IMO, it's cheaper and (importantly) better to do it in the cloud. Better in there's steady improvement to customer experience with zero customer interaction, a cleaner developer "apps" marketplace, and each local device can be smaller, cheaper, and lower power.
Not if it implies violating customer privacy. As a company you either have a customer or an employee - you can't have it both ways. There is one way it is ethical: when you have the informed consent of volunteers. That works for charities because there is no money being exchanged to corrupt the interaction.
IMO the only appropriate context for gaining training data is when you're explicitly paying people for that.
Informed consent does not mix with volunteering to make money for a private company. That people are ignorant does not justify abusing them.
Except that fully local tasks now cannot be done fully locally, more data than needed for the task is leaked to the service provider, and the whole thing becomes a centralized ordeal. It turns into a single service what should be a product giving access to mulitple services. It offers degraded user experience due to latency (reason #1 I don't use voice assistants, #2 being not much useful to be done with them if I can't trivially write code gluing them with stuff I have in my house), stops working completely when there's not Internet connection (Internet isn't and can never be as reliable as power), and I'm doubtful of how significant power benefits are once you account for the power that's used by various routers to transmit your voice to the cloud and get the results back.
Power and cost benefits from the perspective of the end user. I have a device that draws under 10W that is (as far as I'm concerned) capable of world-class and ever-improving voice recognition and interaction. That a service provider or set of service providers are providing me temporary use of kilowatts of computing computing power that I don't ever see a bill for is irrelevant from the perspective of "is it worth me paying $40 and $5 of electricity per year to have a tiny voice assistant?" vs "is it worth me paying $500 and $75/year in power for a fully local, larger, and inferior solution?"
It's unsurprising to me that the market evolved towards the former as the latter would almost surely be a commercial failure.
I'm not surprised by the way the market evolved either, but I consider it a bug, not a feature. I'm a firm believer that if something is to be done, it deserves at least an attempt at doing it right. The product version would be a commercial failure only because it has to compete with the service version; the story would be different if the service version wasn't allowed to exist in the first place.
Also let's not overestimate the involvement of the cloud in the actual voice recognition. With ML models, it's the training of a model that's expensive, not the actual use of that model. I don't see a reason why voice models, once trained, couldn't be run on the device, at a slightly increased cost of compute (of which phones have plenty). It would be fair and good engineering to provide models in exchange for payment.
And no, I don't buy "it's not that simple", because I've worked with MS Speech API over a decade ago, on much slower hardware and on a shitty microphone I soldered myself from parts, and it did work well enough in "fixed grammar" mode after a bit of training with various sources of background noise (read: very loud music). I don't believe there's any technical obstacle; the decision to run this in a cloud is a purely business one, and like many modern business models, this one is geared towards extracting value from purported customers, not giving value to them.
> would be different if the service version wasn't allowed to exist in the first place.
Agreed 100%. However, I'm very reluctant to use laws to prevent people from choosing the things that they decide are best for them on the premise that I/we know better.
The hurdle for my determination that I know better what's good for someone else and to restrict their choice to conform to my superior knowledge is extraordinarily high.
That's a fair point. And although I didn't explicitly said "laws", I guess I meant that.
My view comes from being tired of seeing how the market forces always incentivize the most scummy garbage that can be gotten away with. The only control point we have here is the "gotten away with" part. I'm not very into telling other people how they should live, but a stable and happy society does require some of that.
(That said, I currently have very little influence over this, so I resort to what little of a market signal is created by ranting about this publicly, voting with my wallet, and discouraging people from using solutions that are user-hostile.)
