Stop Killing Sandboxes: Agents Need Durable Disposable Computers

Fly.io introduces Sprites: durable, disposable cloud computers that boot in seconds, auto-idle, provide Anycast URLs, and support instant system checkpoints and restores. Kurt Mackey argues agents don’t want stateless sandboxes; they need persistent machines to avoid rebuilds, externalized state, and timeouts, and to leverage full app lifecycles. He demonstrates this with real examples, concluding that durable disposable computers should replace ephemeral sandboxes.

Key Points

• Ephemeral, read-only sandboxes hinder agent workflows by forcing rebuilds, externalized state, and short timeouts; agents need durable computers, not stateless containers.
• Sprites are fast, durable cloud computers: 1–2s creation, 100GB persistent storage, Anycast HTTPS, auto-idle billing, and near-instant system checkpoints/restores.
• Persistence simplifies development: avoid re-installing dependencies and hacking around state with buckets or plan files; treat rollback like git for the whole system.
• Agents can leverage full app lifecycles when state and logs persist, exemplified by Phoenix.new automatically observing and fixing errors from generated apps.
• Real-world illustration: the author built and runs a personal MDM on a Sprite with Claude; for many small apps, dev is prod, and Sprites make continuous, agent-led iteration viable.

Sentiment

The overall sentiment of the Hacker News discussion is overwhelmingly positive and enthusiastic. Users express significant excitement for Fly.io Sprites' concept and its key features, particularly instant boot, persistent storage, and checkpointing, seeing it as a valuable solution for developer environments and AI agent integration. While some questions and feature requests were raised, they were generally in the spirit of wanting to further explore and adopt the technology, rather than expressing fundamental disagreement or skepticism.

In Agreement

• Sprites offer a "seriously cool" and "fantastic" developer experience and API for sandboxed execution, especially for untrusted code and LLM agents.
• The combination of near-instant creation, automatic spin-down to zero cost, and durable, persistent storage (100GB) addresses a critical need for modern development and agent workflows, moving beyond stateless containers.
• Instant checkpointing and restoration is a game-changing feature, allowing for casual rollbacks, parallel experimentation, and "YOLO mode" development with agents.
• The concept of "disposable, durable computers" (BIC disposable computers) is highly appealing, encouraging a "when in doubt, just make another one" approach.
• There's a strong desire for more secure ways to run agents, and Fly.io's approach to security (e.g., Macaroons, hiding proxies within the VM) is seen as a potential major differentiator.
• The broader vision of more dynamic software and agents iterating on live systems aligns with the potential unlocked by tools like Sprites.

Opposed

• Some initial confusion exists regarding how Sprites truly differ from "fancy VPSs" or containers with durable storage, suggesting the distinction isn't immediately clear to all users.
• The initial lack of regional support is a significant concern for some users, leading to high latency for remote development and typing lag.
• There were requests for clearer documentation on specific operational aspects like GitHub authentication, on-machine APIs, and current usage/cost metrics, indicating some initial friction for adoption.
• Concerns were raised about potential unexpected costs without clear spending limits or scaling controls for Sprites.
• The transition workflow from a Sprite (development) to a more traditional production deployment (e.g., Fly Machine) is not yet fully clear or streamlined for all users.
• Competitors like `exe.dev` are exploring similar problem spaces, indicating that Sprites operate in a competitive market, though Sprites' fast checkpointing is noted as a differentiator.