CORE challenges reactive AI defaults

Timers are automation, not proactivity. CORE reads like an attempt to make assistant behavior a memory problem first, but the harder test is deciding when context should trigger action.

• –The Reddit thread draws a clean line between scheduled checks and contextual awareness, which is the right frame for the next assistant layer. [Reddit thread](https://www.reddit.com/r/singularity/comments/1s622vd/every_ai_assistant_built_is_reactive_by_design_it/)
• –CORE's site describes a user-owned memory graph across IDEs, agents, and apps, so the pitch is bigger than reminders or inbox sweeps. [CORE site](https://heysol.ai/core)
• –The repo presents CORE as a persistent memory layer for AI apps, which matches the idea that assistants need cross-tool context. [GitHub repo](https://github.com/RedPlanetHQ/core)
• –The real unsolved problem is interruption policy: merge related signals, suppress noise, and avoid turning memory into notification spam.