Siri, Alexa, Google - Who's using and having good experiences with voice commands?
Hi there,
Big companies are spending big money on allowing consumers to control everything from their TV, smart speakers, lights, microwaves, etc with voice commands. But do people really want that?
I'm curious because I very much do NOT want to interact with computers vocally. I find it's slow and prone to error, not unlike writing out long passages on a smartphone. It functions, but it's not a great experience and remains a novelty for me.
Bought both kids Echo Dots which is a smart speaker with Amazon's Alexa assistant. It's cool, really innovative. But after a couple weeks both remain unplugged in my kids' rooms - totally lost interest. The only person who I've seen really use a voice speaker is an elderly gentlemen for whom it was the easiest way to interact with a computer (using it specifically to play music and get news).
My friend swears that teens commonly use Siri on their phones to look up information. I can only think of one person I've seen actively use the voice controls and that was an older woman who wanted show her friends how Siri could "rap."
So I was curious how many of you find yourself frequently interacting with computers via voice commands. I personally feel very awkward and self-aware and get easily frustrated because it reminds me so much of the terrible automated attendants on 800 numbers.
Edit: You all are confirming my suspicions. Anyone under age 25 use the voice commands often? I feel like this is all just the first step in designing AI interfacing
I use Siri mainly to call or text people, usually just my partner.
sends
I love her. Okay, off to a bad start. I get it, it's literal, but it's weird to sayto a computer.
half the time Siri will respond
Okay, calling your wife. ... Thanks, Siri.The only good marks Siri gets are in the privacy dept. I trust Apple to keep my information private. That's probably the main reason Siri sucks compared to everyone else, they don't data-mine everything.
We also have Alexa. We have 3, a normal sized one in the living room and a dot in the kitchen and bedroom (1 story house). She's always in listening distance. It weirds us out sometimes when we're talking about the sort of stuff we don't want Amazon to hear. The best thing she does for us is control a few smart switches, we mainly use her for weather, to turn off the lights when we go to bed, and for when we're leaving the house (she leaves on the living room light and plays NPR for the dogs). No complaints (well, the routines are a little lacking functionality-wise), but no praise necessarily either.
I don't trust that Amazon isn't listening in. People say "she doesn't send data about your conversations to Amazon, the only time she is connected to the net is when you speak her name." Isn't it obvious that they could use that time to upload any other data, then? Or when she periodically calls home base? The data transferred could be very small since she does at least some processing on the device itself. So far, though, these privacy concerns have been unfounded.
I have a Google Home mini in the bathroom that we use strictly for listening to music while showering. It's plugged in through a light socket so it turns off when we turn off the bathroom lights. I got it back when we only had one Echo and I wanted to see which direction I wanted to take. It does its job. No issues with moisture either, and we got it used. Had it for probably 8 months? I don't use it very heavily so I don't have much to say.
I haven't owned a HomePod, but I'd love to have one ... again, if Siri wasn't worthless. Plus I don't know if our Kasa hardware (TP-link Home Auto.) would work with HomeKit. Can't afford a $350 glorified Bluetooth speaker. I even have an Apple TV, but I'm pretty sure HomePod doesn't do any voice control with the ATV, and I already have a 5.1 system hooked up so I don't need a wireless speaker for my TV. Pretty upsetting because otherwise we're a 90% Apple family, only non-apple device is my daily driver desktop which is mainly for browsing and games. I would love to have a more versatile Siri experience so we could move to HomePods, but my previous points make it out of the question for a few years at least.
Ninja edit: I never touched on your point about the experience.
I regularly think to myself, as well as voice to my partner, that I hate needing to speak at stuff. I hope that soon we can either somehow get more context-aware actions that emulate knowing what I want, or have some better shortcuts to actions on devices, or something. Alas, voice seems to be the late 2010's trend. I have a feeling it will stick around for a while, but I'm hoping it is either superseded or complemented with other choices soon.
I'm hoping something like Shortcuts in iOS will be able to do everything people currently want from these home "assistants," but based on your phone, not your voice. No one is ever very far from their phone. It should be the universal remote of the home. It already can know that you're home from work, so an ITTT-style shortcut could automatically adjust the thermostat, turn on lights, play music, turn on the TV to the news, etc. without needing voice input, or any input at all other than the initial one-time setup, which could be wizard-based for simplicity.
To answer OP's question, I don't voice interact with anything but my phone and my Apple TV. Siri sucks for most things other than weather, but the search on ATV is actually pretty good. I don't have it listening for "Hey, Siri," I use the microphone button on the remote (and your phone can act as your remote, if you're so interested). You can ask for any show or movie and it'll find it, and every service that has it, and give you a choice where to watch it.
You are concerned that Amazon may be listening you. Aren't you concerned that Siri is doing the same? I mean it uses the same concept as Alexa, doesn't it?
They both uses voice commands to know when activated and respond to you, right? I might be paranoid, but I feel that if Amazon, Apple or Google really want to listen to our conversations they would just do it regardless of our permission.
Following this thinking I do use Google assistant on my not-so-new phone, because at least I have some useful feature.
I find it a bit unresponsive, but I'm sure it is because my phone has this weird tendence to fall on the ground when it slips from my hands and I'm pretty sure it damages it.
I do find the assistant useful when I'm cooking, because I can set a timer without touching the phone with dirty hands, but that's basically it. I still find easier and faster open a tab on the browser and searching from there. I should also add that I'm not an English native speaker, so voice recognition might be not as up to date as the English one.
It's been proven that they do not listen 24/7.
Take Alexa (echo) for example. There is a unit that just listens for the wake up word and only responds to that word. Unless the prelim system hears the wake word the device stays shut down. The system that responds to voice commands is separate from the system that listens for the activation word.
They are not recording constantly.
Now convince me it isn't trivial to hack into the device and turn it back on if you work for three letter agencies. ;)
I disable all of mine. I might consider using voice activation if it worked really well, but others have pointed out the shortcomings elsewhere in this thread. I find I'm more annoyed that all the remotes now have five or less buttons - which is a good thing - but one of them is the 'listen for command' button most of the time.
It's actually, basically impossible. There are a few youtube videos out there that go into it in detail but the "wake word" system and the actual Echo are basically two separate devices in the same shell. The Echo cannot interact with the "wake word" system and that system only listens for the single word. They have done a good job to ensure security and privacy.
Now, that being said, everything you say to her after you have awoken her is 100% sent directly to Amazon. But, that is under the premise that it is used to better the technology.
You can be paranoid about the NSA, CIA, FBI, Homeland, ect turning on your device all you want. But the reality is, that device having or not having voice commands has nothing to do with it. They are more than capable of just using the microphone in your phone or computer if they wanted, Alexa and her friends have nothing to do with it.
To ignore modern advances in technology because of conspiratory theory paranoia is just foolish.
I set up a dual local network at home just to put a stop to all the chatty modern equipment. Computers and devices that have a need for internet access go on the vlan and they can get it. Everything else that's just plugged in can see the rest of the network, but can't get past the firewall. Seemed like the easiest way to make sure nothing I buy is doing anything it shouldn't be doing.
I'll also take some issue with that, since there have been many cases of vendors sending back data, scanning the network, even connecting to vpns. We're well past 'theory', though I doubt it's anything intentionally nefarious yet - just overzealous vendors/developers with a 'collect everything' mentality. I don't expect to to get better either, only worse and more invasive, all in the name of making things 'easier' to use.
It's not so much the "they're still listening" bit that is conspiracy theory-ish, it's the "send it back to intelligence agencies" part.
Which may or may not be possible with smart home devices, but seems inconsequential when compared to the fact you have a microphone on you 24/7 as long as you own a cell phone, which is probably a more tempting mic to try and hack than a limited use smart speaker.
Kind of a tangent, but calling it her sounds way more dystopian to me than anything else in this thread.
Oh please, people have anthropomorphized stranger inanimate objects. Boats, cars, fishing rods, the weather, vacuum cleaners...
Anthropomorphizing objects that don't monitor you or talk to you is fine. But when you do it with Siri, Alexa and so on you end up with this and if you don't find that creepy and dystopian... well, I don't know what to tell you.
Yea and at least this little black box talks to me.
I work for a company that builds security systems, they have cameras you can have in your house, the operator cannot access the camera without you inputting an access code via your app or the hardware in your house, first they have to request access, then you have to approve it, before they can actually look at the footage, I know exactly how it works, how it's extremely secure, but I STILL don't have one in my house.
It's a bit silly, considering fucking everything comes with a built in camera these days, but there's just something about it always watching that I can't forget about, it's teh same with those listening products. Even though I'm certain all the big players are listening through my phone the whole time anyway.
That might stop it from turning on without the wake word but what's stopping the system from staying on after a command and continuing to listen? I'm not trying to be confrontational, I'm legitimately curious.
The system audibly confirms when it is shutting down after performing the requested function and the light on the unit shuts off.
Sure, one could theorize that it stays on longer than that, but there really is no reason to think it does.
If your phone was recording audio or video and sending it off to some remote location in any significant way, it would be almost impossible to hide, and someone would notice it pretty quickly. Lots of people monitor their home networks and watch for specifically this sort of thing. Even if the phone was programmed to only transmit this stuff via the cell network, infosec experts who test these things, and counterintelligence agencies of the world (the real people the NSA wants to fool) would pick up on that in a heartbeat as well.
The only way this kind of widespread mic or camera recording would work is if the phone was programmed to only connect to a secret network of government operated wireless gateways that nobody knows about, using a super secret spread spectrum waveform that nobody can detect. And if they were going to do this, they wouldn't just tie it into a voice command service.
I can sum it up very easily.
I provide audio data because Google provides me with convenience. Google Home rarely lets me down when I ask a question, but I have reasonable expectations for what Google Home can do. We use it to play music, tell us the weather, do unit conversions, play videos, look up specific pieces of information, or showing pictures. All of these things are simple and easy to do, and they can all be done by anyone in the house - if my four year old son wants to listen to music, he can do so with no oversight from me.
Generally speaking, most people opt in to things that you might not like for the sake of convenience. Facebook is a great example of this. It's a very convenient service.
When I bought my first smart TV I've opened it and physically removed the microphone. It defaulted to voice input every time, and recently it's been updated with an always-listening Google Assistant, so that was the last drop.
I haven’t bought a TV since they all became “smart.” Are embedded microphones standard now? I don’t mean to invoke 1984 lightly but that’s getting pretty close to the idea of a telescreen.
I’ll probably be in the market for a 4K TV soon. Can anyone recommend one that’s just a display with some standard video inputs? The last thing I want to do is give it access to the internet (or have it constantly nag me about that if I don’t).
I think your best bet for that kind of thing would be looking for giant computer monitors rather than TVs. I had a very hard time finding a TV that wasn't "smart"
I dunno where y'all are shopping, but places like Walmart have plenty of non-smart options. They're insanely cheap now too, I just saw a TV that was over 50 inches wide selling for under $250.
But are they 4K or are they 1080p? I've only ever saw smart options for 4K running either Android or Tizen. Buying a 1080p TV in 2018 is not a good idea, it's not really future-proof, and the picture doesn't look as good. But you can make a more informed purchase if you'll do some research on the software that your future TV is going to have. Android is preferable because it is the most open and flexible option.
Oh yeah, I have virtually zero interest in 4K, to tell you the truth so the majority were 1080p. That said, there were a few 4K ones in the realm of $500 that didn't appear to be "smart," however I could have missed the label. You'd know more about the difficulty of finding those than me haha.
Yeah, microphones are pretty ubiquitous in smart TVs, mainly because it's a chore to enter any text from a TV remote. But don't be scared to get a SmartTV - you just have to get the one with Android TV on it, it allows you to sideload apps, and you can install great Android apps to it to remove all the shitware. It's more productive to find a TV you can root than a TV that isn't smart. In my case I've got the cheapest decent 4K TV I could get, and it's really nice. I keep it behind the firewall at all times, but I just couldn't block google domains without blocking youtube as well, so I had to remove the mic. If you aren't going to use YouTube on it you can blacklist it from going online at all.
You can find that in the form of specialty commercial displays -- things made for airports, offices, etc. Here's LG's line, for example. With that said, they're expensive compared to consumer panels, and aren't subject to the same reviews and criticism that "regular" ones are.
What I might suggest as an alternative is getting a "smart" one, getting it on the latest firmware via a wired connection, and then cutting it off from Ethernet and have it default to whatever set-top-box/console/etc. you want.
I have carpal tunnel, so yea, voice commands aren't really a choice for me. This is limited to mobile phone usage though - I don't have any smart home style tech.
Not just for free, people PAY to own these wiretaps. Baffling indeed.
I see a lot of Google, Siri, and Echo stuff.
I can't trust any of those companies with huge amounts of my voice data, to be honest.
However, as I'm slowly losing control of my hands, and programming is kinda my career, I've been looking more and more at these various systems.
At the moment, I'm trying to teach Mycroft.
It's open source, you can buy a unit, or install on Linux, Android, or a custom version just for a Raspberry Pi. (No Windows or macOS yet, but maybe in the future).
You can pair it with their online database, for better and faster word processing, like the big boys... But you can also choose to do all that locally.
Being able to more easily add abilities to the assistant is also a huge plus. Modify, deploy, no negotiating with strange and terrible SDKs, no licensing.
I'm definitely checking that out. Great to know there is an open-source approach being worked on.
I have a very similar experience to yours with Google Home.
You probably know this already, but you can wrap up that last batch of commands into a routine as well! When I say "Good night" to it, it knows to turn off the lights, tv, set an alarm for my usual wake time, and then play ambient sounds on the bedroom Mini.
Additionally, the continued conversation feature is a game changer in my opinion. Basically after it answers your first query, it stays listening for about 5 seconds, and if it picks up on another command/query, it'll do that too without you having to say "Hey Google" again. Makes for a more natural conversation.
Very interesting! Thank you for sharing. My understanding is that Google has the best voice recognition of all of them, but I've no doubt that they all improve over time.
For a while I used one voice command: "OK Google, what's the temperature outside?" Sometimes it would actually tell me. But it never worked consistently, and I ended up wasting more time than I would have if I just tapped on the weather app. Fuck it man, I ain't got time for that shit.
I can see how it would be useful for people with certain disabilities though.
I've noticed that Google doesn't know what precipitation is. "Ok Google, what's the chance of precipitation tomorrow?" doesn't work, but "Ok Google, what's the chance of rain tomorrow?" does. Which gets annoying, as precipitation covers rain, snow, sleet, etc. Rain covers 1 thing.
I have an Echo with Alexa in my living room, she controls basically everything in that room. My living room lights, thermostat, and tv all respond to voice commands via the Echo. It's super convenient to come home with armfuls of grocery bags and have the ability to say "Alexa, lights on", I used to think it was dumb until I tried it out and it's awesome. I can turn on the lights, dim them, adjust the heater and put on my favorite show with out touching a switch or remote. It's pretty awesome honestly. I also use her for spotify, news via the echo's "flash briefing" feature and to find out misc local information like traffic and weather.
Eventually I would like to add a roomba, kitchen appliances, security system and door locks that are voice enabled. Go for the full smart house feel.
I know a lot of people have problems with voice command responsiveness, but I haven't had any major issues other than occasionally having to repeat the command because she didn't hear it or I spoke to fast.
For anyone that is distrustful of the proprietary voice data owned by Google and Amazon, I recommend they check out Mozilla's Common Voice.
Mozilla's open source project, Common Voice, is well on its way to becoming the world’s largest repository of human voice data to be used for machine learning.
https://opensource.com/article/18/4/common-voice
How do that compare to Mycroft?
Mozilla has more user data than Mycroft. Mycroft once considered using Mozilla's data.
More info here: https://mycroft.ai/voice-mycroft-ai/
Thank you!
This is very cool and very easy to contribute to! Thanks for putting it on my radar. I've set a goal for myself to speak and validate a few sentences every day in order to help out.
You're welcome my friend. I love to participate in these little citizen science-like projects. They actually have an iPhone app which makes it super easy to contribute: https://itunes.apple.com/us/app/project-common-voice-by-mozilla/id1240588326?mt=8
I use it on occasion. About 20% of the time I use it, my final response to it is, "You're an idiot." I think the last time I used it (Android), I asked, "Tell me if it's going to rain." If it had said, "I'm sorry, but I lack that capability," I would've thought, Okay, fine. I'll just pay attention to the weather (I had put clothes outside to dry). Instead, it issued a web search query with that text. I tried again: "Notify me if it's going to rain." It created a notification called, "If it's going to rain."
I might try it again once I forget about that interaction, but it would take a lot to impress me. The assistants seem to be very limited.
I also simply don't like speaking aloud when I don't need to. Not exactly a popular perspective in my marriage, but I tire of vocalizing the humdrum of daily existence. I like talking as part of conversation, but the rest I find to be a waste of effort. "Do you want a big plate or a small plate with dinner?" This is about the worst question I've ever been asked.
I have only used the Google assistant because I have heard that Alexa is not quite as good and Siri is apparently terrible.
And guess what? The Google assistant is still terrible. It's not the best at understanding what I say, but the worst thing by far is how stingy it is about listening to you. You have to wait after you call for it so that you can hear the chime to know that it is listening, but it sometimes takes a second or two for that to happen. That extra time listening is just enough to forget how you were going to phrase your request. And having that phraising down is vital, since if you pause for over about a second it will stop listening to you!
It's also really bad at understanding your grammar. If you give it a question with more than two things for it to understand, it will probably ask you about something you just told it. then there are the times it simply doesn't understand the words you are trying to tell it....
From my experimentation, the only function that the Google assistant is good for is mad libs.
I just start talking, very rarely does it not start listening in time. I also don't have the chime enabled, so that might have something to do with it.
I have the chime enabled on mine, it doesn't matter. I've fired off entire commands before the chime, and it still responded correctly.
As long as it catches the wake phrase, it'll start listening immediately.
I don’t like voice interfaces because if I have to enunciate clearly and repeat myself multiple times, I’d rather just type the damn thing.
I've used Google home to reasonable success. I like that it can take multiple commands in one go, like, "hey Google, set your volume to 75% and tell me the weather tomorrow and play heart of gold." And the majority of the time it works. I use it a lot whenever cooking to get temperature conversions and weights from American recipes that insist on volumetric measurements for things not easily measure by volume. I use it in the garage to help with unit conversions and calculations as well. The broadcast feature is great when I need my wife's help in the garage and I don't know where in the house she is.
Recently I got some smart lights that I can control by through Home, "HG, turn on the nightlights" and I get 10% lights on from my bedroom to the baby's room. Great for late night feeding.
I use my google homes constantly. We have smart lights in every room of the home and often use it to check temperature outside. The original google home in the bathroom has a decent speaker and plays music while in the shower. The mini in the livingroom/kitchen is vital for timers, conversions, and even listing out ingredients/steps for a recipe (though I'm still searching for a way to add my own recipes). It's also good for playing netflix/youtube/music on the chromecast in the living room. The mini in the bedroom controls lights, playing music on the chromecast enabled speakers, and playing netflix/youtube on the TV in the bedroom. I even periodically use the google assistant on my phone to broadcast when I'm going home and ask about weather when I'm already on my bike (I just have to hold down a button on my headphones).
Pssst, I'm also under 25.
The thing is, I generally won't use voice commands at all when I am out in public, but around the house and especially in the car, I find Assistant to be one of the most compelling reasons to use Android. Simple stuff like playing music or setting kitchen timers I can do without diverting by broader attention, just by yelling at my machine. When I take naps, I love being able to just mumble "ok google, set a timer for one hour" as I doze off, rather than forcing myself to sit up and make a timer. In the car, it's super nice to be able to just say "ok Google, find me a gas station" or "ok Google, play my podcasts." While I'm on the sofa, I love being able to adjust the lights and control playback by voice.
Overall, I was in a similar boat as you where I found the whole thing a bit creepy, but I think it really does add value to the way I interact with my devices, that I'm over the initial repulsion.
I'm not sure I felt it was "creepy" ... more "cringey" but I realize that's more how I feel and less how others perceive the interaction.
I use Alexa mostly to set timers and play music and occasionally ask about the weather.
I mostly use Google for driving directions. Being able to say "OK Google, take me home" and it automatically giving me directions is pretty cool.
When you're using Google to do that, is that on an Android? I'm all about hands-free technology in the car.
Yeah, Android on Samsung Galaxy S8+. I've got it linked to my car via Bluetooth, so it gives me directions through my car speakers. Sometimes I'll mount the phone on the dashboard too so I can see the navigation without looking down.
Oh yeah. get a dashboard mount. They rock. I use Google maps all the time w/ the sound disabled so I can glance over at the directions.
Care if I pick your brain on the S8? I am a slow adopter - still rocking a iPhone s5 - and I'm considering buying a S7 or maybe an S8 next year as I do not like the Apple ecosystem. Can I easily transfer music and videos to the S8? As in, plug it in to the computer via USB and drag-and-drop MP3s and MP4s in the appropriate folder and have that work? Or do I need to go through an iTunes-like interface (shudder)?
I think on pretty much any Android, you can drag-and-drop media. I just copied some music folders from my computer onto my phone and it worked fine.
I used to use Siri for turning on low power mode until they added a way to turn it on in the bottom menu. Now all I use Siri for is to flip a coin. I often have to say it twice as Siri just doesn't get it the first time for no reason.
I use my GHome constantly. I don't have a clock in my bedroom so I often ask the time when I wake up. I'll play music when doing chores, and set timers when cooking. I'm starting to play with home automation like turning on the AC or heater with voice. Maybe lights next?
I find it extremely convenient as an additional input mechanism. Being online-only though, I am careful to always have a physical analog just in case my internet goes down.
Very cool! Thanks for helping me better understand how people are using these devices.