This is just my opinion, based on my experience and personality.

Voice first running entirely on prem with no internet connectivity, sure I would go for that.

Talking to the internet? No way. Not even if I had root on the device with the source code to the OS, firmware and applications.

> Talking to the internet? No way. Not even if I had root on the device with the source code to the OS, firmware and applications.

Can you explain how this is different from running your linux computer with a webcam attached?

Do you really think the people who are concerned about Google Home/Alexa/etc don't also have their webcam covered with a sticker?

I don't leave a webcam attached, or a microphone.

I can also debug any code running on my machines and have some idea of what they are doing. I can also limit when they are allowed to talk to the internet and to what address.

> Not even if I had root on the device with the source code to the OS, firmware and applications.

Can't you debug code and control network access in that scenario?

Maybe if I have the debugging symbols and kernel headers. Even then there are circumstances where it won't help.

For example: Nearly all cell phones check in to their manufactures website from time to time. Nearly all of them look for a header that triggers a debug mode that turns on CarrierIQ. (It isn't called this any more, AT&T renamed it to an empty name to stop people from searching for it).

Trying to find obscure functions and understand how they interact with the servers they check into can be very difficult at best. It often leads to more questions than answers, at least based on my experience.

I guess you’ll have to wait another decade or two.

In the meantime, I’m “feeding” my devices every day.

Hopefully voice-first becomes a solved problem, and you’ll get your device sooner.

What's funny about that is, I had voice recognition software in 1991 that could start up any application on my computer, reboot my computer, etc. Surely there must be a small device with the processing power of my old 386-DX40.

I believe it was made by a company called Covox or something? There were several other companies that made similar software back then.

Could your device recognize multiple speakers, with varying accents and grammar style?

Yes. Not very well, but yes. Each person had to say a few sentences. This is fine for home use unless you run a halfway house.

Certainly not as well as the software we implemented to listen to phone calls in the background on wireless networks in the mid 90's. That was the wildfire project and was fun. We abandoned it, since our lobbyists were able to kick the hands-free laws down the road.

>Voice-first devices are long overdue.

Sure, that's not really anything that anyone disagrees with. It's the acceptance of data collection and exfiltration that is odd.

I would disagree. I despise voice interfaces and despite working for one of the major tech companies that build these devices and interfaces, I use exactly zero. I do trust my employer to do it correctly based on the processes I witness every single day, but I don't own any because I don't like voice interfaces. Remote control on my phone, sure. Clicking the button to change a song, sure. But no voice, thanks.

You dislike the very idea of voice interfaces or the current state of voice interfaces?

Maybe a naive question, but how are they any worse than having phones with non-transparent hardware/radio drivers in your living room? We seem to have accepted that risk, and even carry phones around everywhere.

We use smartphones for the same reason we don't avoid any non-free software like Stallman: a huge amount of additional possibilities. But what Alexa, Siri etc. offer is just a different way to use a limited set of existing features, in exchange for a much higher risk of a privacy breach.

Everyone needs to draw a line somewhere, I just don't see a reason to use voice assistants, definitely not with internet connectivity and where the company keeps all recordings for some reason.

They keep all recordings because they train their voice recognition off them. GOOG-411 kept all recordings too, for the same reason, as do hundreds of other services, but I don't recall anyone caring about them.

Google gives you a UI to view and delete your recordings and suddenly they're monsters.

Has anyone on HN put their phones into Airplane mode with WiFi on and then had wireshark tracking all outgoing packets from the wifi devices in their home?

To see exactly what outbound connections were made by which device and which app when either device was being actively used, or laying dormant?

This is how it was discovered in the UK that SmartTVs were sending viewing data....