Long-running agents aren't in prod because trust infrastructure doesn't exist yet
Arm1end · reddit · 2026-09-05
The author argues that long-running agents still can't run unattended not because of model capability, but because "trust" as infrastructure doesn't exist yet.
Using an aviation analogy: passengers trust the pilot's licensing, recurring checks, maintenance logs and black box—not the person. Decomposed, agent trust means a concrete checklist:
- Permissions that scale with decisions: allow a $50 API call, block at $51, without killing the process;
- Boundaries that hold on their own, not ones humans must remember to check;
- Pause-for-sign-off without losing agent state;
- Precise execution records (the agent's black box) to reconstruct failures.
Nobody is holding agents back on purpose—the infrastructure that makes unattended runs a boring, safe decision just hasn't been built.
More from coding & agent
- GPT Sol plays real-time table tennis via latency-aware agent research loop — wzenus · 2026-10-21
- PostgreSQL 19 slips to November as LLM-assisted code review floods committers with edge cases — sanjaykalra · 2026-10-21
- 9x Vision open-sources video perception for agents: timestamped evidence via MCP and CLI — koto9x · 2026-10-21
- LangChain adds Human-in-the-Loop beta to Managed Deep Agents — LangChain · 2026-10-21
- Voxtral 3 ships a LiveKit-based voice agent SDK streaming listen-think-speak — MistralAI · 2026-10-21
- Open-sourced pixel art normals generator ships as Rust CLI and wasm library — HugoDuprez · 2026-10-20