The new wave of AI is revolutionizing software development, going beyond just code generation. However, applying generative AI within cloud-native workflows involves more than automating code. To explore the impact of generative AI on cloud-native DevOps, I sat down with David DeSanto, chief product officer at GitLab, during KubeCon + CloudNativeCon 2023. DeSanto highlighted the potential of AI in areas like code review, test generation, and planning, which could significantly enhance operational efficiency.
One area where AI can streamline the development process is code review. According to a 2023 GitLab survey, developers spend 75% of their time on tasks other than coding. DeSanto suggests using machine learning to automatically suggest the right reviewers for new code changes, based on their domain knowledge. Additionally, leveraging large language models (LLMs) could help generate AI-powered commit messages or natural language summaries of code reviews, simplifying collaboration.
AI can also assist in various other aspects of cloud-native development. GitLab is exploring experimental AI features, including generating natural language descriptions of complex code bases, summarizing team discussions for newcomers, automating test generation, recalling commonly used Git commands, and applying root cause analysis for failures. These AI capabilities can enhance understanding, efficiency, and decision-making within development teams.
While 90% of organizations already use or plan to use AI in software development, some technical leaders remain hesitant to incorporate AI in cloud-native application development due to security concerns. DeSanto acknowledges the risks associated with large language models, such as hallucinations, privacy issues, and intellectual property leaks. However, market pressures and the adoption of AI by competitors are pushing companies to explore AI-assisted development cautiously.
DeSanto advises starting with simple AI features that have a low impact, such as suggested reviewers. By gradually experimenting with more advanced AI use cases, companies can strike a balance between leveraging AI and managing risks. It is crucial to automate low-risk areas, such as code review and vulnerability identification, while being cautious when granting AI full control over critical processes like automatically modifying a production website.
The interest in generative AI continues to grow rapidly. Developers are increasingly adopting AI, with surveys reporting that 70% of them already use or plan to use AI in their work. GitHub’s Octoverse report noted a 248% increase in public generative AI projects on the platform since 2022, and a third of open source projects have maintainers employing the Copilot AI tool.
Engineers and DevOps teams must consider leveraging AI within their workflows to gain a competitive advantage. However, they should carefully assess the risks, start small, and gradually explore more advanced AI use cases. While concerns about security and privacy persist, the benefits of AI-assisted development are difficult to ignore in today’s fast-paced and evolving software development landscape.
In conclusion, generative AI has the potential to significantly accelerate cloud-native development and collaboration. By applying AI to code review, test generation, and planning, organizations can achieve operational efficiency and enhance development processes. However, concerns about security and privacy surrounding large language models must be addressed. To stay competitive, engineers and DevOps teams should embrace AI in their workflows while carefully managing risks and focusing on gradual adoption.
