Ironclad Journal icon Ironclad BLOG

Playing a New Game: Developing With AI

July 13, 2023 4 min read
abstract image of woman with VR headset on black background

If you lead development teams, and are considering using AI, you have to understand it for what it is: a fundamental paradigm shift. We are living through a revolution. AI is so powerful and capable, and so distinct from everything that has come before, that it forces us to engage in new ways.

Remember the mobile explosion of the early 2000s? Software developers had to adapt to a world in which smartphones and mobile devices had suddenly become ubiquitous. Then, as now, we were faced with a dramatically changed landscape of user behavior and technology capability. We had to rethink where, when, and how people would interact with our solutions, adjusting our approach to very different interfaces and use cases. It forced profound changes in how we think about development.

AI will be even more disruptive… and just as permanent. This is no wave, no trend. Even in these early days, it is clear that AI must change how we operate, collaborate, and innovate. This is the new reality, and the sooner you come to terms with it the sooner you can maximize your solutions and your people.

So we need to change… but how?  What do you need to understand about AI to be effective, and how should you shift your approach to leading and guiding your team?

Rethink your development priorities.

The first and most core dimension of change: AI changes what we should be working on. It completely changes what is possible and practical. You need to bring a completely new lens to how you set priorities.

In many cases, things that were hard or outright impossible are now feasible. In some cases, these previously immense challenges have become almost easy. This not only changes what projects are possible, it also forces us to reconsider existing investments and practices.

Think about how much development time and energy is spent in designing search capability within software. Now, with AI’s game changing ability to interpret natural language text, this is a vastly simpler challenge. It can unlock meaning and patterns within large amounts of unstructured text in much more sophisticated ways than has ever been possible before. These chatbots are not only less resource-intensive to develop, freeing development teams from the burden of designing search interfaces, but also potentially more compelling and “human-like.”

To give another example, say you want to build an app that reacts intelligently to image content. That was possible with previous generations of machine learning, but it was always difficult to code and calibrate. And to reverse the process – to create images based on user input – was much, much harder. Now, it is dramatically easier to interpret and generate images by tapping the power of off-the-shelf generative AI models. Something that might have taken the focused effort of multiple people over weeks is now possible to achieve in a small fraction of the time.

This means that AI should change how you set your team’s development agenda, allocate resources, and plan out your work.

Shift your skill set.

The mobile revolution changed the talents and skills needed on development teams. AI is already doing the same. This is a true “skill reset” moment!

Broadly speaking, you need two kinds of skills to develop with AI:

  1. The ability to develop new AI models that extract meaning and value from data inputs; and
  2. The ability to exploit, tune, and manipulate those models to create powerful user experiences and breakthrough insights.

Both matter, but with the broad availability of highly capable generative AI models, you can focus mostly on the second. The great challenge of recruiting and staffing up for the AI era right now is, of course, that effectively no one has these skills. Just as there were virtually no mobile developers prior to the mobile explosion, there are vanishingly few people with relevant AI experience. While this picture is rapidly changing, it will remain the case for the foreseeable future.

Learn how to “work in the dark.”

This leads directly to the other category of change that AI forces upon development teams… It requires them to work, collaborate, and share information in very new ways. You have to learn how to take a more open-ended and flexible approach to development.

Because no one really has prior knowledge of, or experience with, AI, everyone is in a learning mode. And because team leaders don’t have the luxury of waiting to use AI until their people come up to speed, the reality is that teams are learning as they go. This can be a difficult adjustment for leaders used to operating with more knowledge and expertise.

The other uncomfortable truth about AI is that it involves a high degree of unpredictability. Often, you cannot fully understand or anticipate the responses that AI will generate. These two factors together means the precision- and process-oriented mindset of many development teams simply won’t work for AI-centered development.

The focus can’t be on stamping out bugs, or moving projects through a careful series of development stages. Instead, you need a more fluid, collaborative, open approach. You could call this “working in the dark.” It combines learning, experimentation, and information-sharing. In this new environment, development leaders must throw out their assumptions and principles around design and innovation. Creating a culture of learning and open-ended exploration becomes so important.

A hard change… but one worth making

The bottom line: generative AI, with its breakthrough ability to recognize and act on natural language, is an immense milestone in software innovation. And it is already ushering in a new paradigm to which technology leaders must adapt.

The change is coming. You can either adapt and thrive, or resist and delay and risk falling way behind. And if you want to attract and retain the best talent, the time to shift your thinking and approach is now.

Want more content like this? Sign up for our monthly newsletter.

Book your live demo

Jason Li is Ironclad’s VP of Engineering. Before joining Ironclad, he was the VP of Engineering at SalesforceIQ and Grovo. He holds a B.S. in Electrical Engineering and Computer Science from U.C. Berkeley and an M.S. in Electrical Engineering from Stanford University.