The Rabbit R1 is probably running Android and is powered by an Android app

Original Article ↗ Hacker News Discussion ↗

Revelations that the $200 Rabbit R1 “AI assistant” is essentially a single privileged Android app running on commodity phone hardware have triggered doubts about its value and marketing claims. Commenters largely see using Android as technically sensible, but argue the device offers little that a smartphone app couldn’t, raising questions about whether the bespoke hardware is just a gimmick in a crowded LLM market. More broadly, the exchange reflects skepticism toward a wave of AI gadgets like Rabbit and Humane’s AI Pin, whose cloud-dependent assistants, weak real-world utility, and unclear business models may leave early buyers with expensive e‑waste once the hype fades.

Android under the hood

  • Commenters largely say it’s unsurprising the R1 runs AOSP on a phone-style SoC; many similar devices (Quest, TVs, Portal, etc.) do the same.
  • Using Android is seen as the only practical option for battery-powered, networked gadgets due to drivers, power management, and time-to-market.
  • The “gotcha” isn’t Android itself but that the whole experience is essentially a single launcher-style Android app with elevated permissions.

Why hardware instead of an app

  • Many argue the R1 “should just be an app” on phones people already carry; the device is effectively “a phone with fewer features.”
  • Counterpoints:
    • An app would be lost among countless LLM wrappers; distinctive hardware creates hype, differentiation, and avoids app-store fees.
    • Controlling hardware allows privileged permissions, deep integration, and an always-on assistant without OS vendor limitations.
    • Some people specifically want standalone gadgets and dislike depending on Apple/Google ecosystems.

Scam vs just weak product

  • One camp calls it scammy: marketed as a revolutionary “AI assistant / OS” while being a locked-down Android phone, overlapping existing devices, and arguably misrepresenting what’s new.
  • Others insist it’s not a scam but an overhyped, early, gimmicky product that honestly ships what you ordered, just with underwhelming capabilities.
  • Disagreement hinges on whether exaggerated marketing and broken promises cross into “scam” territory.

Usefulness vs phones and other devices

  • Many see no clear use case: it does nothing a smartphone (or smartwatch) can’t do more conveniently, with better hardware and apps.
  • Supporters like the idea of a less distracting, single-purpose AI device, potential accessibility benefits, and a “toy / desk trinket” with interesting design.
  • Several note current assistants (Siri/Google) already do most of what R1 promises and will likely integrate LLMs soon, threatening this category.

Business model and sustainability

  • R1 is sold as a ~$200 one-time purchase including a year (or more) of AI access; commenters question how perpetual cloud inference and “Large Action Model” features can be financed.
  • Many expect it’s subsidized, sold at a loss, or banking on low long-term usage and hype-driven sales; some predict the service may shut down, bricking devices.

Broader concerns

  • Worries about e-waste, trend-chasing (following NFTs), password handling, and long-term support are common.
  • Some still welcome experimentation in AI hardware to challenge the Apple/Google duopoly, even if R1 itself is seen as half-baked.