Why edge computing isn’t just cloud at the network edge

Three core advantages that set edge apart, and why AI still struggles with systems programming.

Edge computing promises cost savings, privacy, and speed, but most developers still approach it with cloud-first thinking. In this episode, Roja Eswaran, PhD, Senior Engineer on the Cloud Infrastructure team at Geico, joins the We Love Open Source podcast to share why edge environments demand fundamentally different techniques than cloud, and why AI’s limitations in systems programming reveal exactly where human expertise still matters most.

Roja’s edge computing expertise comes from hands-on PhD research at SUNY Binghamton, where she worked with Raspberry Pi and Nvidia Jetson devices trying to boot virtual machines on resource-constrained hardware. That work revealed something critical: The techniques that work for cloud cannot be directly applied to edge because it’s a fundamentally different paradigm. Edge environments are heterogeneous, distributed across geographic locations, and operate under completely different constraints. That realization sparked her passion for edge computing and led her to Zededa, an edge computing startup, before joining Geico’s cloud infrastructure team.

Roja breaks edge computing down to three advantages. First, it’s cost effective. You don’t need massive infrastructure, just a $10 Orange Pi in your basement can handle edge workloads. Second, it’s privacy-focused. Your data and computation stay on the edge device at your location, so you’re not depending on third parties to store your information. Third, it’s fast. Storage and computation happen at the edge of the network, eliminating the latency of sending data over the internet.

Read more: 6 must-read Linux and open source tutorials

The conversation shifted to AI and whether it will replace developers. Roja uses GitHub Copilot herself for boilerplate code but is direct about its limitations. When you go down into operating systems and firmware where code becomes very complex, AI tends to hallucinate and fabricate incorrect information. Her take? Unless we solve the hallucination problem, AI won’t be replacing developers anytime soon. That’s why she tells aspiring developers to explore operating systems and firmware. A strong foundation in how systems work helps you use AI tools effectively while catching their mistakes.

Key takeaways

• Edge computing delivers three advantages cloud can’t match: Cost effectiveness (start with a $10 Orange Pi), privacy (data stays local), and speed (no internet latency). Cloud techniques don’t directly translate to edge environments.
• AI has real limits in systems programming: While tools like GitHub Copilot excel at boilerplate code, they struggle with complex, low-level code where hallucination becomes a serious problem.
• Foundation matters more as AI tools proliferate: Understanding operating systems and firmware gives developers the depth to use AI effectively and recognize when it’s wrong.

Roja also raised concerns about open source sustainability, noting that great projects get abandoned when developers change companies and lose the time or funding to maintain them. Her advice? Build that strong foundation. The developers who thrive will understand systems deeply, know when to leverage AI tools, and recognize what those tools can and cannot do.

More from We Love Open Source

• 15 open source backup solutions to protect your data
• 10 common Linux issues and the fast fixes you need
• Stop guessing, start measuring developer engagement
• 6 must-read Linux and open source tutorials
• 5 forces driving DevOps and AI in 2026