Muse Gadgets

Original Article ↗ Hacker News Discussion ↗

Meta’s new Muse Gadgets SDK, which lets hobbyists connect ESP32-style hardware and home devices to its Muse AI assistant, is drawing both excitement and skepticism. Enthusiasts like the accessible, open-source firmware and potential for playful “embodied” assistants, while critics point to subscription-locked hardware, restrictive SDK terms, and Meta’s track record on privacy and lock-in as reasons to avoid integrating it into their homes. Many argue similar functionality can already be achieved with open tools like Home Assistant, and question why anyone should grant Meta deeper access to their local networks and devices.

Overall sentiment

  • Strong split between excitement about the technical possibilities and deep distrust of Meta as the provider.
  • Many say they would never allow Meta’s AI or hardware on their home network, regardless of how cool the tech is.

Trust, privacy, and safety

  • Repeated concern about giving Meta more access: Wi‑Fi credentials, LAN topology, IoT control, and rich contextual data.
  • A past incident where Muse shared a user’s home address with a stranger is discussed; one comment notes the user had chosen “always approve,” others respond that most people will click through such prompts.
  • People worry about agents controlling physical devices (doors, lights, etc.) and about non‑deterministic behavior on home networks.
  • Several frame the “name of the game” as capturing proprietary context; this announcement is seen as a way to harvest more of it.

Ownership, openness, and lock‑in

  • Marketing around “make it your own” and “open source hardware” is heavily criticized.
  • Key objections:
    • Gadgets require tokens, with device limits and restrictive SDK terms.
    • The Home Link device needs an active subscription; terms say it stops functioning when the subscription ends and can’t be reflashed.
    • SDK is explicitly “not a supported product” and can be changed or withdrawn at any time.
  • Some call this pseudo‑grassroots and liken it to tenant/landlord or sharecropping dynamics.
  • Others respond that API tokens are normal, firmware is Apache‑licensed, and it’s fine as a tinkering toy if you already use Muse.

Technical ecosystem and alternatives

  • Enthusiasm around ESP32, with mild worry about price spikes; others note ESP32 survived prior shortages.
  • Mentions of alternative MCUs and of existing open setups (Home Assistant, ESPHome).
  • People are already flashing the firmware to ESP32 boards; voice features were reportedly fixed quickly.
  • A community TypeScript client was built from the protocol described in the docs.
  • Some prefer fully local/offline agents and propose building independent devices (e.g., custom “AI routers” or keychain assistants).

Use cases and product strategy

  • Interest in embodied assistants for small, ambient tasks: status lights at desks, morning brief displays, simple home automations.
  • Some view this as Meta smartly seeding a hardware ecosystem and competing by products/platforms rather than pure model strength. Others see it mainly as developer‑oriented PR that will eventually “enshittify.”