Good synthesis of the Routines + CMA stack. One architectural piece I would add to the magic framing: CMA is a hosted meta-harness, so the durable contract is the resource shapes (sessions, agents, environments, vaults), not the loop itself, which is deliberately thin so harness assumptions do not go stale as models improve. And the boundary worth naming for adopters: self-hosted sandboxes move where tool calls run, but every tool result still transits Anthropic's context window for the next inference step. Execution sovereignty, not reasoning sovereignty - that distinction carries most of the enterprise adoption decision.
The brain/hands/log decomposition is right as far as it goes, but the May 19 updates move a boundary it doesn't capture: self-hosted sandboxes hand the Environment back to you (your VPC, your egress rules), while the loop itself — session orchestration, tool routing, the next inference step — still runs on Anthropic's control plane. I decomposed CMA from the wire up after the London event, and the result is that 'execution sovereignty' is only half right: owning the container is not owning the harness. The vault-proxied credentials are the tell — the real trust boundary sits between your sandbox and their router, and the new MCP tunnels make that seam load-bearing. At Sentry scale the question stops being build-vs-buy on infra and becomes which failure modes you're allowed to audit.
What’s interesting here is that Routines and Managed Agents quietly change the unit of software from “app” to “ongoing obligation.” Most SaaS tools still assume a user shows up, clicks around, and initiates work. But a routine is basically a promise: “this thing will keep happening, with context, unless something breaks.” That sounds small, but it creates a very different product surface.
Amodei is credible precisely because he has skin in the game on both sides. The question isn't whether to take it seriously — it's what the 1-5 year window actually means for people in transition now. Retraining programs historically take 3-7 years to mature. The timing gap is the real problem.
How are you handling state management between agent handoffs in these routines? That's consistently been the breaking point in my setups: the model loses critical context at exactly the wrong moment mid-task. Curious whether Managed Agents changes that or just moves the problem upstream.
Good synthesis of the Routines + CMA stack. One architectural piece I would add to the magic framing: CMA is a hosted meta-harness, so the durable contract is the resource shapes (sessions, agents, environments, vaults), not the loop itself, which is deliberately thin so harness assumptions do not go stale as models improve. And the boundary worth naming for adopters: self-hosted sandboxes move where tool calls run, but every tool result still transits Anthropic's context window for the next inference step. Execution sovereignty, not reasoning sovereignty - that distinction carries most of the enterprise adoption decision.
The brain/hands/log decomposition is right as far as it goes, but the May 19 updates move a boundary it doesn't capture: self-hosted sandboxes hand the Environment back to you (your VPC, your egress rules), while the loop itself — session orchestration, tool routing, the next inference step — still runs on Anthropic's control plane. I decomposed CMA from the wire up after the London event, and the result is that 'execution sovereignty' is only half right: owning the container is not owning the harness. The vault-proxied credentials are the tell — the real trust boundary sits between your sandbox and their router, and the new MCP tunnels make that seam load-bearing. At Sentry scale the question stops being build-vs-buy on infra and becomes which failure modes you're allowed to audit.
What’s interesting here is that Routines and Managed Agents quietly change the unit of software from “app” to “ongoing obligation.” Most SaaS tools still assume a user shows up, clicks around, and initiates work. But a routine is basically a promise: “this thing will keep happening, with context, unless something breaks.” That sounds small, but it creates a very different product surface.
Amodei is credible precisely because he has skin in the game on both sides. The question isn't whether to take it seriously — it's what the 1-5 year window actually means for people in transition now. Retraining programs historically take 3-7 years to mature. The timing gap is the real problem.
How are you handling state management between agent handoffs in these routines? That's consistently been the breaking point in my setups: the model loses critical context at exactly the wrong moment mid-task. Curious whether Managed Agents changes that or just moves the problem upstream.