4 ms·

One thing I haven't seen discussed is the poor affordance / discoverability of speech technology. Google Home can do some clever things however, it also does n

by sida 6y ago

One thing I haven't seen discussed is the poor affordance / discoverability of speech technology. Google Home can do some clever things however, it also does not have the ability to do some very basic stuff. As a user, how do you know what Google Home can do and cannot do? It is just trial and error. And if Google Home introduces a new feature to be able to complete new types of queries. What now? How does a user know that last month it wasn't able to do something and this month it is able to do that thing. And lastly, the interface of voice is very clunky. It has no concept of temporal memory like Me: "Ok Google, navigate to the nearest safeway" Google: "navigating you to the nearest gas station" The natural thing to say is "no, I meant safeway nott gas station" however, I now have to say "Ok Google, navigate to the nearest safeway". This is analogous to if keyboard had no backspace and you have to retype everything everytime you have a typo. Well that's the state of speech technology right now

gugagore 6y ago
What you say about "temporal memory" is not exactly true for Google's assistant. You can try two separate queries 1. "Who is the president of the United States?" 2. "What is his wife's name?" And it will resolve the deictic pronoun. I haven't tried this feature out extensively, but it has worked for a few years now.
dalbasal 6y ago
Agreed. More broadly, I'd say that no one has made a good UI yet. The mac/lisa/star had a UI that people could learn. iOS... In some ways a voice UI has bigger problems to deal with than PC GUIs or iOS. Those UIs were replacing pre-existing UIs (eg blackberry, dos, unix, norton) and they could target whatever tasks a smartphone/PC needed to do. For voice UIs, it's a cold start. It's not even obvious what an audio only computer should do. Our mental model for a "virtual assistant" is a person-2-person exchange, and computers still aren't great at communicating like people. FWIW, I think slipping into existing niches is the way to go. That's where a useful voice ui will be discovered. Car stuff, accessibility software, living room controls.. At least these have clear goals. Voice operating spotify, netflix or just an iphone is something people actually need and will use if its useful.
RunningDroid 6y ago
>As a user, how do you know what Google Home can do and cannot do? Amazon's workaround for this problem is to have it tell you when new "options" are available for a command (ie: if you set an alarm it would confirm setting the alarm then tell you it can wake you up to the sound of birds then it provides an example command) and Amazon sends out a "what's new with Alexa" email every so often that's 90% example commands.
whimsicalism 6y ago
Once they start hooking them up to conversational language models that can also submit queries, I think it is going to get a lot better. The conversational model results are starting to look very good.
- petra 6y ago
  Where can i learn more about the leading edge of those conversational language models ?
  - whimsicalism 6y ago
    https://arxiv.org/pdf/2004.13637.pdf https://arxiv.org/pdf/2004.13637.pdf

by sida 6y ago

gugagore 6y ago

What you say about "temporal memory" is not exactly true for Google's assistant. You can try two separate queries 1. "Who is the president of the United States?" 2. "What is his wife's name?" And it will resolve the deictic pronoun. I haven't tried this feature out extensively, but it has worked for a few years now.

dalbasal 6y ago

Agreed. More broadly, I'd say that no one has made a good UI yet. The mac/lisa/star had a UI that people could learn. iOS... In some ways a voice UI has bigger problems to deal with than PC GUIs or iOS. Those UIs were replacing pre-existing UIs (eg blackberry, dos, unix, norton) and they could target whatever tasks a smartphone/PC needed to do. For voice UIs, it's a cold start. It's not even obvious what an audio only computer should do. Our mental model for a "virtual assistant" is a person-2-person exchange, and computers still aren't great at communicating like people. FWIW, I think slipping into existing niches is the way to go. That's where a useful voice ui will be discovered. Car stuff, accessibility software, living room controls.. At least these have clear goals. Voice operating spotify, netflix or just an iphone is something people actually need and will use if its useful.

RunningDroid 6y ago

>As a user, how do you know what Google Home can do and cannot do? Amazon's workaround for this problem is to have it tell you when new "options" are available for a command (ie: if you set an alarm it would confirm setting the alarm then tell you it can wake you up to the sound of birds then it provides an example command) and Amazon sends out a "what's new with Alexa" email every so often that's 90% example commands.

whimsicalism 6y ago

Once they start hooking them up to conversational language models that can also submit queries, I think it is going to get a lot better. The conversational model results are starting to look very good.

petra 6y ago
Where can i learn more about the leading edge of those conversational language models ?
- whimsicalism 6y ago
  https://arxiv.org/pdf/2004.13637.pdf https://arxiv.org/pdf/2004.13637.pdf