AI agents are having their Kubernetes moment: agent logic sinks into the platform layer
bibryam · x · 2026-10-17
Bilgin Ibryam argues AI agents are undergoing a shift mirroring microservices' move to Kubernetes a decade ago: model/tool-calling logic is migrating from app code into reusable harnesses.
Key points
- Harness-based agents execute code in sandboxes, configure skills/plugins/MCP connections via files, and keep session state on the filesystem.
- This demands stronger runtime isolation and lifecycle management, including filesystem persistence.
- Network controls must govern what data a nondeterministic agent can access and where it can send it.
Diagram architecture: Ingress (Envoy/Agent Gateway: auth, policy checks, input filtering, secret removal) → Agent Harness (planning, agent loop, context management, evaluation, guardrails; e.g. Claude Code, Codex, ADK, Strands) → Egress (tool calls, LLM calls, secret injection, policy enforcement), all on an agentic infrastructure substrate (sandboxing, persistence, checkpointing, suspend/resume, execution environments, lifecycle management).
More from coding & agent
- No accepted agent loop yet for Jev-style decision models plus LLM, Reddit thread finds — emersoftware · 2026-10-17
- AI influencer hands an agent full control of a $500 day out — and it's going "horribly wrong" — alliekmiller · 2026-10-17
- Microsoft Copilot Studio adds agent accounts, memory and triggers for persistent business-process agents — clamanna · 2026-10-17
- Parser-free code index for AI agents cuts grep bills 42% with 4x smaller tool outputs — microbass · 2026-10-17
- Claude now writes 90% of Claude Code's code — yet Anthropic still lists 67 open research roles — robleclerc · 2026-10-17
- Inside The Very Next Day: Asteria animators use a custom LTX IC-LoRA on open weights to carry 2D keyframes — lmoroney · 2026-10-17