Hacker News Reader

Comment by hyperbovine | Hacker News Reader

This is awesome. Nobody will use it. People feel stupid talking to computers. Remember how you embarrassed you felt the last time you were forced to bark at one of those computerized phone operators in a public place? That's this, only all the time.

Would you not feel like the biggest doofus in the world whipping out your Droid 2 and performing "note to self [note]" in front of people?

View on HN

You are being shortsighted and silly. Here are some excellent examples to show you that speech technology could be pretty damn amazing! The first one <strike>could</strike> will save lives.

(While driving) "Should I turn left or right at this intersection?"

(While shopping, examining a carton of eggs) "Were these eggs cheaper at the corner store?"

(While coding) "Any meetings soon?"

..... Do I need to continue? The applications are endless, assuming it actually works. People have felt stupid in the past because the conversation would go like this: "note to self"....."I said, NOTE TO SELF not NOTICE ELF you fucking idiot machine!"

The point is...if Google can do it well, it will take off.

Shortsighted perhaps but I don't think he is silly.

The problem is about trust. Do I trust the system to give me proper feedback. What if I am yelling because I am stressed, what if it's noisy etc. What if we are two people right next to each other.

Abd then there are quite a few mental reasons why voice recognition isn't just going to be a consumer reality, even if/when the technology is there.

I clearly stated that speech technology would only take off if it works. If you don't trust it, or it can't handle you yelling, then it doesn't work and I agree that it won't take off.

I'm curious about the "mental reasons" you hint at. What are those reasons? Do you mean something like Uncanny Valley? Or something else entirely?

you can expect a person to ask if something is unclear. no speech recog system has done that yet.

speech recog is orders of magnitude behind humans in accuracy.

understanding is another thing entirely. i don't mean philosophy but the problem of even parsing a sentence. just look at implementations of automatic language translation to get a feeling for this problem.

speech recog has a long way to go because AI has a long way to go.

you don't get human-like machine interaction out of bare speech recog.

I wouldn't feel stupid if I never had to repeat myself. For example, I use iPhone voice dial all the time. Even on really complicated or foreign names, like my wife's for example, it rarely makes an error. So rarely I can't think of an instance in recent times. The search space is really small, but it takes advantage of that in a way that makes it useful.

Android does not do a good job with this. When I use it to do a voice dial, it has two major problems. First, it must be connected to the internet, because the sound or a signature of it is going to be sent over the wire. Especially when I am in a low signal area or on Edge, this is a deal-breaker. The second problem is that it's much less accurate than iPhone's voice dial. It typically will only get one word of the name of the person I'm trying to dial -- the surname. As a result, if I'm trying to call someone in my family, it pops up an alphabetical list of all of their names. So I would never use Android voice dial when I was on Android, because I invariably had to repeat myself or interact with the screen, which if I'm already doing that then why use voice dial at all.

(1) I have an iPhone, but I'd really love this type of functionality. I agree I'd probably feel weird saying this to the cell phone. However, when I'm using my earphones with microphone built-in, it's actually extremely handy to issue a command (like switching songs, calling someone, etc) without having to actually whip out the phone.

(2) It's interesting that they trust the voice recognition so much that you can say "Call X Company" and they will google it and call it. If they get it wrong, you'll end up dialing some really weird place instead. This is a personal problem I have with the voice recognition commands on my iPhone. I've ended up calling completely unintended people because the phone misunderstands me, and I can only imagine adding an extra factor of error may make things worse.

I really hope this type of stuff catches on.

You used to look like a giant dork checking your calendar on a PalmPilot.

I use the voice search in my phone. Not everywhere, because sometimes I would feel like a doofus, but when I'm walking it is so much easier to talk to the phone instead of trying to type into it.

I can see this being really convenient for text messaging. It's not perfect, but accurate enough to be useful.

Considering the number of mistakes I make typing on my cell phone, this could actually be a vast improvement :S

Well, this isn't a totally new thing for Android; it's more an expansion of the voice recognition capability Google's already been adding into the system for some time now.

I'll be perfectly honest -- my gut reaction is to agree with you. That said, I've been surprised by how many people I've encountered that talk about how much they love using voice-to-text for things like sending SMSes while driving.

I just updated my phone with it, cant w8 to be driving and be able to use my phone's functionality without pressing buttons and whatnot

I wonder if you can modify the action commands. I'd make the note to self command a self referential 3rd person "note to elblanco".

That is a novel solution to the issue of malicious, or at least irrelevant commands.

http://dilbert.com/strips/comic/1994-04-24/

(while searching for that comic, I found coverage of a MS speech recognition issue, whereby malicious sound recording could speak commands...)

Did you just copy and paste a past comment of yours?

I can say that I absolutely will be using it. Frequently.

Typing on a touchscreen keyboard is not enjoyable, and this makes the device far more usable for casually looking up restaurants, getting directions, etc. Android already had a wide range of voice recognition, but this (from at least a cursory set of tests with it -- experiences might sour trying to use it in the real world) seems to be a HUGE improvement.

Do I think people will use it in public places or at the workplace around peers? I certainly hope not. But in the car, in your own house or office...this is a killer application.

Reply on news.ycombinator.com