https://future-os-blog.github.io/ Engineering notes from the FutureOS team: agent internals, the RPC wire contract, loop orchestration, and the desktop/mobile/terminal surfaces built on top of them. en 2026-09-21 https://future-os-blog.github.io/posts/why-grpc-not-acp.html https://future-os-blog.github.io/posts/why-grpc-not-acp.html Mon, 21 Sep 2026 00:00:00 +0000 The question came up while we were wiring the desktop to the agent: why are we hand-rolling a gRPC protocol when the Agent Client Protocol (ACP) already exists for exactly the "UI talks to an agent" problem? It was a… protocol grpc acp desktop architecture https://future-os-blog.github.io/posts/screenshotting-a-desktop-app-with-no-display.html https://future-os-blog.github.io/posts/screenshotting-a-desktop-app-with-no-display.html Mon, 14 Sep 2026 00:00:00 +0000 Every product screenshot and demo video in FutureOS — including the figures in this blog — is rendered by driving the real React UI in headless Chrome against mock data. No display, no running app, no hand-drawn mockups. The same harness doubles as a UI regression tool. desktop tooling internals https://future-os-blog.github.io/posts/sandboxing-an-agent-on-three-oses.html https://future-os-blog.github.io/posts/sandboxing-an-agent-on-three-oses.html Sun, 06 Sep 2026 00:00:00 +0000 The product promise is one line: the agent should not touch what you did not allow. Delivering it takes three completely different OS mechanisms — Seatbelt, Bubblewrap, and a restricted token with NTFS ACLs — and an honest list of places where the promise cannot be fully kept. agent desktop internals https://future-os-blog.github.io/posts/phone-desktop-e2ee.html https://future-os-blog.github.io/posts/phone-desktop-e2ee.html Sat, 29 Aug 2026 00:00:00 +0000 FutureOS Mobile controls sessions on your desktop over a NATS relay. The relay is on the public network, so it has to be treated as hostile — able to drop, reorder, replay and fabricate messages, but never able to read a command or forge one. Here's how the v2 channel achieves that with a Noise handshake and a ChaCha20-Poly1305 record layer. agent mobile desktop security e2ee https://future-os-blog.github.io/posts/loop-engineering.html https://future-os-blog.github.io/posts/loop-engineering.html Sat, 22 Aug 2026 00:00:00 +0000 A chat loses context. future-loop turns "keep an eye on this for a week" into a durable goal — a todo graph, human gates, per-step evidence, and a verifiable definition of done that survives sessions, restarts, and parallel workers. This is how the control plane works, and how it carried the Matilda tiling run. agent loop multi-agent architecture https://future-os-blog.github.io/posts/solving-combinatorics-with-four-llms.html https://future-os-blog.github.io/posts/solving-combinatorics-with-four-llms.html Thu, 13 Aug 2026 00:00:00 +0000 We gave a 16×16 tiling problem to four different models working in parallel under the future-loop control plane. Over 4.5 hours and 15 rounds the answer went from 30 to a verified 21 — and the two decisive breaks came from different models disagreeing with each other, not from any single model reasoning harder. agent loop multi-agent llm experiments https://future-os-blog.github.io/posts/context-compaction-measured.html https://future-os-blog.github.io/posts/context-compaction-measured.html Tue, 04 Aug 2026 00:00:00 +0000 We ran three context-compaction strategies through the same 178-question exam. FutureOS kept 83% of the values that had been compacted away; OpenCode kept 47%, Codex 38%. The difference isn't a better summary — it's what each system decides a compaction should keep. agent compaction llm experiments