JFrog CEO: No One Cares About Source Code Anymore
JFrog CEO Shlomi Ben Haim told approximately 500 software engineers today that in the age of artificial intelligence (AI) no one will soon care much about source code anymore. Speaking at the company’s swampUP 2026 conference, Ben Haim said AI tools are now creating executable binaries directly without relying on source code that historically existed […]
JFrog CEO Shlomi Ben Haim warned at the swampUP 2026 conference that in the age of artificial intelligence (AI), no one will soon care much about source code. AI tools are now creating executable binaries directly without the need for human-readable source code. Ben Haim emphasized that scanning application binaries has become the focal point for securing software supply chains in this AI-driven era.
He noted that the volume of code generated using AI coding tools far exceeds existing DevSecOps workflows, highlighting the need to focus on securing executables stored in registries like JFrog Artifactory. Ben Haim forecasted that by 2030, 2.5 billion active agents would complete 459 trillion actions, making it increasingly challenging for DevSecOps teams to update and replace software packages.
He proposed a shift towards building applications that are self-healing, with each executable validating its cryptographic security without human intervention. The role of DevSecOps teams may evolve, focusing more on securing prompts and specification frameworks used to create executables. While the exact evolution of DevSecOps teams is uncertain, Ben Haim stressed that the fundamental assumptions about securing software in the AI era need to change.
The challenge lies in the ongoing development of agentic engineering, lacking well-defined best practices for securing software supply chains in the AI era. Despite the escalating rate of vulnerabilities in both existing and new applications, there is a growing urgency around application security, a sentiment that has been lacking for too long.
Written by urgent.news from DevOps.com's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.