CNCF Chief: AI Inference Will Drive Increased Cloud Native Software Consumption
The executive director of the Cloud Native Computing Foundation (CNCF) is predicting that the rise of artificial intelligence (AI) inference models is about to spur a massive increase in the consumption of cloud-native software.
Speaking at an event in New York to commemorate the 10th anniversary of the founding of the consortium, Jonathan Bryce told attendees that he expects the CNCF to become the center of gravity for inference as more AI models are deployed on Kubernetes clusters. “This is where we have a huge opportunity,” he said.
Additionally, as the performance of open foundational AI models continues to improve, many organizations will be using them to create smaller AI models that are trained to more accurately automate a narrower range of tasks, noted Bryce. Most of those models will be deployed on inference engines running on Kubernetes clusters, he noted.
The assumption is that AI inference models and the agents deployed with them will be invoking more cloud-native software running as a backend service. As a result, there will be less emphasis placed on the need to develop graphical interfaces for each open source project being advanced under the auspices of the CNCF. Instead, AI agents will eventually become the primary means for invoking what will become a portfolio of headless backend services.
Orchestrating the consumption of those services will also require new tools and systems, said Bryce. Existing code review workflows and continuous integration/continuous delivery (CI/CD) platforms are not going to be able to keep up with the rate of change that AI coding tools introduce, he added.
There will, however, still be a need for the expertise of a human software engineer, especially as the amount of sloppy code being created using AI coding tools continues to increase, noted Bryce. However, the role of the software engineer will evolve as more organizations with the advent of AI embrace platform engineering as the primary methodology for building and deploying applications at scale, he added.
The CNCF currently oversees more than 240 projects. It’s not clear to what degree all those projects will remain relevant in the age of AI. In many instances, AI agents that have already been trained using open source software are creating bespoke versions of open source software to provide a similar capability.
In the meantime, maintainers of open source projects are already being overwhelmed by developers who are using AI coding tools to contribute more code than ever. Referred to as “Eternal September” in a nod to freshmen that show up on college campuses every fall knowing little to nothing about any process, this uptick in contributions is creating a need than ever to expand the ranks of the maintainers of open source software projects, noted Bryce.
It’s too early to determine with absolute certainty what impact AI will have on the open source community writ large. The one thing that is clear is that there is now more code of varying quality moving through pipelines that will, for better or worse, wind up being incorporated into open source software. Whether more is better, however, remains to be seen.
