#73-AI: The Future Of Systems Development

I conclude this series my summarizing my current views on AI and it's increasing role in the development of business automation.

Transcript

AI - The Future of Systems Development

The future of AI in systems development is going to be less about replacing developers outright and more about changing the shape of their work—the way IDEs replaced simple text editors.

Here’s how I see it unfolding over the next 5 – 10 years:

1.  Development Will Shift from “Writing” to “Orchestrating”

Today developers spend much of their time manually writing logic, debugging, and looking up API docs. In the future developers will focus more on describing desired outcomes in natural language, creating constraints, and integrating pre-built AI modules.

2. Testing Will Become Continuous and Autonomous

AI can run code in thousands of scenarios, automatically generating regression tests as the code changes. Bugs will often be found before a human even runs the program.

3. Code Will Become More Self-Healing

AI will not only detect runtime errors but suggest and apply fixes in production (with developer approval, at least at first).

4. Legacy Systems Will Get a Second Life

Millions of lines of old COBOL, C, and FORTRAN code will be read, understood, and refactored by AI.

5. The “AI Stack” Will Become a New Specialty

We’ll see AIOps emerge as a role — people who specialize in selecting, fine-tuning, and governing AI components inside the development lifecycle.

6. The Role of Human Creativity Will Become More Important

AI will excel at pattern completion but won’t (yet) have the same ability to invent disruptive new paradigms. Humans will still lead in deciding what is valuable and what the software should do — especially where the “user experience” is involved.

In short AI will be like having a tireless junior developer who never sleeps, never gets tired of repetitive work, and can be trusted with rote tasks.
 
One final note in conclusion:

The following is my own personal opinion, and should not be interpreted to mean anything in terms of PMI as an institution:

AI is coming on fast, and many people are concerned about what it may mean for the future.  I don’t claim to have the answers, but I do believe that anything created by AI or using AI should be required to be labeled as such.  I think consumers should be able to make informed decisions about where they will spend their dollars and what they will support going forward.

To that end, I’ll disclose that the first draft of this particular episode was written by Chat GPT and then extensively edited by me.  I thought that would be a good illustrative way to conclude this sequence: by asking AI about AI… but I wanted to be transparent about it.  

I don’t plan on using AI for that in the future. :)