I'm terra tauri. I run platforms at scale: Kubernetes, networks, and now the AI agents that maintain them.
Previously at Grafana Labs I ran the platform under Mimir, Loki and Tempo, across 10,000+ nodes, zero-downtime CNI migrations, and softIRQ bottlenecks hunted at a million packets a second.
The default answer to "I need more from my AI" is still a bigger model on a bigger GPU. But every piece of the other answer, horizontal scaling for inference, already exists in the research. Here's how small, specialized, cache-friendly workers make AI cheap instead of scarce.
Bolting AI onto an organization built for walking pace doesn't make it faster. The bottleneck was never the typing. It was coordination, and AI only shone a light on it.
Why coding agents need a sandbox the agent cannot reach, an audit log the agent cannot write, and signed instructions the agent cannot rewrite.
I'm terra tauri. I'm enabling platform engineering at Bit Complete. I live rurally in the Nelson area, BC, and I'm involved with the Nelson Pride Society. Most of what I build is public: a fleet manager for coding agents, an offline tool for rewriting text without the AI tells, eBPF rate limiting for Kubernetes, and local inference on Apple Silicon.
Your game saves in this browser.
Too small to play
Make this window bigger to carry on, or go back to the site (Esc).