing projects usually took months.

Then came agile methodologies, along with more reusable frameworks for web and cloud development. The process was loosely broken up into defining behaviours that a product or service needed to satisfy, and projects could begin as soon as a backlog was set up. Development typically took weeks (or sprints).

Soon after that, more frameworks emerged, making it even easier to set up project scaffolding and reusable components. Suddenly, even Agile felt outdated. Development followed a much looser-Agile process, and everything sped up even more.

At the frontier of AI-enhanced development IDEs, we now face a new paradigm: development that is even faster! It would be easy to assume that this means we need fewer developers for the same output that we have today, a common fear often propagated in the news or on social media. But sticking to the status quo with AI-enhanced development (and far fewer developers) would greatly miss out on the opportunity for growth, growth, growth!

AI enhanced growth

With AI, the pace at which it takes build an IT solution can be massively accelerated. So what shall we do with this pace… Well… all of those innovative ideas we have in a backlog, waiting for an approval cos their ROI is unclear and the investment (pre AI) is too high. Lets do those!

The legacy systems that are slow and clunky, and nobody had too much interest in trying to rewrite or replace… Well, let’s start rewriting them.

Technical debt, bugs... let’s sort all of those out in a jiffy!

Modernisation… enhancing features… seasonal solutions… experiments… lets do them all.

The keyboard is not the bottleneck

Once upon a time, the phrase “typing is not the bottleneck” was coined in support of the practice of pair programming. But now, all of a sudden, typing really isn’t the bottleneck. If a competitor started on a project 3 years earlier and it seemed like they had raced too far ahead for your team or company to catch up — well, maybe not anymore!

Fast solution delivery might put the focus back on creative thinking, innovative processes, and a keen understanding of changing customer needs and demands to keep up.

Increasing innovation... and turning up the heat for competition

Living in a world of trial and error, if the cost of implementation is drastically reduced, organisations will be pressed to try more — more features, more seasonal ideas, bolder experiments to win market share, customer engagement, and growth.

Maybe… the death of SaaS

The advantages of using SaaS offerings to date have largely been because the IT costs to develop like-for-like services were too high and distracting from a company’s core business. But what happens when this cost is greatly reduced? Can companies dive back into the world of custom development for the systems they would otherwise rely on SaaS for?

To compete with the lowered entry barriers to developing in-house solutions, will SaaS companies be forced to dramatically reduce their license fees?

Custom solutions could be making a revival — uniquely tailor-made to how companies want to work, with their own unique processes and characteristics. This would be in contrast to having to fit their workflows around those enforced by generic, standardised SaaS solutions.

The profile role of developers

There could be a shift from deep expertise in specific areas to an era of generalists. Instead of focusing solely on deep expertise in things like UI engineering, the emphasis might shift toward generalists who understand the full stack and can intuitively switch between programming language families. This way, they can take advantage of AI doing the heavy lifting, rather than relying on experts for deep technical knowledge.

Closing

Software and solution development has exponentially become faster over the past 20 years. AI technologies now offers the landscape to move at an even faster pace, so much so, the cause and effect from having an idea to having a solution in production will evolve the landscape on how companies make decisions and processes they adopt to run and grow their businesses and compete with one another.

It is often reported that AI will drastically take away many jobs, but this would only be true if we want to maintain the current output of organisations. A more realistic landscape for the future will be a dramatic change in pace and using AI to do more, more and even more, in seemingly fractional times as pre AI times.