Google debuts Google Assistant SDK in preview, letting developers run Assistant on their own hardware
Ron Amadeo / Ars Technica :
Context & Ripple Effects
A year after Google unveiled Assistant as the successor to Google Now, the company is handing developers the tools to run it on hardware Google doesn't make — turning a phone feature into an embeddable platform. Weeks later, Google would lay out its plan to make Assistant ubiquitous against Siri and Alexa, both of which hold strong incumbent advantages (The Verge's breakdown of that strategy).
The SDK is the distribution half of that plan: rather than waiting for users to buy Google-built devices, Google lets third parties ship Assistant inside speakers, appliances, and gadgets of their own design. The related coverage shows this was only the start of a multi-year build-out.
First-order effects
- Device makers gain a supported path to ship Google's conversational assistant in their own products, giving Google a presence in hardware categories it doesn't manufacture for.
Second-order effects
- Amazon's Alexa faces a licensed rival in the same third-party hardware channel where it had been running largely unopposed, and Google followed through by expanding the SDK to Australia, Canada, Germany, Japan, and the UK with device-specific smart-home voice control (the December 2017 expansion).
Third-order effects
- If the pattern holds, voice assistants consolidate into a small set of licensable operating layers — Google, Amazon, Apple — with hardware makers choosing a platform the way they once chose mobile OSes, and service discovery flowing through whichever assistant owns the device.
The trend: Voice assistants are evolving from built-in features of first-party devices into licensed distribution layers that third-party hardware carries into the home.