Agent Phones Enter the System-Era
智东西 · wechat · 2026-07-17
This article explores the Agent Phone, detailing why "AI Phones" are shifting from simply integrating large models to restructuring the entire OS around AI agents.
Using StepFun's STEPXNeo / StepAOS as an example, it illustrates three key capabilities being developed:
- Memory: Cross-app, cross-task long-term memory, separating user and agent domains.
- Execution: Breaking down app features into directly callable "atomic capabilities," freeing agents from simulating UI taps.
- Security: Introducing TEE and a "trusted, visible, controllable, reversible" mechanism to support execution tracking and undoing errors.
The article stresses that the real bottleneck for agent phones lies in the ecosystem, not just the OS. StepFun has partnered with Ctrip, Alipay, DiDi, and Meituan to integrate agent capabilities into travel, lifestyle services, and office scenarios. The conclusion: AI smartphone competition is evolving from "model integration" into a phase defined by "system architecture + task execution + ecosystem collaboration."
More from Embodied
- Skyfall AI's SKY-001 world model turns sketches into CAD builds in 3.9s, vs 199s for GPT-6 Astra — rohanpaul_ai · 2026-10-27
- Tesla FSD approved in Latvia, the 10th European country; Musk praises smaller nations for moving fastest — elonmusk · 2026-10-27
- WanGP v18.81 adds Tencent's Hy-Nomas world model, playable on 8GB VRAM — cocktailpeanut · 2026-10-27
- NVIDIA open-sources FLORA, a robotics data engine ~125x cheaper than GPT-6 Astra — nvidia · 2026-10-27
- Embodied AI startups need a different operational discipline, investor says — ___Mufasaa · 2026-10-27
- π0.5 study: raising demos from 10 to 438 lifts Place Ball in Bowl success from 7% to 83% — stingning · 2026-10-27