HA has a voice assistant, but itâs not in the same league as Alexa. Maybe someday but itâs a long way from that today.
Back before 2020, SmartThings also let you device what devices you could share with Alexa. That was back when SmartThings was so much better than it is today. Just before they got rid of the faster, more reliable app, they released an update to the Alexa integration that got rid of the selecting devices. It was a botched fix to some other bug they had.
I vaguely remember some SmartThings devs somehow thought it was better because it got rid of the chore of selecting devices to share.
There was a huge uproar on the ST community. It was ignored ("those are just power users, most of our real users don't care" was the response).
It was one of many small cuts that made me stop using that platform.
I think part of the question is what do we expect from a voice assistant. There are a lot of potential use cases that have nothing to do with HA or Hibitat. Strictly from a controlling HA setup it shouldn't need much. It just needs to be able to control decices or look up entities details and reapond. From what i saw that looked very doable.
Exactly what it should do, the problem I saw with even the latest version of it, you would have to repeat multiple times to even wake the Voice Assistant up, then having it do what you want was also very painful. I have no doubt that it will eventually be working well but for now if you are having problems with the 3 major connected ones, you don't want to go the HA/Local route right now. Just saying this based on what the OP asked for, if you want to try it out for yourself, go ahead, could be great as a second VA with another major one for now. Keep the WAF high and permit you to test it out without involving the rest of the family that might not tolerate it not responding multiple times in a row for a simple light to be turned on.
Is this just conspiracy theory stuff or can you provide reputable sources for this information?
I'll let @SmartHomePrimer offer his own view, but I would expect the likelihood of peer-reviewed academic-based analysis of this would be unlikely or at the very least limited, also given the commercial implications involved... But I would suggest that anecdotal evidence at least would support changes in the responses from these organisations after certain phrases or opinions are expressed....
I can't blame them, that's what their business models are based on...
I donât doubt that Google, Meta, etc., do exactly that, but the comment about our phones âspyingâ on us (while sitting idle on the table next to me) is more my point. âŚand âevery word we typeâ even in RCS or iMessages?
Fair enough... My comment did also (in my head) cover changes in responses based on conversations in the presence of a phone... I feel like that has also happened for me with a Google phone
But you are right, I at least have no reputable source for my claim, if I am not considered reputable ![]()
![]()
When it comes to Hubitat, I find you very credible!
I just find it hard to believe that, it covert phone eavesdropping is a fact, someone with the appropriate âsnifferâ devices hasnât been able to show it to the public. Same with âend-to-end encrypted messagingâ.
Thanks. Yeah, I'll tap out when talking more generally about mobile device security. I'm sure others here are at least more confident than me to comment..
Based on what. All i see above is the OP asking for alternatives, and then what looks like a post someone got flagged and hidden which may of had more details about why. I can infer that based on the conversation there is some concern about being used for training AI or his info being used to do so. If thay is the case why not look at a solution were the OP could use a LLM locally with a small computer with a built in GPU or a jetson Nano.
I have a Raspberry Pi 4 sitting around and have been dabbling with running ollama on my server so i think i will give it a try and test it out. My suspicion is the problem may be mostly finding good items to build the device is the biggest problem.
I use the HomeKit integration and Siri since my ecosystem is all Apple. Alexa isn't allowed in my house.
Aww, whatâd I miss?
I also wish there was a privileged group who had parental consent to view flagged content....![]()
On a more serious note, it is a shame that a conversation like this descends into that kind of territory... It is a very pertinent question for those concerned about local control and the large companies involved in these voice-based systems... I hope I don't derail the conversation too much
Iâm guessing the OP thinks that posts by me, @aaiyar or perhaps others are deliberately sarcastic or patronizing, and is responding in kind.
Absolutely wasnât my intention, so I do hope we can get back on track too.
I still havenât seen much of anything from the OP describing their needs, goals, priorities, etc, though.
A violation of the community guidelines ![]()
The English language is sufficiently complex to permit one to make a point without being snarky, rude, or offensive.
Hard to comment with their recent response flagged...
I agree, but would be nice to have some time in "the southern states" to assess that for ourselves.... Not knocking the great work done by those while I sleep to keep the Community humming and in good order, but would be nice to at least get a preview of some of the more fruity content posted overnight for us down here....
Hopefully OP will post again and keep it on topic.
In the meantime, Iâd also like to know more about what @SmartHomePrimer was referring to with respect to âeverythingâ we speak and type being subject to consumption by an AI training model.
âCan and doesâ implies different things to different people.