Uploaded August 2026 | Updated September 2026, 6 hours ago
In this AMA livestream, Viktor and Scott tackle a wide range of questions from the community, covering everything from AI-assisted development to Kubernetes infrastructure decisions. The conversation kicks off with a thoughtful debate on whether AI agents are making programming less fun, with both hosts agreeing that while AI handles the mundane work, it actually frees developers to focus on the complex, interesting challenges. They also discuss practical workflows — using tools like Cursor in ask mode as a sounding board, maintaining coding proficiency, and knowing when to let agents run autonomously versus staying hands-on.
The second half of the stream dives deep into infrastructure topics, including choosing the right Kubernetes distribution for on-prem environments (with shoutouts to Talos, VKS, and NKP), the SaaS versus self-hosted debate for Kafka and CouchBase, and the role of RAG pipelines versus MCP integrations for querying internal documentation. Viktor and Scott also cover Kubernetes security hardening — touching on container image scanning, runtime protection with eBPF, admission policies with Kyverno and OPA, and the importance of SBOMs, image signing, and attestations. Rounding things out, they share advice for those looking to break into DevSecOps and platform engineering roles, recommending open source contributions to projects like Kubernetes Cluster API as a powerful way to build real-world experience.
▬▬▬▬▬▬ ⏱ Timecodes ⏱ ▬▬▬▬▬▬
00:00 Intro (skip to first question)
05:59 AI agents making coding more effective but less fun
10:55 Best Kubernetes flavor for single app on-prem production
13:22 Kafka and CouchBase: self-hosted vs SaaS for 100M events/day
16:30 Self-hosted RAG pipelines vs MCP for data retrieval
18:51 How to get hired in DevSecOps and platform engineering roles
22:25 Best open source projects to build DevOps experience
23:18 Three-day bug solving no longer justified with AI?
24:14 Best open source projects recommended for DevOps jobs
24:14 Kubernetes SRE agents for troubleshooting production errors
28:15 Virtual Kubernetes clusters per customer: best solution
28:53 Are PostgreSQL operators ready for 10TB and 100K+ users?
30:17 Network overhead with CSI drivers in Kubernetes
39:58 Using AI as a second brain: storage format and workflow
43:23 Kubernetes security hardening: common day-one mistakes
47:56 SBOM vs KBOM for container security
50:14 IDEs vs TUI terminals for AI agent workflows
In this AMA livestream, Viktor and Scott tackle a wide range of questions from the community, covering everything from AI-assisted development to Kubernetes infrastructure decisions. The conversation kicks off with a thoughtful debate on whether AI agents are making programming less fun, with both hosts agreeing that while AI handles the mundane work, it actually frees developers to focus on the complex, interesting challenges. They also discuss practical workflows — using tools like Cursor in ask mode as a sounding board, maintaining coding proficiency, and knowing when to let agents run autonomously versus staying hands-on.
The second half of the stream dives deep into infrastructure topics, including choosing the right Kubernetes distribution for on-prem environments (with shoutouts to Talos, VKS, and NKP), the SaaS versus self-hosted debate for Kafka and CouchBase, and the role of RAG pipelines versus MCP integrations for querying internal documentation. Viktor and Scott also cover Kubernetes security hardening — touching on container image scanning, runtime protection with eBPF, admission policies with Kyverno and OPA, and the importance of SBOMs, image signing, and attestations. Rounding things out, they share advice for those looking to break into DevSecOps and platform engineering roles, recommending open source contributions to projects like Kubernetes Cluster API as a powerful way to build real-world experience.
▬▬▬▬▬▬ ⏱ Timecodes ⏱ ▬▬▬▬▬▬
00:00 Intro (skip to first question)
05:59 AI agents making coding more effective but less fun
10:55 Best Kubernetes flavor for single app on-prem production
13:22 Kafka and CouchBase: self-hosted vs SaaS for 100M events/day
16:30 Self-hosted RAG pipelines vs MCP for data retrieval
18:51 How to get hired in DevSecOps and platform engineering roles
22:25 Best open source projects to build DevOps experience
23:18 Three-day bug solving no longer justified with AI?
24:14 Best open source projects recommended for DevOps jobs
24:14 Kubernetes SRE agents for troubleshooting production errors
28:15 Virtual Kubernetes clusters per customer: best solution
28:53 Are PostgreSQL operators ready for 10TB and 100K+ users?
30:17 Network overhead with CSI drivers in Kubernetes
39:58 Using AI as a second brain: storage format and workflow
43:23 Kubernetes security hardening: common day-one mistakes
47:56 SBOM vs KBOM for container security
50:14 IDEs vs TUI terminals for AI agent workflows
