google duplex

It’s been years since Google I/O has been so impressive. Not since the parachuting Google Glass stunt has the new Google really truly surprised everyone. The in-betweens were all smaller iterations of Google products, tweaking through the Android versions, Google messaging apps, and slowly building basic machine-learning termed AI into their apps. Even all the recent AI push has all manifested as machine-learned models simplified down to single functions done well – ie smart identification in photos.

IO18 was a stunner. https://youtu.be/ogfYd705cRs?t=34m58s
Nobody anticipated Google showing off this incredible demo of a Google Assistant triggered Google Duplex voice call to real humans, and holding a full call conversation, so smoothly, intelligently and realistically. None of the journalists knew, none of the fans knew, and it seemed like even the Silicon Valley people working on AI did not see this coming – that is how far ahead of all other announcements this is. All other companies are working on machine learning bits, just trying to optimize large data sets, and these guys literally blew everyone out the water on so many fronts.
1. How smooth the voice sounds – even current Google Assistant on Google Home is nowhere near. And sheesh they only just announced basic improvements like 6 more voices, continued actions stringing multiple commands. If your GA is so advanced, why are we only getting such simple options. It’s like us here using language talking to a 2 year old and over there is a 6 year old hiding behind.
2. How smooth the conversation is – contextually it is pretty darn coherent. Especially considering it wasn’t a basic simple idea conversation flow but rather with some jumping, clarification, repeats, etc. Does this mean they mastered the basic flow long ago? How long ago?
3. How long the conversation – all we have with GA right now is like a 2-3 sentences max, and Google touts how good their GA is at contextual compared to Alexa. And yes GA is super good at it compared to Alexa and Siri and Cortana. So where did this beast of a full conversation come from? It’s 10 times longer, and that’s exponentially harder.
4. How realistically it mimics humans – in the (fake) pauses, non-word sounds, acknowledgements, intonation.
5. Comprehension of the human responses – the heavy accents, incomplete sentences. Wow why do we still need to talk to GA like a baby.

It is So Good that the reactions to the demo aren’t even about which portion it did well – it did everything so darn well – that the comments are whether it is a good thing to have and whether it needs to declare in advance that it’s a computer.

It deserves a standing ovation for the sheer accomplishment, and also audacity to show this to such a large audience.

Amazon and Microsoft only just announced basic ability to share commands with each other. Fairly straightforward computer API calling. Siri is dead silent. GA is just running it’s own race now.

There are some really good comments on it:
In response to “GA needs to start the call by saying it is GA”, and “We should make AI sound different from humans” or “voice assistant should stick to sounding like a computer”, these are like such standard instinctive human responses to overwhelming new technology.
And then someone said if humans get so dependent on the above cues, would you realize if a rogue AI drops the declaration? And then what even if you create an “AI accent”, it is only going to trend and people / hippies / kids will just adopt the new fangled AI accent.
And what would you do after that?
Trying to stop something artificially or regulate it, will just trigger other disrupters, hackers, etc.

Whatever it is, a new front has been revealed, like when Boston Dynamics showed off their walking running jumping cartwheeling bots.

When these people said AI is their new focus and is the future of Google, they already knew this, long ago.

https://www.theatlantic.com/technology/archive/2018/05/humans-acting-like-robots-vs-robots-acting-like-humans/559955/

But maybe small talk has a purpose. The urban theorist (and hyper-observer of the city) Jane Jacobs thought that all these dumb interactions were the social fabric. She wrote in The Death and Life of Great American Cities:

The sum of such casual, public contact at a local level—most of it fortuitous, most of it associated with errands, all of it metered by the person concerned and not thrust upon him by anyone—is a feeling for the public identity of people, a web of public respect and trust, and a resource in time of personal or neighborhood need. The absence of this trust is a disaster to a city street.

After Jacobs published her book in 1961, most American cities suburbanized, cutting people off from this kind of casual contact with the people around them. But at the same time, they connected through other means, the telephone primary among them.

Already, the push-button service framework, Uber for everything, has eroded this last bastion of local chitchat. Google Duplex will simply extend that trend even to those businesses who have not given in to the computerization of their reservation systems. And I don’t except myself from this trend: I hate making these little phone calls as much as anyone else.

But if Jacobs is right that simply making conversation with our fellow human beings—“most of it associated with errands”—is what generates trust in the world around us, then what happens when no one is ever quite sure if Alexis or Duplexis is calling?

Having spent more time in old-style Singapore HDB neighbourhoods over the past 2 years, going more to the small shops, kopitiams, sub neighbourhood centres, it’s been a homecoming from 10 years of mainly shopping malls? For a period I was mainly interacting with large neighbourhood malls and central shopping districts. It was kindof fresh to go to wet markets, brick and mortar shops. It was cheaper, more friendly, less refined, less upscale, more casual, more interactive, less predictable, less reliable, less convenient, more authentic. yeah. But I’ve also done alot more online shopping hah.

But I find that the value applies more to food outlets than other shops.

And requoted from http://thoughtwax.com/2018/05/google-duplex/

Kurt Vonnegut – buying an envelope
Anyway, I take my pages and I have this thing made out of steel, it’s called a paper clip, and I put my pages together, being careful to number them, too, of course. So I go downstairs, to take off, and I pass my wife, the photo journalist Jill Krementz, who was bloody high tech then, and is even higher tech now. She calls out, “Where are you going?” Her favorite reading when she was a girl was Nancy Drew mysteries, you know, the girl detective. So she can’t help but ask, “Where are you going?” And I say, “I am going out to get an envelope.” And she says, “Well, you’re not a poor man. Why don’t you buy a thousand envelopes? They’ll deliver them, and you can put them in a closet.” And I say, “Hush.”

So I go down the steps, and this is on 48th Street in New York City between Second Avenue and Third, and I go out to this newsstand across the street where they sell magazines and lottery tickets and stationery. And I know their stock very well, and so I get an envelope, a manila envelope. It is as though whoever made that envelope knew what size of paper I’m using. I get in line because there are people buying lottery tickets, candy, and that sort of thing, and I chat with them. I say, “Do you know anybody who ever won anything in the lottery?” And, “What happened to your foot?”

Finally I get up to the head of the line. The people who own this store are Hindus. The woman behind the counter has a jewel between her eyes. Now isn’t that worth the trip? I ask her, “Have there been any big lottery winners lately?” Then I pay for the envelope. I take my manuscript and I put it inside. The envelope has two little metal prongs for going through a hole in the flap. For those of you who have never seen one, there are two ways of closing a manila envelope. I use both of them. First I lick the mucilage—it’s kind of sexy. I put the little thin metal diddle through the hole—I never did know what they call them. Then I glue the flap down.

I go next to the postal convenience center down the block at the corner of 47th Street and Second Avenue. This is very close to the United Nations, so there are all these funny-looking people there from all over the world. I go in there and we are lined up again. I’m secretly in love with the woman behind the counter. She doesn’t know it. My wife knows it. I am not about to do anything about it. She is so nice. All I have ever seen of her is from the waist up because she is always behind the counter. But every day she will do something with herself above her waist to cheer us up. Sometimes her hair will be all frizzy. Sometimes she will have ironed it flat. One day she was wearing black lipstick. This is all so exciting and so generous of her, just to cheer us all up, people from all over the world.

So I wait in line, and I say, “Hey what was that language you were talking? Was it Urdu?” I have nice chats. Sometimes not. There is also, “If you don’t like it here, why don’t you go back to your little tinhorn dictatorship where you came from?” One time I had my pocket picked in there and got to meet a cop and tell him about it. Anyway, finally I get up to the head of the line. I don’t reveal to her that I love her. I keep poker-faced. She might as well be looking at a cantaloupe, there is so little information in my face, but my heart is beating. And I give her the envelope, and she weighs it, because I want to put the right number of stamps on it, and have her okay it. If she says that’s the right number of stamps and cancels it, that’s it. They can’t send it back to me. I get the right stamps and I address the envelope to Carol in Woodstock.

Then I go outside and there is a mailbox. And I feed the pages to the giant blue bullfrog. And it says, “Ribbit.”

And I go home. And I have had one hell of a good time.

Electronic communities build nothing. You wind up with nothing. We are dancing animals. How beautiful it is to get up and go out and do something. We are here on Earth to fart around. Don’t let anybody tell you any different.

This is super entertaining. I like how the author is described as “black humor, satiric voice, and incomparable imagination”.

google duplex

Leave a Reply

Your email address will not be published. Required fields are marked *