GitLab outlines AI-assisted future, as Gartner claims DevOps platforms will dominate over traditional toolchains

GitLab outlines AI-assisted future, as Gartner claims DevOps platforms will dominate over traditional toolchains

GitLab introduced Duo at its launch event for GitLab 16 last month, a suite of AI capabilities for DevSecOps (Development, Security and Operations) which looks like its answer to GitHub Copilot.

The name Duo, like that of its rival GitHub Copilot, is intended to describe how the AI works alongside the developer forming a twosome. Duo features include code suggestions, issue summaries, merge request summaries, vulnerability explanations, test generation, and explaining source code. Duo is currently in preview with pricing yet to be determined.

Whereas Microsoft is aligned with the Azure cloud platform and looks to Open AI for its various Copilot offerings, GitLab is partnering with Google Cloud and building on its Vertex AI platform.

GitLab add-ins for Visual Studio Code and just announced Visual Studio are available, including Duo code suggestions as well as other GitLab features. Code suggestions are “routed through Google Vertex AI Codey APIs,” according to the documentation, which means it can be used with the 14 programming languages the Codey APIs support.

Duo code suggestions are not as mature as Copilot. Known limitations include “specific situations that can produce unexpected or incoherent results,” according to the docs, while early adopters commenting here mention issues including not playing well with existing Intellisense code completion in VS Code, difficulty getting suggestions working, or distracting from code flow. “Automatic code suggestions make it hard to stay in the zone because I have to constantly evaluate code that is not mine and lose concentration,” one dev complained, asking that it be shown only on demand.

Despite these kinds of limitations, GitLab sees AI assistance across its platform as essential for boosting productivity, with chief product officer David DeSanto making the extravagant claim that “Our goal is to help you achieve a 10x improvement in workflow efficiency.”

GitLab has long aimed to be a comprehensive DevOps platform and AI has become an essential ingredient in what that now means.

According to a Gartner report last month, DevOps platforms like GitLab and GitHub are set to take increasing market share from traditional DevOps toolchains, involving multiple products. “By 2027, 75% of organizations will have switched from multiple point solutions to DevOps platforms to streamline application delivery, up from 25% in 2023,” the analysts state.

The key features of a DevOps platform include product planning, version control, continuous integration, test automation, release orchestration, security and compliance policy automation, monitoring and observability. When these are combined into one platform, “prebuilt integration between different components … reduces cognitive load and leads to improved visibility, auditability and traceability,” Gartner’s research says.

Few if any platforms tick all these boxes though, with GitLab and Microsoft’s GitHub combined with Azure DevOps being the two that come closest. Regarding GitHub though, Gartner cites lack of support for “flow metrics and software performance metrics,” as well as market confusion between GitHub and Azure DevOps; while GitLab lacks strong content collaboration, and is weak in remote development environments, according to the analyst.

Others though are further behind. Only GitLab, Microsoft and Atlassian make it into the sought-after top right corner of Gartner’s “Magic Quadrant”, showing completeness of vision as well as ability to execute, but Atlassian suffers from low adoption of its CI/CD capabilities, according to Gartner.

AWS has a suite of DevOps services including CodeCommit, CodeDeploy, CodePipeline and CodeArtifact, as well as CodeCatalyst as a unified offering, which the analysts call a good fit for those looking for integration with AWS services – and by implication, a poor fit for more general use. JetBrains is strong on the coding side but according to Gartner, “lagging operations capabilities.”