Show Notes
The most useful coding agents can mutate their environments by downloading packages, writing files, and connecting to services across the network. However, that freedom also presents dangers, and promises to usher in a new wave of security threats.
Docker recently announced Docker Sandboxes, which give each agent its own isolated micro VM while preserving the familiar ergonomics of a container. A standard container shares the host’s kernel, but a micro VM emulates hardware and runs its own kernel, giving a stronger security boundary around code that cannot be trusted.
Mark Cavage is the President and COO of Docker, and he previously worked at companies including Stripe, AWS and Oracle. In this episode, Mark joins Gregor Vand for a wide-ranging conversation that includes why agents break the immutability assumptions containers were built on, how micro VMs differ from both containers and traditional VMs, and the still-unsolved challenge of giving agents scoped, trustworthy access to sensitive services and data.
Sponsorship inquiries: [email protected]
The post Docker and Sandboxing AI Agents appeared first on Software Engineering Daily.
More from Software Engineering Daily
Codebases in enterprise organizations often span hundreds of thousands, or even millions, of lines of code. In many cases, changes must follow established patterns precisely, or risk introducing costly security or operational issues. A common challenge with AI coding tools is that they struggle at this scale: LLM context windows cover only a fraction of a large codebase, and output quality degrades the longer an agent runs.
In recent years, the main impact of AI has been on the digital world, and most notably software engineering. However, its longer-term effect on the physical world may prove even more profound, reshaping how cars drive themselves, how factories run, and how robots move through homes and warehouses.
Software testing is widely understood as essential to building reliable systems, and it’s one of the hardest things to get right. A fundamental issue is that tests are good at confirming the behavior a team expects, but the most damaging failures tend to hide in the edge cases no one thought to check during development. Those blind spots are only growing as AI produces more code that no human has closely reviewed.
SED News is a monthly podcast from Software Engineering Daily where hosts Gregor Vand and Sean Falconer break down the biggest stories shaping software engineering, Silicon Valley, and the broader tech industry.
The popular story of Steve Jobs is that he co-founded Apple, was pushed out, wandered through a failed venture called NeXT, then returned to save the company and usher in the iPod and iPhone. That version is memorable, widely repeated, and, in an important sense, wrong. His years at NeXT are usually treated as a footnote, when they were among the most consequential of his life and career.