A few years ago I momentarily set mine to Spanish just to see how it worked. Now, once every few weeks I randomly get a Spanish response. 99.9% of responses are English, just the odd ball Spanish one. I'm never going to file a bug report; it makes me smile. Part of me thinks I keep using it because of this.
I use Siri to set timers almost every day, to play music, check the weather forecast, and to turn groups of lights on and off.
All of these assistants have limitations because they don’t have any real context to the requests. People think that they are failures because they are not really intelligent but they do have small uses on a daily basis as long as you understand the limitations.
The lack of context is exactly what I noticed was the problem when using Google Assistant. I had to be very explicit about what I wanted it to do. It couldn't use context of what I was doing at the time to figure the details out on its own.
True I guess, I know my brother and sister-in-law use Siri to set alarms and reminders and such because they frequently have hands full with kids. At the same time I don't think that they're worth getting a distinct device for (lets be honest if you're buying an assistant device you have either an android or iPhone with google assistant/siri)
An alarm clock that doesn't stand a pretty good chance of randomly broadcasting your bedroom audio to the internet on a regular basis. You can even get ones that work when the power's out!
If you replace "assistant" with "internet connected room mic" when pondering them, I find it helps make the decisions far clearer. The best place for them is never having been shipped in the first place. The second best place is the trash. Down the list from there is "unplugged and forgotten." And somewhere at the bottom is "Powered on, connected, and waiting for a plausible excuse to broadcast fairly clear room audio out to the internet."
Guess what, your phone can also start randomly broadcasting your bedroom audio to the internet. You could call a phone a "internet connected room mic" or "personal gps tracker". Does that stop you from using one?
I've abandoned smartphones for the time being (but have a PinePhone to hack on for sport and challenge - it in no way resembles a functional device at this point in time) and have gone back to a flip phone device with a removable battery that generally lives on the kitchen counter, and has a pretty poor room mic. It gets rebooted (full power cycle with battery removal) often enough, and I frequently don't bother bringing it with me places, or at least bringing it with powered on. I've found since I'm looking at it far less, it falls out of mind, so I will quite literally head into town and not realize I don't have it until I get nearly back home. It's fine, I did that for years before cell phones existed, I see no reason why I can't go back to that.
"No phones in the bedroom" is just a standing good policy for a range of reasons anyway.
However, I generally trust phones a bit more than I trust the "room mic" class of "assistant" devices. First, they're on battery, and widespread abuse will mean rapid battery drain and a warm phone. A phone, warm in the pocket, is cause for either turning your phone screen off, or alarm if it's already screen off and you don't know why it's warm.
But they also have a less effective microphone. Most of the ones I've experimented with over the years won't pick up a normal conversational tone across the room with any accuracy - but you can walk into a room, query for an Alexa/Google/etc device, and the one across the room will generally respond to you. They have far better microphones than a typical phone does for room audio.
I'm aware that the various devices are not supposed to be a live stream of room audio to the internet and generally don't seem to be doing that unless triggered, but the reports of people who have done a deep dive into their triggered captures have been that it captures a lot of nonsensical chunks of audio for no discernible reason beyond, presumably, "When in doubt, capture and send" - ease of use, of course, meaning that false negatives in detection are far more serious than false positives.
But neither do I trust the companies involved to not be doing a range of audio analysis locally and sending the information up in random updates. I don't know that they do this sort of thing, but I don't trust them not to. You could do things like unique voice count in the room detection to identify how many people typically live in a place, or if people are having a party. You could do vocal stressor analysis to determine if people are frequently angry, upset, etc. And I'm sure there are many papers written on what else one can use this sort of stuff for. I object to the capability being present, even if not used. Hence my general trend of not keeping smartphones around me regularly - but even there, at least Apple devices do tend to indicate if something is abusing mic privileges, and I can remove audio privileges from apps. I can't do that with the... skills? whatever the audio helper additional functionality code is called.
I use Siri daily for setting appointments/reminders, adding items to the shopping list, or hitting play on Podcasts because I'm already driving. It's pretty game changing to not have to open a device and punch buttons for a couple/few minutes but rather just say a few words to book an appointment or remind myself to do x in y hours/days.
That’s not too surprising. Alexa is the kind of thing that would be a miracle if it was the only consumer computing device instead of nearly worthless to anyone with a smartphone or laptop. I too just use it to play very random playlists and set kitchen timers. I’m sure there are some people who don’t have the ability to use traditional devices who rely heavily on it. I had an uncle who went blind quite late in life and so never learned how to cope and would mainly sit and listen to the radio. It would have been a wonder for him.
I've used mine everyday since I got it a few years ago. It's mostly for podcasts and news. I've automated a few lights too. But, I've yet to find an app that's useful to me on a regular basis. It seems that there should be one but no luck.
For light control it's very useful and pretty cheap so if I only used it for that I would be happy.
Much of the problem stems not only from the limited understanding capability of voice assistant devices, but also from the lack of trust we have in the big tech vendors backing them. I have multiple voice assistant devices, a few from Amazon, and I not only disabled but unplugged them once I heard about Amazon's efforts to share my bandwidth without my consent. Limited utility plus bad privacy practices aren't a compelling combination.
I have a Google Home assistant and I only use it to play music since it has fairly decent sound quality. My Amazon echo and show devices will stay unplugged.
i ran the max rate lora numbers, for very close by devices, & i dont fully remember but i think it was like 250MB a month. again assuming 100% channel utilization at max rate.
i really am very happy to find fault, to elaborate misuse. but the reaction to amazon using a very little bit of bandwidth (usually under 50MB i expect) feels entirely emotional, entirely reactionary. in any technical sense we can measure it is for all intensive purposes going to be irrelevant.
this feels like a social dilemna. we want to activate ourselves against amazon. want to tear down the big, the seeming man. but this feels petty & small. that we so readily make a stand here says that we arent really discerning or tasteful or making sensible assessments & rationalizations & choices. we're acting in aggressively zero sum hostile premises, & causing a stir over that which is nothing.
generally i think you're larger point about mistrust is dead on. these systems are black boxes, running in far off data centers. this is anti-personal computing, unobservable by the primary stakeholder, & it's a fucking abominable form of computing that i think more & more people realize has no respect for them, that treats them like end users, not masters of their home & it's systems.
The amount of bandwidth being used isn't the concern. It's the lack of control and consent over what is being transmitted. How would you feel if the car you bought granted someone the right to come over once a month just to open your car doors and see what's in your glove compartment? Oh, it's just once a month and they're not doing anything, just rifling through your glove box. This is how I feel about the sharing of Echo bandwidth.
It has nothing to do with the "big guy". If some no-name startup sold me software or a device and told me they're going to use my bandwidth for whatever reasons, I'd be just as concerned. I use plenty of Amazon on a day to day basis, no need to "tear down the man". But these Echo and Show devices will stay turned off.
it's acting as a general gateway for semi-public use. it's not transmitting any of your own data unless you tell the device to send it. but other people who do want to keep in contact with their devices can send a couple bytes every now & again.
claiming that this is like giving access to your car to a stranger is ridiculous. i cant see the fauntest shred of truth to this FUD.
I don't want to talk to machines. It's why we still use keyboards IMO. Talking to machines makes me feel stupid and embarrassed when the wife asks what I said. I only want to talk to other humans or the cat. When I call an automated line that wants me to talk them I find it demeaning so press random buttons until it let's me type a number. Is it just me?
I'm increasingly seeing automated phone systems simply hang up if they don't understand whatever buttons you press, or you have the gall to say "representative" instead of "customer service". I guess it must improve metrics somehow, but it's infuriating to crawl through the tree multiple times until you find the right sequence of precise inputs to get to a representative who can transfer you to the place you actually wanted to go, an hour later.
Recently I was at a budget/hertz (my first mistake) with a reservation and found they had no cars for my reservation and no timeline by which they would.
I was told by the person manning the desk they could not help me, only the representative on the phone could and no, they did not have a direct line.
After 30min on hold person on the phone said they could not tell me what hertz/budget would have a car for my reservation though they were the reservation desk. They transferred me to customer service where I waited for 90 min to no avail.
Even worse, I was in a total loop at booking.com. Enter reservation number pls: easy. Enter pin: there isn't one. Immediate disconnect. No representative, no series of numbers, no alternate phone number available.
It was the same thing with Bluetooth headsets a decade and a half ago. People were walking around randomly talking to themselves. Nowadays it's just normal.
I personally don't want to talk to machines all that much either, but in for your reason. It's because the utility of such an interface sucks. The things can't actually do much other than bring something up on a screen. And they make mistakes half the time.
I've found that with these automated phone things, if you spam enough words with broad enough targets at them, and speak in somewhat unintelligibly, it will just error and connect you with a human.
When voice works, it's nice. For example I do enjoy asking my fire tv to play music or find a show. It's not perfect but it's easier than typing one character at a time on the TV remote, or doom scrolling.
I might not mind talking to computers if they could understand me and there was an easy to use list of words to use. I can't easily process the recordings telling me what to say for X; not sure why, but it's easier to push a number for a topic from a list; part of it is you can expect to interrupt the prompt by pushing buttons, but I don't expect to talk over a prompt (and if I do, I dunno if it will hear me anyway)
I stopped using Alexa a few months ago after it almost burned my house down on Thanksgiving day, then it was virtually impossible to report to Amazon, and then Amazon never sent me an email even acknowledging that they almost burnt down my house.
My daughter put a pizza into the Amazon Alexa Enabled oven/microwave. You can probably guess where this is going. She’s done this many many times, put it on the metal rack, and let me know it was good to go.
I opened the Alexa app on my phone and said, very precisely (I saved the log) “Alexa, bake at 400 degrees Fahrenheit for thirty minutes on first oven”.
Alexa interpreted this as “Alexa, fig at 400 degrees Fahrenheit for thirty minutes on first oven”.
A
Alexa interpreted “fig” as “microwave”, and proceeded to set the microwave/oven full of metal cookware to microwave for 30 minutes.
For some reason I thought being extremely precise and including degrees and things that would only make sense if you wanted to bake, or the fact that I have never microwaved something for that length of time, would inform Alexa and keep this from happening.
Around 21 minutes later my wife smelled something, as we were upstairs, and we had to open all the windows and leave a few hours early for our relatives house. Thank goodness we weren’t hosting.
Anyway that’s the last time I’ll use smart appliances like microwaves.
Sorry, I'm not really familiar with these oven/microwave combos but is it like... one compartment? Or do you store metal in your microwave when you're not using it?
I was wary of such a device too but wanted to try it for the convenience factor. Our other daughter is wheelchair bound and needs to have someone with her at all times, so it was nice that our 6 year old could put things into the oven even if she couldn’t operate the controls.
Oh wow, I can see how that could be convenient for certain cases (such as yours) but as a general consumer device that sounds like a phenomenally bad idea.
Maybe I’m crazy, but to me, having a heating element that can burn my house down, directly able to be controlled from the internet seems like the first issue.
I get the Internet of Things dreams (I understand, not agree), but who would want this? You still need to be there to load and consume the food. It’s not like the HVAC that can be of use to me while I’m not there yet.
I dont know... When I put something in the oven, I'm already right there. Turning the knob is at least at fast as saying what I want, certainly at the precision of OPs statement. I keep failing to see the use case.
This was an Amazon branded device. They apparently no longer sell it. It makes me more upset that it seems like no human ever acknowledged that my $300 microwave malfunctioned so spectacularly. Especially considering how specific I’d learned to get with my commands to it.
We had our daughter who can’t consistently work the microwave put things in then we’d have Alexa cook it. We’ve got a disabled daughter who needs to be monitored at all times so yes, it made things way easier and we’d used it a hundred times before with no issues. It never microwaved when it was supposed to bake before that.
Pacemakers these days do have BLE and automatically communicate over the internet through bridges such as a smartphone app or dedicated night stand device.
Maybe I'm overly paranoid, afraid of everything, and wrong... but I've always considered machine learning and AI a probability problem. In other words any solution is a probability with an error (the amount of which is unknown to the user). That's why I treat these toys with extreme caution, because I know that there is an error that is not negligible. Your experience reinforces my idea. Another example would be a self-driving car with the driver sleeping inside.
i have never used any of these devices, so what's confusing me here is that is there no option to have the device repeat your command in it's own words to make sure it was correctly understood?
i would be scared to use such a device to control appliances or make orders without doble checking that it is actually going to do what i want. even for something seemingly harmless as playing a movie if a misunderstanding could play something inappropriate for the kids.
"The assertion that Alexa growth is slowing is not accurate,” Pearsall wrote to Bloomberg. “The fact is that Alexa continues to grow—we see increases in customer usage, and Alexa is used in more households around the world than ever before.”
End-quote.
So that sure sounds like some board-member speak there.
That said, I personally use Siri. I typically use it to check the weather or play music. Nothing else.
I also have a HomePod mini. It’s fine for what it is.
I just don’t think that voice controlled software is going to go beyond general information that you Can access quickly by asking a simple question, especially if you are doing something else at the same time.
Notes? Faster to check it manually since most people can read and find information faster than a voice telling you your notes one by one.
Searching for info? Right now it’s easier to do it yourself. Even Siri will say “here’s what I found on the web for #thing”, forcing you to look at your phone.
I don’t know, maybe there is potential there. But people will always do what’s easier and faster. A voice just isn’t faster than reading and doing tasks yourself in many situations.
I use Siri for some simple things, but one trick that keeps me coming back is it understands whispers really well. I can be looking at my screen and see a number, and whisper “dial 345-three oh two-fourtyfivehundred” or any manner of numerical words so quietly I can barely hear it, and it usually works.
The idea of having a computers integrated with the home seems great, but to me the voice interface is awkward at best.
I’d much rather have a uniform interface for controlling house devices via their APIs as wall mounted touchscreens (connected to mics, speakers, etc. as desired)
An iPad is close hardware wise but a uniform and customizable UI would be ideal.
Just seems like a lot has been built via voice because a voice OS didn't exist, as opposed to voice actually being the best way to receive commands and output information.
I’ve used Alexa since shortly after its release. We use it daily for music, lights, timers, and random questions like what internal temperature to cook chicken to. All our use cases are either when our hands aren’t free or the request is room-related.
I used to use google assistant on my phone, back when it first became a thing.
These voice assistants are sparkling novelties and nothing more. Since cell phones and ramping up with smartphones, people have this weird hype based relationship with technology. It permeates everything. Screens everywhere, capacitive touch everything, wireless where it isn't necessary, it's all novelties that don't actually improve anyone's life. And the voice assistant is probably the biggest example.
And that's what these products are supposed to be, pretty on a shelf, playful in the commercial, a spunky young woman smiling while making breakfast for the kids or going on a jog, all to get you to buy them, not to get you to use them. As long as they're plugged in in your house tracking everything you do they're generating revenue. When you step back from the hype, what you really are doing is inviting a spy into your home in exchange for an overpriced and overengineered egg timer.
I find use for small things. Dictating a text while driving, setting that egg timer is fast, unit conversions, dialing a phone number I’m reading off another screen. But the whole “tie me into home automation” is lost on me.
No one knows how much it really helps, but I have the “raise to talk” and “Hey Siri” features disabled to keep the “always on” buffer as short as possible.
people around me are using alexa way more than usua
and by around me, it's globally, on various networks (discord, irc etc), i see people mention/talk about alexa way more than google, even more than siri..
What does Amazon care? They know having an eavesdropping device in every home is table stakes in their league so I am sure they would keep pushing Alexa shit even if it was losing them a lot of money in the short run
I'm not at all surprised. Far, far too often, the hype of the tech, and the "oooh, new shiny!" excitement wears off and you figure out that your "new, exciting, expensive labor saving device" is, in fact, more trouble than it's worth, the "smart" bits are better described as "dumb as rocks, but in a very un-rock-like way," that there's a sucker in the transaction who paid a lot of something valuable (money, time, attention, whatever) for the device, and that you're that sucker.
I grew up watching Star Trek (TNG), and even many years later, talking to computers just feels somewhere between "wrong" and "the Primary Processor Cortex is infected with Clamanthian Neurovirans!" - it's gotten dumb and they've got until the end of the episode to somehow fix it. It probably involves tachyons and the main deflector dish.
The more they try to sound like actual humans (and voice synthesis has been improving vastly over the past 20 years, I still recall thinking Dr. Sbaitso was impressive), the more jarring it is when they... don't respond remotely like an actual human would. Someone else in this thread shared a story of Alexa doing the same thing it had done many other times, wrong, and setting a microwave going with a metal oven rack in it. If I've told someone how to bake a pizza 30 times, it's fairly unlikely that they'll screw it up the 31st. That's not true talking to a computer, it seems to have the same low probability of screwing it up every time you do it.
That's all before you get into the fact that they're an absolute privacy and security nightmare, that plugging your whole house of "smart" devices in pretty much guarantees that you're broadcasting your every action to data aggregators (I consider actions taken in my home to be my and my family's business, not the property of whoever has the strongest clause in their EULA), that the devices will fail to function long before a simple mechanical device would have (resulting in huge piles of ewaste), and are just generally human-toxic trash.
I work in the low weeds of tech, and a comparison I saw some years back is entirely apt (paraphrased, sorry, I don't know the original source):
- Technophile: Has all the smart devices connected to his phone, can control his shower from across the planet, argues online over which hue of RGB light goes best for each programmed situation, and nothing in his house ever quite seems to work right.
- Firmware/CPU Engineer: The smartest thing in the house is his printer, and he keeps a loaded 12 gauge leaned up against it in case it makes an unexpected noise.
It's very true. Most of the people I know in the deep weeds have a deep, abiding mistrust for more and more of "consumer tech," and a deep hatred of "things" lately.
Personal voice assistants are great for tasks where using a device directly is involved inconvenient: setting a timer while cooking, playing a song r some kind of music spontaneously, turning on all the lights and lowering all the shades in the living room, or getting direction in the car. It works really well for these tasks and I’m totally hooked. But the other stuff, like buys things, where I could just go and do on an iPad much faster, it doesn’t make much sense.
My children developed their own tastes in music well before they could read which I credit to voice tech.
It's not so useful to me personally but I feel it has enriched the family experience.
We have the explicit music filter on due to the young <10 children in the house.
Sometimes they ask Alexa to play a song, but everything seems to require the music plan on top of Prime these days, so it will play a different ‘related’ station instead.
It’s beyond strange when kids asks for a random kids song, but it is not available, so it starts playing obscure rap songs dropping F- and N- bombs (so obscure it bypasses explicit filters?).
It’s so hilariously broken in these cases (has happened two or three times over the past couple years), but I would love to know the breakdown in logic that these kids are finding…
they use it to play songs, our 6 year old uses it to spell words when she is writing cards, they use it to hear animal sounds, tell stories, talk between rooms when friends are over,etc.
It's objectively worse than a switch or a remote control. Home automation is only nice when it is passive. The moment you need to interact with it, it stops being automation and becomes multiple layers of abstraction to deal with to do an otherwise simple task.
I have several blind friends, they use voice assistants, but don't particularly like them. In all honesty, a lot of voice-based stuff isn't that advanced versus what were using 15-20 years ago with Wildfire. For productivity for disabled users, they're quite mediocre in many ways. Not saying they're terrible, but I am saying there's a very long road ahead, and being a minority means it is low priority and low understanding for most corps. Their focus is on gimmicks over substance.
68 comments
[ 24.7 ms ] story [ 4124 ms ] threadAll of these assistants have limitations because they don’t have any real context to the requests. People think that they are failures because they are not really intelligent but they do have small uses on a daily basis as long as you understand the limitations.
An alarm clock that doesn't stand a pretty good chance of randomly broadcasting your bedroom audio to the internet on a regular basis. You can even get ones that work when the power's out!
If you replace "assistant" with "internet connected room mic" when pondering them, I find it helps make the decisions far clearer. The best place for them is never having been shipped in the first place. The second best place is the trash. Down the list from there is "unplugged and forgotten." And somewhere at the bottom is "Powered on, connected, and waiting for a plausible excuse to broadcast fairly clear room audio out to the internet."
I've abandoned smartphones for the time being (but have a PinePhone to hack on for sport and challenge - it in no way resembles a functional device at this point in time) and have gone back to a flip phone device with a removable battery that generally lives on the kitchen counter, and has a pretty poor room mic. It gets rebooted (full power cycle with battery removal) often enough, and I frequently don't bother bringing it with me places, or at least bringing it with powered on. I've found since I'm looking at it far less, it falls out of mind, so I will quite literally head into town and not realize I don't have it until I get nearly back home. It's fine, I did that for years before cell phones existed, I see no reason why I can't go back to that.
"No phones in the bedroom" is just a standing good policy for a range of reasons anyway.
However, I generally trust phones a bit more than I trust the "room mic" class of "assistant" devices. First, they're on battery, and widespread abuse will mean rapid battery drain and a warm phone. A phone, warm in the pocket, is cause for either turning your phone screen off, or alarm if it's already screen off and you don't know why it's warm.
But they also have a less effective microphone. Most of the ones I've experimented with over the years won't pick up a normal conversational tone across the room with any accuracy - but you can walk into a room, query for an Alexa/Google/etc device, and the one across the room will generally respond to you. They have far better microphones than a typical phone does for room audio.
I'm aware that the various devices are not supposed to be a live stream of room audio to the internet and generally don't seem to be doing that unless triggered, but the reports of people who have done a deep dive into their triggered captures have been that it captures a lot of nonsensical chunks of audio for no discernible reason beyond, presumably, "When in doubt, capture and send" - ease of use, of course, meaning that false negatives in detection are far more serious than false positives.
But neither do I trust the companies involved to not be doing a range of audio analysis locally and sending the information up in random updates. I don't know that they do this sort of thing, but I don't trust them not to. You could do things like unique voice count in the room detection to identify how many people typically live in a place, or if people are having a party. You could do vocal stressor analysis to determine if people are frequently angry, upset, etc. And I'm sure there are many papers written on what else one can use this sort of stuff for. I object to the capability being present, even if not used. Hence my general trend of not keeping smartphones around me regularly - but even there, at least Apple devices do tend to indicate if something is abusing mic privileges, and I can remove audio privileges from apps. I can't do that with the... skills? whatever the audio helper additional functionality code is called.
I don’t need an internet connected spying device for this though. So ended up switching to a HomePod mini.
It sets timers without internet, and plays music from my phone which is somewhere around the house. It can ask my phone for weather info as well.
For light control it's very useful and pretty cheap so if I only used it for that I would be happy.
I have a Google Home assistant and I only use it to play music since it has fairly decent sound quality. My Amazon echo and show devices will stay unplugged.
i really am very happy to find fault, to elaborate misuse. but the reaction to amazon using a very little bit of bandwidth (usually under 50MB i expect) feels entirely emotional, entirely reactionary. in any technical sense we can measure it is for all intensive purposes going to be irrelevant.
this feels like a social dilemna. we want to activate ourselves against amazon. want to tear down the big, the seeming man. but this feels petty & small. that we so readily make a stand here says that we arent really discerning or tasteful or making sensible assessments & rationalizations & choices. we're acting in aggressively zero sum hostile premises, & causing a stir over that which is nothing.
generally i think you're larger point about mistrust is dead on. these systems are black boxes, running in far off data centers. this is anti-personal computing, unobservable by the primary stakeholder, & it's a fucking abominable form of computing that i think more & more people realize has no respect for them, that treats them like end users, not masters of their home & it's systems.
It has nothing to do with the "big guy". If some no-name startup sold me software or a device and told me they're going to use my bandwidth for whatever reasons, I'd be just as concerned. I use plenty of Amazon on a day to day basis, no need to "tear down the man". But these Echo and Show devices will stay turned off.
claiming that this is like giving access to your car to a stranger is ridiculous. i cant see the fauntest shred of truth to this FUD.
I was told by the person manning the desk they could not help me, only the representative on the phone could and no, they did not have a direct line.
After 30min on hold person on the phone said they could not tell me what hertz/budget would have a car for my reservation though they were the reservation desk. They transferred me to customer service where I waited for 90 min to no avail.
Even worse, I was in a total loop at booking.com. Enter reservation number pls: easy. Enter pin: there isn't one. Immediate disconnect. No representative, no series of numbers, no alternate phone number available.
Usually, the robot will give you something like "i didn't understand you, can you please repeat that?" and blow 2 or 3 more times into the mic...
Afterwards, i'm usually transferred pretty quickly to a human being - presumably this is due to some accessibility requirements....
I personally don't want to talk to machines all that much either, but in for your reason. It's because the utility of such an interface sucks. The things can't actually do much other than bring something up on a screen. And they make mistakes half the time.
My daughter put a pizza into the Amazon Alexa Enabled oven/microwave. You can probably guess where this is going. She’s done this many many times, put it on the metal rack, and let me know it was good to go.
I opened the Alexa app on my phone and said, very precisely (I saved the log) “Alexa, bake at 400 degrees Fahrenheit for thirty minutes on first oven”. Alexa interpreted this as “Alexa, fig at 400 degrees Fahrenheit for thirty minutes on first oven”. A Alexa interpreted “fig” as “microwave”, and proceeded to set the microwave/oven full of metal cookware to microwave for 30 minutes.
For some reason I thought being extremely precise and including degrees and things that would only make sense if you wanted to bake, or the fact that I have never microwaved something for that length of time, would inform Alexa and keep this from happening.
Around 21 minutes later my wife smelled something, as we were upstairs, and we had to open all the windows and leave a few hours early for our relatives house. Thank goodness we weren’t hosting.
Anyway that’s the last time I’ll use smart appliances like microwaves.
I get the Internet of Things dreams (I understand, not agree), but who would want this? You still need to be there to load and consume the food. It’s not like the HVAC that can be of use to me while I’m not there yet.
Maybe I'm just a Luddite.
What next, Internet connected pacemakers?
i would be scared to use such a device to control appliances or make orders without doble checking that it is actually going to do what i want. even for something seemingly harmless as playing a movie if a misunderstanding could play something inappropriate for the kids.
"The assertion that Alexa growth is slowing is not accurate,” Pearsall wrote to Bloomberg. “The fact is that Alexa continues to grow—we see increases in customer usage, and Alexa is used in more households around the world than ever before.”
End-quote.
So that sure sounds like some board-member speak there.
That said, I personally use Siri. I typically use it to check the weather or play music. Nothing else.
I also have a HomePod mini. It’s fine for what it is.
I just don’t think that voice controlled software is going to go beyond general information that you Can access quickly by asking a simple question, especially if you are doing something else at the same time.
Notes? Faster to check it manually since most people can read and find information faster than a voice telling you your notes one by one.
Searching for info? Right now it’s easier to do it yourself. Even Siri will say “here’s what I found on the web for #thing”, forcing you to look at your phone.
I don’t know, maybe there is potential there. But people will always do what’s easier and faster. A voice just isn’t faster than reading and doing tasks yourself in many situations.
I’d much rather have a uniform interface for controlling house devices via their APIs as wall mounted touchscreens (connected to mics, speakers, etc. as desired)
An iPad is close hardware wise but a uniform and customizable UI would be ideal.
Just seems like a lot has been built via voice because a voice OS didn't exist, as opposed to voice actually being the best way to receive commands and output information.
These voice assistants are sparkling novelties and nothing more. Since cell phones and ramping up with smartphones, people have this weird hype based relationship with technology. It permeates everything. Screens everywhere, capacitive touch everything, wireless where it isn't necessary, it's all novelties that don't actually improve anyone's life. And the voice assistant is probably the biggest example.
And that's what these products are supposed to be, pretty on a shelf, playful in the commercial, a spunky young woman smiling while making breakfast for the kids or going on a jog, all to get you to buy them, not to get you to use them. As long as they're plugged in in your house tracking everything you do they're generating revenue. When you step back from the hype, what you really are doing is inviting a spy into your home in exchange for an overpriced and overengineered egg timer.
No one knows how much it really helps, but I have the “raise to talk” and “Hey Siri” features disabled to keep the “always on” buffer as short as possible.
people around me are using alexa way more than usua
and by around me, it's globally, on various networks (discord, irc etc), i see people mention/talk about alexa way more than google, even more than siri..
I grew up watching Star Trek (TNG), and even many years later, talking to computers just feels somewhere between "wrong" and "the Primary Processor Cortex is infected with Clamanthian Neurovirans!" - it's gotten dumb and they've got until the end of the episode to somehow fix it. It probably involves tachyons and the main deflector dish.
The more they try to sound like actual humans (and voice synthesis has been improving vastly over the past 20 years, I still recall thinking Dr. Sbaitso was impressive), the more jarring it is when they... don't respond remotely like an actual human would. Someone else in this thread shared a story of Alexa doing the same thing it had done many other times, wrong, and setting a microwave going with a metal oven rack in it. If I've told someone how to bake a pizza 30 times, it's fairly unlikely that they'll screw it up the 31st. That's not true talking to a computer, it seems to have the same low probability of screwing it up every time you do it.
That's all before you get into the fact that they're an absolute privacy and security nightmare, that plugging your whole house of "smart" devices in pretty much guarantees that you're broadcasting your every action to data aggregators (I consider actions taken in my home to be my and my family's business, not the property of whoever has the strongest clause in their EULA), that the devices will fail to function long before a simple mechanical device would have (resulting in huge piles of ewaste), and are just generally human-toxic trash.
I work in the low weeds of tech, and a comparison I saw some years back is entirely apt (paraphrased, sorry, I don't know the original source):
- Technophile: Has all the smart devices connected to his phone, can control his shower from across the planet, argues online over which hue of RGB light goes best for each programmed situation, and nothing in his house ever quite seems to work right.
- Firmware/CPU Engineer: The smartest thing in the house is his printer, and he keeps a loaded 12 gauge leaned up against it in case it makes an unexpected noise.
It's very true. Most of the people I know in the deep weeds have a deep, abiding mistrust for more and more of "consumer tech," and a deep hatred of "things" lately.
Sometimes they ask Alexa to play a song, but everything seems to require the music plan on top of Prime these days, so it will play a different ‘related’ station instead.
It’s beyond strange when kids asks for a random kids song, but it is not available, so it starts playing obscure rap songs dropping F- and N- bombs (so obscure it bypasses explicit filters?).
It’s so hilariously broken in these cases (has happened two or three times over the past couple years), but I would love to know the breakdown in logic that these kids are finding…
they use it to play songs, our 6 year old uses it to spell words when she is writing cards, they use it to hear animal sounds, tell stories, talk between rooms when friends are over,etc.
Adults just use it for weather and timers…
I have several blind friends, they use voice assistants, but don't particularly like them. In all honesty, a lot of voice-based stuff isn't that advanced versus what were using 15-20 years ago with Wildfire. For productivity for disabled users, they're quite mediocre in many ways. Not saying they're terrible, but I am saying there's a very long road ahead, and being a minority means it is low priority and low understanding for most corps. Their focus is on gimmicks over substance.