As AI coding agents ship code faster than teams can watch it, the question facing platform and SRE teams is no longer just what is broken, but whether the telemetry behind the answer can be trusted. That tension is at the center of ControlTheory’s story heading into KubeCon + CloudNativeCon North America 2026 in Salt Lake City. The company’s Dstl8 platform is an AI SRE built on what it calls Telemetry Distillation: it distills noisy telemetry, finds root cause, and hands the evidence to the engineer or coding agent, such as Claude Code, Cursor, or Codex, that shipped the change. Dstl8 became generally available in September, and the company says it is bringing a major new release to the show.
In this VMblog pre-show Q&A, Bob Quillin, CEO and Co-Founder of ControlTheory, explains why legacy observability’s “Telemetry Tax” leaves AI SREs reasoning over data that was cut down to control costs, and why he believes the next phase of the category will be defined by trust. He also shares a production result from an enterprise fintech, where Dstl8 detected 336 incidents across 13 Kubernetes clusters and the team resolved 328 of them with zero alert rules written. Quillin also covers the company’s open source work with Gonzo, how attendees can try Dstl8, and where to find ControlTheory at booth 188 in Hall 1, Solutions Showcase, at the Salt Palace.
++
VMblog: What’s your elevator pitch for KubeCon + CloudNativeCon North America 2026? If attendees only remember one thing about your company after visiting your booth, what should it be?
Bob Quillin: Observability without feedback is just watching. Dstl8 by ControlTheory is an AI SRE built on Telemetry Distillation: it distills your telemetry, finds root cause, and hands it to the engineer or coding agent who shipped the change. If attendees remember one thing, I hope it’s this question: if your AI SRE tells you what is wrong and drives the fix, on telemetry it can trust, what is your observability platform for?
VMblog: This is KubeCon’s return to Salt Lake City — what makes the North American cloud-native community particularly strategic for your business objectives this year?
Quillin: Kubernetes and OpenTelemetry are the foundation for AI-native computing. Dstl8 and our open source tool Gonzo were built on and for these CNCF projects. KubeCon North America brings together SRE, development, and platform engineering leaders shaping AI SRE and recasting observability for the AI era.
VMblog: Where exactly can attendees find you on the show floor at the Salt Palace Convention Center, and what’s the one must-see demo or experience you’ve built that showcases your technology’s real-world impact?
Quillin: The ControlTheory booth is 188 in Hall 1, Solutions Showcase, at the Salt Palace. The must-see demo is Dstl8, our AI SRE, distilling noisy telemetry, finding root cause, and handing it to Claude Code, Cursor, or Codex. We’re also introducing a major new release, and booth 188 is the place to see it first.
We’ll run live Dstl8 demos all show, plus our Prompt Wars game and 15-minute 1:1s you can book in advance. In Prompt Wars, players write a prompt for an AI bot that battles rival bots. You can play now at kubecon.dstl8.ai; the leaderboard resets when KubeCon opens, and prizes (a Mac mini and a Raspberry Pi kit) go only to players who visit booth 188.
VMblog: Are you launching any new products, unveiling significant features, or announcing partnerships at KubeCon NA 2026? What should attendees watch for?
Quillin: Yes. Dstl8 became generally available in September, and we’re bringing a major Dstl8 release to Salt Lake City. We’re saving the details for the show, but if you’ve followed how we think about runtime feedback, expect the loop to get a lot bigger. Stop by booth 188 or learn more at dstl8.ai.
VMblog: Let’s get technical — what specific architectural challenges or operational bottlenecks does your solution address for cloud-native teams heading into 2027?
Quillin: Legacy observability charges to ingest, index, and retain every byte, so teams sample traces, drop debug logs, and shorten retention to control the bill. We call that the Telemetry Tax, and any AI SRE built on that data can only reason over what survived the cuts. Dstl8 uses Telemetry Distillation to keep the facts that matter close to the source, so teams get complete answers without paying real-time prices for an archive.
VMblog: As we head toward 2027, what’s your unique value proposition? With dozens of vendors claiming similar AI-infrastructure capabilities, why should CTOs choose your solution?
Quillin: Most AI SREs arrive after the hard part is over, reasoning over data that was cut down to control cost. Dstl8 starts earlier. It distills the telemetry itself, correlates it across the deploy, and hands root cause, with the evidence, to the engineer or coding agent who shipped the change. Owning the telemetry is what lets an AI SRE show its work.
VMblog: Can you share a compelling customer story or use case that demonstrates tangible outcomes — whether that’s cost savings, performance improvements, or accelerated delivery cycles?
Quillin: In production at an enterprise fintech, Dstl8 detected 336 incidents across 13 Kubernetes clusters over two months, and the team resolved 328 of them, with zero alert rules written. Nobody had to guess at thresholds in advance, because Dstl8 worked from the telemetry itself.
VMblog: AI coding agents are shipping code faster than teams can watch it. What breaks first in production?
Quillin: Visibility breaks first. Code volume and velocity are climbing at the same time, and they compound, so the problems that reach production are the ones nobody wrote a test or an alert for. Alert rules and dashboards assume a system that changes slowly, and maintaining them pulls platform teams away from building leverage. When agents ship all day, teams need runtime feedback that goes back to the engineer or agent who shipped the change while the code is still moving.
VMblog: CNCF has introduced a brand-new “AI Inference + Agentic” track for 2026, covering GPU scheduling, model serving, and observability for production AI systems. Where does your company fit into that conversation, and what’s driving your investment there?
Quillin: Trust is the most critical factor affecting AI inference and agentic deployments today, particularly when it comes to observability. Production AI systems, and the agents that operate them, are only as good as the telemetry underneath. A data gap is unknown, not zero, and an agent that can’t tell the difference will give confident, wrong answers. That’s what drives our investment: Dstl8 uses Telemetry Distillation so people and agents reason over trusted, distilled telemetry instead of a raw dump or a sampled one.
VMblog: AI agents are starting to investigate production issues on their own. How do you keep an agent trustworthy when the stakes are high?
Quillin: Trust starts with evidence. Möbius, Dstl8’s AI agent, connects root cause across the deployment chain, keyed to the deploy event, pinpoints code issues down to the line, and shows the evidence behind every diagnosis. It then hands what it finds to the team’s own coding agent, such as Claude Code, Cursor, or Codex, and an engineer reviews and approves the change, or lets their agent act autonomously within the parameters their organization sets.
VMblog: Give us your executive pitch — how do you articulate ROI and business value to C-suite leaders who need to justify cloud-native and AI infrastructure investments?
Quillin: The ROI case is simple: pay for fewer terabytes and get faster answers. Telemetry Distillation cuts the Telemetry Tax because teams stop paying to index everything just in case, and Möbius gives engineers an evidence-based diagnosis instead of digging through dashboards. My advice to executives is to ask what they actually get for their observability spend.
VMblog: Beyond the booth presence, what’s your broader community engagement strategy for KubeCon NA 2026? How are you contributing to the cloud-native ecosystem’s growth?
Quillin: We contribute through open source and build on CNCF standards. ControlTheory created Gonzo, an open-source OTLP terminal UI with more than 2,700 GitHub stars, and Dstl8.lite adds a free local GUI for it. Dstl8 is built on an OpenTelemetry foundation and speaks OTLP natively, so teams already running the Collector can add Dstl8 as one more exporter and keep the standards they use today. Kubernetes support goes deep: pod and container logs, Kubernetes events, cluster topology, and app and infra visibility, correlated by pod, deployment, namespace, and cluster.
VMblog: How can attendees try Dstl8 or Gonzo before or during the show?
Quillin: Dstl8 has a free 14-day trial with no credit card required and setup in minutes, at dstl8.ai. Gonzo is open source, so brew install gonzo gets you started in a terminal. At the show, you can book a 15-minute 1:1 at controltheory.com/news/kubecon or just stop by booth 188.
VMblog: Looking beyond KubeCon NA 2026, what emerging trends or technologies should cloud-native practitioners be preparing for? What’s on your company’s innovation roadmap that signals where the industry is heading?
Quillin: Observability is splitting in two. AI SRE is taking over the live half, acting on what is happening right now, and the rest, while still important, is just storage and compliance. The AI SRE category is near the peak of its hype cycle, and the next phase will be about trust, because an AI SRE can give wrong answers when the data it runs on is incomplete. Practitioners should ask what their AI SRE actually sees, and what happens when nothing fires an alert. Our roadmap is aimed squarely at the live half and its role in the new AI SDLC, so stay tuned for more new features at KubeCon in Salt Lake City. See you there!
##
