Can AI and developers work in harmony?

As AI becomes more intelligent, it can code faster than developers. That puts developers in a quandary; we love to write code, but if AI will take that over, where does that leave us?

There was an interesting meme circulating recently, that for the life of me I can’t find again, where someone was stating that they love to create art and write and hate dishes and laundry. They were annoyed because AI was taking away their art and writing, so they had more time to do laundry and do the dishes!

Shouldn’t it be the other way around?

The exact same issue is starting to plague software development teams. As AI becomes more intelligent, it can code faster than developers. That puts developers in a quandary; we love to write code, but if AI will take that over, where does that leave us?

The concern isn’t that AI is replacing developers, but that AI will replace the parts of their jobs they love the most!  

So, how do we use artificial intelligence to develop embedded systems without losing the parts of our jobs that we love the most?

Can AI and developers work in harmony?

Recognizing AI’s Strengths and Weaknesses

AI excels in tasks that require pattern recognition, data analysis, and repetitiveness. Developers often find these tasks monotonous or time-consuming. In fact, we often find them exciting the first few times we do them, but after that, they become a chore!

For example, the first time I wrote a USART driver, I was ecstatic. It was an exciting problem that taught me how USART peripherals work, how to configure their registers, and how to implement appropriate error handling. It required some hard work and creativity to find an elegant and reusable solution.

Fast-forward 15-20 years, and I’ve written dozens of USART drivers. Do you think I get excited about writing USART drivers today? No! I’ve been there and done that, and I want to work on other challenges!

Instead, I can feed in my design patterns and examples and ask an AI to make a driver for the X microcontroller using the examples I’ve provided as a template. Twenty seconds later, I have a driver ready to go, but it probably would have taken me several hours to hand-code it!

By leveraging AI for these types of tasks, we can free up our time to focus on the creative and challenging aspects of software development that we truly enjoy—not the activities that we’ve done a thousand times and no longer find joy in.

Enhancing, Not Replacing, Human Creativity

AI can be a powerful tool to augment our capabilities rather than replace them. For example, AI-driven code generators can handle boilerplate code, allowing developers to focus on architecture, design, and problem-solving.

Embedded software developers often think of themselves as electrical or software engineers. Instead, perhaps we should consider ourselves embedded systems engineers!

I might be biased, given that my master’s degree is in systems engineering, but systems engineering is where I think the real value is, anyway! Systems engineers must architect and solve system-level integration issues, dive into the lowest level of a subsystem, and deliver an entire system.

If embedded teams shift their focus from the low-level details to the higher level, they can better utilize their creativity and problem-solving skills and let AI manage the tedious parts of coding. Humans are freed up to spend more time innovating and exploring new ideas!

At the end of the day, if there is some part of a coding project that you want to do, there wouldn’t be any reason you couldn’t! AI is there to remove the tedious activities and not take away the joy of development.

(There may come a time when trade-offs must be made to deliver projects. Company management will need to learn that maximizing profits needs to be balanced with employee satisfaction)

Collaboration Between Humans and AI

One of the most effective ways to use AI in embedded systems development is to view it as a collaborator. AI can offer a lot of value to a developer and team by:

• Providing suggestions for troubleshooting
• Optimizing code
• Generating test cases
• Resolving bugs
• Predicting potential issues

At the end of the day, the human touch is essential for nuanced decision-making, understanding complex requirements, and crafting elegant solutions. The symbiotic relationship between human intuition and AI efficiency can lead to higher quality and more sophisticated software. 

Taking your Next Steps

Ultimately, the goal is to maintain the joy and satisfaction of creating something unique and valuable. By offloading mundane tasks to AI, we can preserve the most fulfilling aspects of software development. Whether it’s the thrill of solving a complex problem, the creativity involved in designing a new system, or the satisfaction of seeing our code come to life in a real-world application, AI should be an enabler, not a detractor.

Developers can collaboratively use AI to improve their work experience. The majority of projects are delivered late and over budget. Understanding when and where to use AI can help teams accelerate their project deliveries, allowing them to do more with less while still doing what they enjoy at work.

If you aren’t using AI to develop your embedded systems today, you should be! Take some time this week to identify the mundane and repetitive tasks you perform.

Ask yourself, “Is there a way to offload this to AI?”

Jacob Beningo is an embedded software consultant who specializes in real-time, microcontroller-based systems. He actively promotes software best practices through numerous articles, blogs, and webinars on topics from software architecture design, embedded DevOps, and implementation techniques. Jacob has 20 years of experience in the field and holds three degrees including a Masters of Engineering from the University of Michigan.

Related Contents: