GitLab All in on AI: CEO Predicts Increased Demand for Coders
GitLab is all in on AI, with CEO and co-founder Sid Sijbrandij calling it “one of the most exciting technology developments of our time” and making an unusual prediction that it will create demand for more programmers.
“AI represents a major shift for our industry. It fundamentally changes the way that software is developed,” Sijbrandij said on GitLab’s earnings call Monday. “We believe it will accelerate our ability to help organizations make software, faster. I am excited about this next wave of technology innovation.”
GitLab plans to incorporate AI at all levels of its DevSecOps platform, he added.
“We believe that an AI-powered platform focused solely on the Developer persona is incomplete. It is missing essential Security, Operations, and Enterprise functionality,” Sijbrandij said. “Remember: developers spend only a small fraction of their time developing code. The real promise of AI extends far beyond code creation.”
During the first quarter of the year, GitLab delivered five new AI features, followed by five more in May with the release of GitLab 16 — including a beta of Code Suggestion, as well as security testing and analysis, observability and proactive vulnerability detection. Additional AI-powered features available include Suggested Reviewers for code review, Explain This Vulnerability for vulnerability remediation, and Value Stream Forecasting for predicting future team efficiency. Code Suggestions does what its name implies, making code suggestions to developers as they type.
“We’re proud to have 10 AI features available to customers today, almost three times more than the competition,” he said, adding that applying AI to a single data store, for the full software development life cycle, also creates compelling business outcomes and is something he believes can be done with GitLab.
GitLab continues to iterate on Code Suggestions and expects to make it generally available later this year. The company has also boosted language support from six languages to 13, so more developers can use it, he added.
“Code Suggestions is uniquely built with privacy first as a critical foundation,” he said. “Our customers’ proprietary source code never leaves GitLab’s cloud infrastructure. This means that their source code stays secure. In addition, model output is not stored and not used as training data.”
AI Support for Development Teams
Also later this year, the company plans to introduce an AI add-on focused on supporting development teams, which will include Code Suggestions functionality, across all GitLab’s tiers at an anticipated price point of $9 per user per month, billed annually, he said.
He noted that they’d had many conversations with senior-level customers, but one comment from the CTO of a top European bank stood out.
“When the conversation moved into AI, the CTO said something extremely interesting. He said: Code generation is only one aspect of the development cycle. If we only optimize code generation, everything else downstream from the development team — including QA, security, and operations — breaks. It breaks because these other teams involved in software development can’t keep up,” Sijbrandij said. “This point — incorporating AI throughout the software development lifecycle — is at the core of our AI strategy.”
Companies Reevaluating Strategies in the Wake of AI
Customers are also reevaluating their own software supply chain through the AI lens, he added. Additionally, chief information security officers are also engaging with AI, and applying governance, security, compliance and audit-ability to it.
He predicted that AI will increase GitLab’s market for three reasons. First, AI will make writing software easier, which in turn will expand the audience of people — including junior and citizen developers — who build software. Second, as developers become more productive, software will become less expensive to create, which will fuel demand for more software and require more developers to meet the additional need. Third, the company expects more customers will turn to its solutions as they build machine learning models and AI into their applications.
“As we add ModelOps capabilities to our DevSecOps platform, this will invite data science teams as new personas, and will allow these teams to work alongside their Dev, Sec, and Ops counterparts,” he said. “We see ModelOps as a big opportunity for GitLab.”
Sijbrandij also shared how global security and aerospace company Lockheed Martin used GitLab to reduce their toolchain and complexity while reducing costs. The Lockheed Martin team has reported 80x faster CI pipeline builds and 90% less time spent on system maintenance, he said, adding they’ve also retired thousands of Jenkins servers. They’ve also moved from monthly or weekly deliveries to daily or multiple daily deliveries.