So there you have it. As of October 4, Google Now has a clear lead in terms of the sheer volume of queries addressed, and more complete accuracy with its queries than either Siri or Cortana. All three parties will keep investing in this type of technology, but the cold hard facts are that Google is progressing the fastest on all fronts.
Not surprising, really, considering Google’s huge information lead. Still, I have yet to find much use for these personal assistants – I essentially only use Google Now to set alarms and do simple Google queries, but even then only the English ones that do not contain complicated names.
I personally would totally love being able to use a voice assistant while driving, especially so if I could combine it with Sygic GPS; it’d be great to be able to tell commands to Sygic or an address I’d like to navigate, or similarly, I’d love to be able to make phone-calls or send messages just by speaking. Alas, as things stand I have to stop the car somewhere to do any of that and that’s.. well, inefficient. (I refuse to do any of that in a moving car while behind the wheel, I have no interest in putting pedestrians or other drivers in danger.)
The thing is, all these damn voice-assistants I’ve tried either assume that I only want to use the same company’s apps, like e.g. Google Maps whereas I prefer Sygic, or they don’t do Finnish :/
Android=more users=more queries for Google Now=big deal. Just another attempt to start another Apple-Microsoft-Google argument, but I’m tired of those. All three of these “assistants” are pretty darn good at what they do, but they all have work to do and much improvement is needed to make them truly useful. Until then, I really don’t care how many people use them, I am more concerned with what Apple, Google, and Microsoft are doing to make them more useful, but that isn’t sexy and doesn’t generate page views.
Read the link; the authors share your concern…
The “sheer volume” in the snippet refers to the number of queries successfully answered in the test. So Google Now provides enhanced results for a higher percentage of tested queries and *also* provides more complete answers in those enhanced results.
I think it’s pretty clear that Google Now has the lead. I don’t agree that all of them are pretty good. Google Now is just barely getting to useful stage, and the others are fine for simple things, but often fail spectacularly with other queries.
That said I just don’t use a lot of voice queries. Maybe once I can rely on the answer being correct I will.
Wait, people actually use that feature? I thought it was just Apple, Google, and Microsoft trying to out-gimmick each other…
It just… I just thought people only used Siri to make their phone say some snarky response. :/
i’ve been working with computers day in day out for almost 30 years now, and i’ve never successfully used voice control. even when it “works” it’s a slow, inaccurate, cringe-worthy experience. humans shouldn’t talk to machines and really expect them to understand.
voice has so many issues as a UI:
–i listen to music or radio most the day, so there’s that
–pronunciation issues,
–slang,
–the use of non-mainstream names or words,
–grammar,
–the need for a verbal confirmation of every step which is also often times misunderstood
–issuing singular, serial commands requiring constant review and approval is so slow
and I’m a native american-english speaker, i could only imagine how much worse it is for those with accents/languages different than the company that makes the voice software!
the best use case for voice is hands-free and in private with time to kill, which is why people cite the car as the perfect place for it.
last time i issued a voice command to my phone in the car it called a woman i hadn’t spoken to in years. it was nothing close to what i asked it to do. i yelled at it to hang up and it wouldn’t, i yelled at it to call my wife instead and it just kept ringing this woman’s line, and i immediately thought “great, thx personal assistant for really screwing up my week here” as i have to explain to two women why i made that call, and all i have to say for myself is “my phone did it!”.
it also made me grab the phone to manually override what the voice assistant was incorrectly doing, and swerve in the process, so no thanks voice if you are going to be so volatile.
I think it depends on the person. My dad has this southern drawl and can’t get computers to understand his voice for anything. I, on the other hand, don’t have any issues with it. If I recite long sentences into my phone, it’s gonna screw up eventually. But for short queries and text messages, it works adequately.
I personally find Siri more useful than Google Now (I haven’t tried Cortana) because you can toggle settings on/off with your voice without turning on the phone. Just hold down the home button and speak. I think there’s a way to do that on Android with Tasker, but I haven’t tried it. (I probably should.)
This guy uses voice control to code, as a way of getting around RSI:
https://www.youtube.com/watch?v=8SkdfdXWYaI
Agreed
Multiple microphones coupled with smart software solves that. A webcam could also help.
These all seem the same to me and I think swiftkey has that solved mostly.
These seem the same to me too. Startrek already showed how you should do commands and feedback.
Question of training. It is already solved.
Disagree. I can think of a lot of times when voice control could be a great option. Any command you don’t use often shouldn’t have UI on the screen.
I never ever want to use voice commands or voice input for anything, especially the least used feature of my phone. For the least used features, I want to see what the options are explicitly, I don’t want to have to remember which quirky phrase activates it the way I like.
I agree, less used features should not be prominently placed, but if you make them voice only, I’ll just buy a device from some other manufacturer.
I’ll go out on a limb and predict that voice will never be the primary way we interact with devices. It’s just not efficient. I can type and click much more quickly and privately.
I guess that is sarcasm because nobody in their right mind would view voice as the primary interface except when nothing else would work.
When your hands are occupied voice is faster.
I can’t really think of a time when my hands are occupied that my eyes / brain aren’t also occupied in a way that makes voice command of anything completely redundant. Voice command is a gimmick at best. It’s slow and difficult to do complex tasks with. I travel a lot for work, so making an appointment involves looking at the calendar to see if I not only have time, but am in the right area. Putting calendar items on my calendar by voice, stupidly inconvenient. Doing it with fingers, quick and easy. I scroll through my calendar, see what’s already there and pick the right time.
If I want to find the nearest coffee or pizza or gas, I need to be situated so my eyes can be focused on the device that will be providing me the answer. If I am able to focus my eyes, I can use my hands as well. Voice is just a slower option to do what I can already to well without voice.
Voice command is pretty much a useless gimmick, voice recognition is nice for dictation, but not much else.
Please keep this in mind: I find voice control right now useless. But I can see the possibilities and I don’t think they are far off.
You can’t think of it
Assertion
You can’t think of a way of how it could work
Straw man/Red herring
Assertion
Assertion
Assertion
Just asserting something as truth does not make it so. Please give reasons, evidence or examples. Using your lack of imagination as an argument makes me scratch my head.
Given that you have previously used a work of fiction as part of an argument that voice command can work, I suggest you might want to start over. If you find yourself using “Star Trek” as a reason for anything, you are probably going down the wrong track.
My device will never know more about my preferences than I do, which means I will always need to be looking at it when I am doing anything meaningful (other than boring dictation, for that voice is useful). If I am required to look at it, I can’t be doing other things, and I might as well use a faster method of data input. Talking at my device will never rise to the most effective option, only the most Star Trekky, which is of no actual value, only a gimmick.
You seriously lack imagination.
You’ve never driven a car and found out that you need to find a gas station, for example, but you don’t have any idea where the nearest one is?
Some hands-free use cases: high altitude work like a lineman or bridge repair, biohazard situations, underwater situations, cooking, bathing, handicapped users……
As far as the gas station – aren’t they always on the intersections and by the highways? I think GPS is ruining people’s ability to be interesting and see anything in the world around them.
Finding our way somewhere is a basic animal skill that we should not try to move past.
Not always, no.
Not all of us can help it. I’ve *always* had abysmal sense of direction and there’s nothing I can do about it.
I understand that, we all have our crutches. I have always been pretty good at it, but I also grew up on the south shore of a very large lake so I only had 3 directions to worry about.
I think good map software should include a text summary option, aka “the old way” of giving directions, something like:
“Take a left out of here and go to the end of the street. Thats Stewart Road. Take a right and go about 5 miles to Mill Street. Go south on Mill, down the hill. Take that to Route 471, go North until exit 32.”
IMO that’s easier to read, to understand, and to memorize than any of the current output methods used by map software.
to add to the concise direction software idea —
they should also tailor the size of the directions to your screen, so you can maintain your positional focus.
for instance, a short block of text should stay on the screen in entirety:
“Left out of here to the end of the street. Thats Stewart. Take a right and go about 5 miles to Mill. Go right/south on Mill, down the hill. Take to 471 North until exit 32.”
depending on the display mode, there could be more detail added:
“Take a left out of this parking lot and go to the end of the street (by the Makebelieve Mall). That’s Stewart Road. Take a right on Stewart and go about 5 miles to Mill Street, across from the Wal-Mart. Go right/south on Mill Street, down the hill, and stay in the right lane. Take that to Route 471 North. Stay on 471 North until exit 32, Johnson Ave, about 8 miles.”
or zoom the focus to show the previous 1 and the future 2 steps, in this case we’ve just turned onto Stewart:
(out of view) “Take a left out of this parking lot and go to the end of the street (by the Makebelieve Mall). That’s Stewart Road.
(in view)
Take a right on Stewart and go about 5 miles to Mill Street, across from the Wal-Mart. Go right/south on Mill Street, down the hill, and stay in the right lane. Take that to Route 471 North.
(out of view)
Stay on 471 North until exit 32, Johnson Ave, about 8 miles.”
This method would allow you or a navigator to commit it to memory, and think in small basic chunks. You also would look at the physical world for more clues and milestones.
Edited 2014-10-10 18:20 UTC
Doesn’t Google Maps already do this? I don’t use Maps, but when I tried it the last time it gave me textual directions in addition to showing the route on map.
EDIT: https://dl.dropboxusercontent.com/u/11811685/2014-10-10-21-27-04.png — I don’t know if it scrolls along as you go, but it sure displays some sort of textual instructions.
Edited 2014-10-10 18:29 UTC
yeah that’s not bad. of all the ones out there that i’ve seen, the pic you showed is the best current map interface IMO.
but i still think their basic scrolling list can be condensed and improved (i don’t read that language so forgive my assumptions) :
— the directional icons can be improved with correlating text
— it’s showing distance but no time estimates
— it’s not showing any local place information, like attractions and directional pointers (campus, hospital, police, etc.)
— it’s not showing anything geographic
— i see the total distance of 354km but i’m only viewing 2km on this screen. that would need updating very quickly.
— being that this screen would only stay relevant for a minute at most, it makes it feel much more crowded
— there’s no room for anything visual for quick glance while looking out the windows and driving.
— it needs upcoming basics in bold so you can commit to short-term memory and focus on the road
So i hope the developers of such apps are really working on an end goal of refining the delivery of directional data to us in the least intrusive way.
the apple watch comes into play here, as it promises to tap you on the left or right side of the watch to give you directional cues. if they work that out (i see some issues in practice) the future of digital assistants might be far less intrusive than the current state.
Edited 2014-10-10 20:07 UTC
I’ve never been driving and needed anything where using voice commands and looking as a screen on the dash would be any more efficient or less intrusive than just taking an exit or puling over and doing it by hand. I’ve driven a friends car with the voice command for everything and found it to be slow and inefficient. It’s a gimmick, but even he has stopped bothering to use it, and he’s in the car every day.
At the rate some digital personal assistants are developing, it’s only a matter of time before some dude in Japan marries his smartphone. http://www.youtube.com/watch?v=anhhs3HRbOk
Edited 2014-10-09 15:18 UTC
http://arstechnica.com/cars/2014/10/driving-with-voice-activated-in…
Not a big surprise here. Distraction is the name of the game today and car makers are a big part of it. I recently rented a nice BMW and it had a seriously complicated computer with about 5 different input methods and one of the strangest UI’s i’ve ever worked with.
stare at the screen. look at it now. check it again. stare at it’s animation. try to use the 5 different ways to move the cursor. hey, input some text by scrolling through the letters, video game high score style. remember there’s 10 top level screens with 3-5 sub-level screens below. commit to remembering this and navigating through it. oh yeah, also drive this $70k car down the highway without killing anyone.
i think the gov should consider banning screens beyond a certain size and resolution on the dashboard of cars, or if the car is in gear it should be forced to use a very simple interface.
too many screens, you used to be able to get in the car to avoid the screens and think.
The problem is that car manufacturers still don’t figured out why a car needs a infotainment system at all other to run GPS. So their investment on it is pretty minimal and the engineering teams are all geared to “cram features” on it without devoting a single minute to things like ergonomics or consistence.
So, a infotainment system is more a disjointed gimmick than a real feature, something that is inside the car just to show to potential buyers that it is “modern”.
True this. The engineers working on parts of this understand, but they aren’t given the resources/authority to make those kinds of decisions. Often times the design comes from the label, but the work gets done by a third party supplier whos only concern is getting something that kind of works for the lowest possible cost.
A highly automated modern passenger aircraft requires two pilots with thousands of hours of training and experience to manage a marginally higher workload than a modern car.
TBH sometime it feels like Google know to much about me.
they track every activity, it feels scary, so i tends to use all other alternative services like firefox, DDG, etc. Only service use from google is gmail and play store.
i haven’t use Google Now a single time on my Android.
By using gmail, it means that Google already have access to a huge chunk of your personal life, perhaps the single greatest private repository about yourself that you can have on internet.
So i don’t know why you are scared of the remaining Google services.
BTW, if you use GPS on your phone, it is very likely that you enabled aGPS features to get a faster lock (Google location services), so Google know where you live, where you work, and where the places that you go often.
its scary because it can be used for wrong purpose.
I got your point, for maps at least Now I have alternative for Google maps, HERE maps
http://www.xperiablog.net/2014/10/03/nokias-here-maps-beta-for-andr…
I live in India, so offline maps are welcome
I use Gmail only for professional work. Most of my personal stuff is on good old yahoo mail. Also i always try stay away from this cloud crap.
But I don’t know how much I can hold on like this.
The best example of useful voice assisance is with the Moto X and two less advertised features, driving mode and home mode (which is basically the same as driving). When I get a test, and when the Moto X determines I’m driving, it will ask me if I would like to hear the text, and reply in a conversational way. This is just fantastic when driving, and those of us stateside outside of bigger cities tend to drive a lot. It really is a killer feature, and Siri has nothing like it.
Edit: I should add that while the phone can’t easily be activated with voice (unless you say it EXACTLY right), the driving mode does respond well, and accurately.
I can’t speak to Cortana, because I haven’t used it, though some of the contextual reminders they advertise have really piqued my interest.
Edited 2014-10-09 20:39 UTC
Really my moto X responds to “OK, Google now” almost flawlessly. It’s great for a quick answer or setting a reminder, but even useful in the car for placing calls.
The other day I asked Siri – “What do you think of Cortana?” I must not have been clear enough, she said “noun: a female dog”. Wha…?
fretinator,
http://www.yourdictionary.com/bitch
I didn’t realize Siri was programmed for insults, anyone know of others?
Edited 2014-10-09 22:07 UTC
Hi, I’m from Stone Temple, the people who did this study. When we asked that question, Siri answered, “I’m not at liberty to say.”
That was a joke. 1…2….3… OK, now everyone laugh!
Your comment after the article: “Still, I have yet to find much use for these personal assistants ”
You really have to read the caveats at the beginning of the linked article. They are not testing the quality of the results or the personal assistants. They are only testing if they resulted in good knowledge panels. And the questions were picked to get good results.
“these were not random queries. In fact, they were picked because we felt they were likely to trigger a knowledge panel. In addition, this was a straight up knowledge box comparison, not a personal assistant comparison.”
I really like PA’s, but only when I am alone and/or in the car. Just press a button, ask a question, get an answer. Or press a button, give a command, get the result (alarms are indeed my most used) but making the navigation/calling/sms/mail work by voice is great
Apple clearly has much work to do, but note that iPhone, iPad, and iPod touch users can really supercharge their knowledge box results with the free WolframAlpha Viewer which allows users to explore Wolfram|Alpha results from Siri.
A more comprehensive test of knowledge boxes would have graded Siri with the WolframAlpha Viewer installed which would have significantly improved Siri’s results.
Google Now’s ‘Remind me to do <something> at <time>’ is a killer feature for me. I use it all the time with great success. Google Now went from barely recognising anything I say to getting it right 80% of the time.
I find myself wondering if this is improvement to the software itself or whether some sort of machine learning is going on for it to learn to understand me better over the past year or so.
The only information I ask Google Now tends to be about movies. Its pretty good at that.