Hacker News Reader

Hacker News

Comment by dheera | Hacker News Reader

Parent Full thread

dheera

Not to mention the GPS doesn't do half the things I want it to do.

"OK Google zoom in the map by 2X"

nope

"OK Google don't hide all the businesses while on navigation mode"

nope

"OK Google show all street labels don't hide them from me"

nope

"OK Google rotate the map to the right 30 degrees for 5 seconds"

nope

"OK Google show me satellite maps instead"

nope

"OK Google pause notifications from Facebook while driving"

nope

... I bet most people have to fiddle with their screens due to OK Google's utter incompetence.

Meanwhile the people on the OK Google team are busy figuring out how to arrange a damn haircut which I never needed.

View on HN

That's brilliant, and so funny because it's true.

I've been wondering why we haven't really had any advances in our voice commands for nearly a decade - things like zooming and such are actions I've long wanted to be able to perform.

Even that rotate map.. when your car is shown at the bottom and map is hard locked to current heading, when navigation is about to do a sharp right you end up with zero visibility of the rest of the route as they still haven't bothered doing a projection and modifying rotation dynamically - they're not even trying.

All your examples are exactly the kinds of things that an intelligent system should be able to respond to, in order to be useful.

Getting that haircut ordered is the _last_ thing we need.

Because it takes a dedicated team reviewing commands all day, every day, and dev team implementing event handlers, coordinating with other product teams. It takes lots of dedication.

I used to run a chatbot that screened people. We thought we were asking simple enough questions, and yet two million chats in, we still discovered at least a Dozen unexpected/misdetected intents that required care. A day. And here is the problem: management will ask: "so if we add handling for 'rotate the map 30 deg', how many people will appreciate it?". The answer is usually "one person every 400k sessions". For an MbA trained manager that means there is no reason to do it. And its the same with every other similar discovery.

Running an NLP based tool requires staffing a 10-20 people full time, telling them to keep the quality high and leaving them the fck alone. Thats anathema for most companies that proclaim to be "metric based". Google is a prime example of this behavior.

> The answer is usually "one person every 400k sessions".

But you control that to some extent. If more people are aware of the command and believe it will work (I wouldn't) then you'll probably get higher incidence rates. I wouldn't believe it would work because like the person two parent comments above, less than half the novel crap I try works. So I try novel crap less and less frequently compared to things that I know will work.

I agree. But its a hard sell to management types.

Especially that those systems can usually be "proven to work" because they succeed in millions of cases a day.

> they're not even trying.

Wasn't their core business Advertising? Guess they are just not into these sidetracks much.

I just disabled Google assistant and all of its functions, it's beyond useless and probably the most frustrating thing I have ever used on my phone. And I don't mean just the voice controls, but all the "helpful" tips and suggestions are just bad and I can't stand them.

siri's worse than google's voice system... Siri understands what I'm saying, and then CHANGES the correct words into something stupid.

I just risk a fine and use my hands for a short moment instead of spending 2 minutes being distracted by siri.

When driving I don't bother to use Siri for anything other than "call <person>" or "text <person>".

The most hilarious example of how bad Siri is for me, is that literally 100% of the time that I'm playing something on Apple TV, if I say "Hey Siri, pause the TV" (or any variation, calling it Apple TV, including the name of the room, etc.) it always replies that there's nothing currently playing. And if I tell Siri on my phone to play on the Apple TV, all it can do is start Apple Music playing a random song on the TV. It can however respond OK to "resume on Apple TV" which restarts whatever I was watching that had been paused.

OK Google is awful for me, it repeats back exactly what I said but rarely interprets it, at all. "Don't know how to help with 'navigate home'", "'send a message'", "'find a supermarket'". Bleurgh. This is via Android Auto. Next minute it's reading me a text and asking if I want to reply by "SMS or WhatsApp".

None of these systems show you enough 'under the hood' to know why it's failing and so react. Clearly, here the voice recognition, tts are working well; some optional visual feedback would be great - like show on screen a list of responses:

- Keyword heard

- OK Google responding

- Voice1 recognised

- Input "navigate home"

- Input sent to maps

- Maps didn't understand

- Reporting failure

I guess Google know all this, at the time of the query, but as they don't know what I'm trying to do (though 'navigate home' is pretty clear, perhaps I was actually saying 'buy a gnome') ... doesn't seem like there's a proper feedback loop to help direct/correct the system?

"OK Google, navigate home" has never failed to work for me for a couple of years now. Maybe you had opened the app to a profile that doesn't have your home address configured? Maybe you had been logged out?

The sole thing I use it for is setting alarms/timers, or starting the process of a calendar appointment. Those are the two use-cases I've found useful.

Say "navigate to X" also can kinda work with less fumbling than typing half the time but it's a bit messy.

Last time I tried, I couldn't even figure out how to exit navigation using voice. I'm sure it's possible, even easy, but the combination of not knowing the right incantation, car noise, language issues (does it expect German or English commands?) and overall stress meant I just ended up muting the sound system, which is a physical button.

Would you like to see a map of the nearby roads with no lables on it? I fyou zoom in to the point where a road is half the width of the screen I'll deign to add a label on one of the little side roads you don't care about.

Never will I ever tell you what the arterial route you're currently on is called though. Sucker.

It's not just Google Maps though. Also Apple Maps, Music, Podcasts... I find it so frustrating to use voice commands on my phone in the car that I mostly just don't even try. I set things up before I head off as much as I can and mostly stop somewhere before I make changes.

Reply on news.ycombinator.com