Hacker News Reader

> Identify or find objects with Alexa+.. use the camera on Echo Show devices to identify objects and get help with daily tasks. For example, reorder pantry items by showing them to the new Alexa, ask Alexa to identify a type of plant and its care instructions and set reminders for when it needs to get watered, or get fashion or décor ideas. This new feature is particularly game changing for people who are blind or have low vision

This has so much potential, but it will require a workflow for Alexa to learn about specific objects and layout (of multiple objects) within the customer's home. Apple's "live audio descriptions of video" had similar promise at launch, but hasn't evolved beyond the launch demo, https://support.apple.com/en-ca/guide/iphone/iph32deb9296/18.... Could Alexa+ enable self-service RLHF on home video/images?

It's a testament to the latent market opportunity that Amazon has sold 500+ million devices, despite the obstacles that greet customers trying to customize Alexa for their specific needs. With open developer interfaces, Alexa could have been the "IBM PC" of voice AI, instead of just another walled garden.

Alexa use cases for elderly: https://news.ycombinator.com/item?id=41062989

In theory, Home Assistant voice hardware could be integrated with local LLMs for private voice control, https://www.home-assistant.io/voice-pe/

> Fully open software, firmware, and hardware.. Grove port for connecting sensors and a 3.5mm headphone jack for connecting external speakers

View on HN