Blog Layout

Will AI-generated code be normalized in 2025?

Bill Doerrfeld | November 1, 2024

This story is in progress: I'm collecting input here.


Over the last two years, we've seen monumental promises around AI’s impact on developer productivity. Tools like GitHub Copilot are becoming ubiquitous, while OpenAI continues to push boundaries with groundbreaking features. Yet, the AI landscape is still a Wild West and must mature. We've witnessed the downsides of hallucination, questionable code qualitybloat & technical debt, copyright/IP restrictions, friction with data regulations, security issues with LLMs, confusion about what open-source AI is, and more. Not to mention the high energy and environmental costs of AI processing.


Valuable use cases with AI are still rare. Some analysts say the genAI bubble has popped, and some tech leaders are pulling the plug on major AI investments. There is also competing research on the actual productivity gains of using AI agents for code development. For consumer-facing AI-driven apps, "Disclaimer: results may not be accurate" is the saying of this age.


Despite this, there are signs of AI-generated code improving. Engineering leads are increasingly embedding it in development workflows, and developers are beginning to expect AI assistance. 2025 may be a breakout year for AI-generated code to become more stable, secure, and normalized. What will be done to get there?


I'm putting together a feature to discuss both the potential for widespread AI adoption and the practical steps necessary for overcoming current limitations. By forecasting these advancements, I hope to predict whether AI-generated code will truly be part of mainstream development by next year.


The deadline for submission is Friday, 11/15/24, 5 pm PST.


I’m looking for commentary from respected individuals in the software development field, offering practical examples (no product pitches). For efficiency's sake, I'll only be accepting input here on the below. I cannot guarantee the publication of any or all responses. Setting a maximum 1k characters (~150 words) per response. Looking forward to reading your thoughts!


This is probably for InfoWorld but may run elsewhere.


Form: Will AI-generated code be normalized in 2025?
Bringing internal engineering gems into the light
By Bill Doerrfeld March 28, 2025
Some internal projects are too good to stay hidden. For DirectorPlus, Spotify's Pia Nilsson shares how to identify and externalize internal platforms — like they did with Backstage.
Why AI-generated code is not enough
By Bill Doerrfeld March 17, 2025
My latest feature for InfoWorld looks into the gaps surrounding AI-generated code and what leaders believe we need to improve the status quo.
Can LLMs accelerate requirements engineering?
By Bill Doerrfeld March 6, 2025
New research shows LLMs outperform humans in software requirements engineering, reducing weeks of work to seconds while improving completeness by 10.2%.
Infoworld feature on ebpf and container networking
By Bill Doerrfeld March 3, 2025
Containers are designed to be isolated—that's great for security but tricky for networking. My feature with InfoWorld explores how eBPF is evolving container networking.
Agentic AI DirectorPlus New Relic
By Bill Doerrfeld February 28, 2025
I recently synced with Camden Swita, Head of AI and ML Innovation at New Relic, to see how software engineers are implementing agentic AI in practice.
Study reveals growing technical debt in AI age
By Bill Doerrfeld February 19, 2025
The 2nd annual code quality report from GitClear found 10x more duplicated code than two years ago and fewer signs of code reuse than ever before.
Kubernetes usability InfoWorld pilot cockpit Doerrfeld
By Bill Doerrfeld February 10, 2025
My feature on InfoWorld explores the state of Kubernetes usability, highlighting various advancements across workload types, support for edge and AI, and new features like observability and security.
Carving out time for large-scale engineering chores
By Bill Doerrfeld January 31, 2025
It's up to leadership to help prioritize large-scale engineering updates that keep software running smoothly. Kent Wills, Director of Engineering at Yelp, provides insight on the latest DirectorPlus.
5 potential use cases for Arazzo
By Bill Doerrfeld January 30, 2025
Italian for “tapestry,” Arazzo is aptly named since it can be used to weave together sequences of API calls to illustrate a specific business pattern.
How genAI can aid Kubernetes troubleshooting
By Bill Doerrfeld December 23, 2024
I caught up with CTO Itiel Shwartz about leveraging finely tuned models like Komodor’s KlaudiaAI agent for ultra-specific DevOps challenges, like Kubernetes error diagnosis and remediation challenges.
More Posts
Share by: