By 2030, software developers will be using AI to cut their workload 'in half'

Prepare for the HyperAssistant of the future, maybe

Half a decade hence, software development will be transformed by AI assistance, argue four academics from the University of Lugano in Switzerland.

In a preprint paper titled, "From Today’s Code to Tomorrow’s Symphony: The AI Transformation of Developers' Routine by 2030," researchers Matteo Ciniselli, Niccolò Puccinelli, Ketai Qiu, and Luca Di Grazia, describe how they foresee AI assistance evolving in the years to come, with an eye toward encouraging the work necessary to realize that vision.

Looking specifically at the implementation phase of the software development lifecycle, they propose an successor to AI code suggestion tools like GitHub Copilot, OpenAI ChatGPT, Google Gemini, and Tabnine called HyperAssistant.

HyperAssistant is imagined as a more capable version of today's automated programming support tools, with a remit that extends beyond source code creation. The proposed AI aide would support developers' mental health by recommending work breaks and suggesting activities. It would be better at bug and vulnerability detection, and at code optimization.

We emphasize AI as a complementary force, augmenting developers’ capabilities rather than replacing them

As envisioned, HyperAssistant would coordinate meetings with relevant team members, in whatever mode the developer prefers. And it would support the creation of new features and the development of new skills through learning guidance.

"We envision HyperAssistant, an augmented AI tool that offers comprehensive support to 2030 developers, addressing current limitations in mental health support, fault detection, code optimization, team interaction, and skill development," the authors explain.

"We emphasize AI as a complementary force, augmenting developers’ capabilities rather than replacing them, leading to the creation of sophisticated, reliable, and secure software solutions."

Generative AI is not quite there yet. When it proposes code, it can make mistakes, or hallucinations in marketing terminology. As the authors observe, current tools don't address the mental aspects of programming, don't optimize code very well, don't do much to facilitate team synergy, and fail to consider developer's unique skill sets or needs. What's more, developers may compound the limitations of AI by putting too much trust in its suggestions and failing to verify them.

The authors outline how a day in the life of a programmer might look in 2030, if the anticipated HyperAssistant ever materializes as imagined.

"Ashley, the developer in 2030, arrives in the office and immediately notices that some code has changed since yesterday," they describe in their paper. "However, thanks to HyperAssistant, a concise summary is presented to her, highlighting only the pertinent edits. With this efficiency, she swiftly comprehends the updates and is ready to begin her tasks."

"As she starts coding, an intelligent bug detection system notifies her of an error she inadvertently introduced. The system not only reports the bug, but also suggests potential fixes, streamlining the debugging process."

It also presents a notification about how the code differs from what's described in the corresponding javadoc comments, with an alignment recommendation pulled from code written by a senior developer at the same operation.

HyperAssistant even goes so far as to schedule a meeting between Ashley and the senior developer, augmented by relevant documentation. And along the way the AI is watching for typos or other metrics that might indicate the need to take a break. With such assistance, the authors suggest, Ashley will require only half a day in 2030 to do what took her a full day in 2024.

Considering that by 2040 other researchers have suggested most code will be written by machines, HyperAssistant doesn't seem like that much of a stretch.

The authors argue that AI has the potential to make software development more fulfilling and productive. Perhaps so, as long as it's HyperAssistant and not HyperManager. ®

More about

TIP US OFF

Send us news


Other stories you might like