Amazon (AMZN.O) is planning a major revamp of its decade-old money-losing Alexa service to include a conversational generative AI with two tiers of service and has considered a monthly fee of around $5 to access the superior version, according to people with direct knowledge of the company’s plans.
What info are they getting from me telling it to turn on the lights?
The service it provides I would expect to either pay a reasonable marginal fee, or do everything locally.
If the Home Assistant voice Appliance stuff can get its shit together and I can get one for reasonable prices I will move to that (or something like it) instead.
More than just “ripcord likes to have lights on at 6:00 pm,” surprisingly.
It knows what brand lights you have, who’s interacting with it, who you might be with if anyone speaks in the background, what times and days you’re typically home… it’ll even infer your mood based on how your voice sounds.
Unfortunately, Amazon isn’t required to disclose every bit of personal data they take from you, so only so much is known about it. If you consider though that data collection is a new, multi-billion dollar industry, and how effective hundreds of PhDs in data science and social-engineering can be with near infinite resources to develop tools to extract as much information from these devices as possible, it starts becoming more believable.
Here’s a good paper I found: http://arxiv.org/pdf/2204.10920
it really depends on how much you trust amazon on what it records as alexa is an always on(in terms of microphone) device.
It shouldn’t take a subscription to manage turning on lights.
You can very easily do it locally.
With voice control?
Yes
https://www.home-assistant.io/voice_control/
Which is why I said
Unfortunately, when I looked most recently it still wasn’t even remotely close to being ready. Particularly the hardware options.
They say that you can build one for $13.
https://www.home-assistant.io/voice_control/thirteen-usd-voice-remote/
They also have on their roadmap that they’re working to see if they can build or engineer
outor whatever an all in one, easy to set up voice satellite hardware as one of their next up priorities.https://www.home-assistant.io/blog/2024/06/12/roadmap-2024h1/
Once the second thing happens, assuming it’s any good, then I will look into switching again.
Until then, there don’t seem to be many great options.
You could also argue Apple is heading in an interesting direction with on-device AI. Im ready to switch to Apple TV for fewer ads, as soon as they release a new version capable of on-device AI
I agree. I do keep considering it, but the additional value to me right now vs cost hasn’t been worth it.
Same with moving from Roku to Apple TV.
Also, not having much in the Apple ecosystem is a factor. Down to just one occasionally-used Mac (and other macs that just serve as servers in the homelab)