“Siri is chapter one of a much longer, bigger story,” says Dag Kittlaus, one of three of the original creators of Apple’s virtual personal assistant. The team, originally acquired by Apple as part of its $200M purchase of Siri, has now left the company to form a new startup, Viv Labs, to work on the rest of that story.
The vision described by the team in a lengthy piece in Wired is certainly ambitious. The problem with Siri, they say, is that it can only do things it has been explicitly programmed to do.
Though Apple has since extended Siri’s powers—to make an OpenTable restaurant reservation, for example—she still can’t do something as simple as booking a table on the next available night in your schedule. She knows how to check your calendar and she knows how to use OpenTable. But putting those things together is, at the moment, beyond her.
What Kittlaus and his team want to do is create a personal assistant which can learn to do new things for itself …
The secret, they say, is giving the system access to as many sources of data as possible and allowing it to intelligently combine data from each to make sense of natural-language requests.
One engineer explains how he has been refining Viv’s response to “Get me a ticket to the cheapest flight from SFO to Charles de Gaulle on July 2, with a return flight the following Monday.” In the past week, the engineer added an airplane-seating database. Using a laptop-based prototype of Viv that displays a virtual phone screen, he speaks into the microphone. Lufthansa Flight 455 fits the bill. “Seat 61G is available according to your preferences,” Viv replies, then purchases the seat using a credit card.
Unlike Siri, they don’t see Viv as something that would be tied to a particular piece of hardware, but rather made licensed to anyone from websites to car companies to TV manufacturers – possibly operating on a commission rather than charging a licensing fee.
Kittlaus cites a factoid about Match.com that he learned from its CEO: The company arranges 50,000 dates a day. “What Match.com isn’t able to do is say, ‘Let me get you tickets for something. Would you like me to book a table? Do you want me to send Uber to pick her up? Do you want me to have flowers sent to the table?’” Viv could provide all those services—in exchange for a cut of the transactions that resulted.
That, they argue, would be a truly intelligent virtual assistant.
The Viv team make no secret of the fact that they are at an early stage in development, and cannot say when they might be ready to launch. But one influential outsider who has seen it in action says he was “blown away” by it.
Vishal Sharma was until recently VP of product for Google Now. When Cheyer showed him how Viv located the closest bottle of wine that paired well with a dish, he was blown away. “I don’t know any system in the world that could answer a question like that,” he says. “Many things can go wrong, but I would like to see something like this exist.”
Viv are not the only former Siri team members working on next-generation personal assistants, but they have certainly set themselves the most ambitious goal. The full piece over on Wired is well worth a read.
FTC: We use income earning auto affiliate links. More.
factoid: ˈfaktɔɪd: noun
an item of unreliable information that is reported and repeated so often that it becomes accepted as fact.
It’s also come to mean a trivial fact. There’s more than anyone could ever want to know about it here: http://en.wikipedia.org/wiki/Factoid :-)
What I like about Siri is that it DOESN’T pretend to be something that its not….an actual human being. It can perform simple tasks, it can be very helpful with a lot of things. But its not a gimmick. It actually works at the things you’d expect it to be able to do. That is not to say there isn’t any room for improvement, but I am never going to ask my iPhone to “book a table on the next available night”. It just doesn’t work that way. That is asking it to do something that requires a lot of personal thought.
I do look forward to the time I can tell my iPhone “Arrange lunch with Jo.” It knows when I’m free and where I’ll be each day, and it knows what I like to eat. It can request the same info from Jo’s phone, which it hands over as we’ve both given permission. I just want back a response that says “Lunch arranged for 1.15pm Thursday at Taylors.”
That would be awesome!
That would be useless for the vast majority of people though, as the vast majority don’t put every facet of their life into the calendar app.
Totally agree. And there’s so many levels to integrate from there… It arranges lunch at Joe’s at 1:15 because Joe’s is located only 5 minutes away from your previous appointment versus a restaurant on the other side of town that would be impossible to reach in time.
Yes and then you can take your iPhone to the café and put it on the table and leave. Then your buddy can come 1 minute later and do the same. Then the two phones can meet and talk and you can go home without any human interactivity.
Sad world when humans will lose the very thing that makes them human: the human touch.
I hate Siri. It always gets at least one of my words wrong in a request. I hate it! Siri is useless around people, around music, in a noisy car, or anywhere that isn’t silent. Useless I say. Google Now is much better, but still doesn’t work where it isn’t quiet.
I have found Siri to work the best on my iPhone 5, its much better than what it was on my 4S. On my Nexus and Moto X however i find Google Now to not really work the way i want to, it is very nice to say okay google anytime on my Moto X but if it doesn’t work the way i want it to then why would i use it?
I have iOS 8 and the hey Siri feature when the phone is plugged up to a charger is great! Works just the way i would expect it to.
Dare I say it is a lot better in iOS 8….but nothing can ever be done about most of your complaints. Why you expect it to work when the microphone itself cannot discern between your voice and noise that is louder than your voice, is unknown to me.
It’s probably his accent…
That’s where a headset comes in handy: running into a noisy airport/bus/train terminal, heavy luggage in both hands: “Hey Siri, what time is it? What time’s my train home? On what platform?”. These can all be done with Siri/app integration.
And it’s only a matter of time before things like “my fastest route?” and “the cheapest flight?” and possibly intelligent combinations of price, duration and other factors become features included in smartphones.
I actually enjoy struggling to get the right response out of Siri, kinda like training an infant!
And I get a great buzz out of hitting on the right presentation of words that gets a correct response from Siri. It’s great fun when colleagues at my workplace, scoffing at each failure, suddenly see the potential of the system as an unexpectedly brilliant response impresses us all with its correctness.
I’m guessing Apple will progress onto including more on-board AI with each device which, combined with server support via the recently built-out $100m Content Delivery Network (CDN), will gradually meet and eventually exceed our expectations with regard to information requests using logical operators like “this AND NOT that”, “this BUT NOT that OR that”, not to mention contextual stuff.
Siri’s future’s bright, so long its creators continue to “Light a Candle rather than curse the Darkness).
Why couldn’t they have worked on this while at Apple?
Internal politics probably. I remember being involved in the Java community in Silicon Valley in the mid 1990s. There were fully working demos (all could have been implemented there and then) of real-time dynamic websites, such as airplane ticket booking with interactive seat selectors, same with movie theatre seats, and countless even more impressive stuff.
None of it ever cam to fruition until now, with html5 and Flash. Why? Because Microsoft were terrified of the web and Java, and instructed corporates to reject it, so 15 years later, we’re back to square one trying to make it work with html5, but we’ve given up and gone back to apps.
The irony!
Either way, futuristic concepts (even if working) rarely make it out at prototype stage, sometimes it can be up to 15 or even 20 years! Consider HP’s 1970s smart watch, as mentioned today elsewhere on a gadget blog.
With Siri, either Apple pushed for a simpler easier to achieve solution or the technology wasn’t ready then.
Or there are philosophical disagreements between Apple and Viv’s team that originated from the discussion of how much privacy the user should have to give up to enable a truly, next generation, personal assistant experience.
I am guessing Apple wanted to protect the user’s privacy more than this app is willing to adhere to.
All of those demos you speak about first arrived from us at NeXT with WebObjects 3.0 and its Openstep/Cocoa Objective-C frameworks. Whether Dodge, United Airlines, USPostal Service and much more, NeXT had it.
What changed? Instead of going public [a week prior to this happening] we merged with Apple. Apple after Steve came back realized the Consumer was dying and switched focus from The Enterprise to the Consumer.
It’s just now getting back to the Enterprise with IBM.
We changed from leading the Enterprise with such tools to knee-capping WOF with Java and ignoring it.
I feel like Apple has been Very lazy with Siri. People and location based reminders (as seen on Cortana commercials) should have been added 2 years ago. Right now Siri is in Last place compared to google now and cortana.
I would never use my phone to book reservations for dinner and I would bet that 1 in 1000 people actually use that feature. Most people using the phone are average Joe’s and don’t need a reservation to eat.
I have a $49 Nokia 520 running 8.1 and cortana does 99% of what siri can do and some things siri can’t do.
the 520 is a second device. iPhone 5S as main.
Siri can do location based reminders… you just have to have the location by address.
What Siri needs more than anything else is functionality like Watson (supercomputer used on Jeopardy). Apple should have a supercomputer like that which sends info quickly back to your device. Siri is bad in that most inquiries simply ask you to search the web, or it simply doesn’t understand. It needs to be so powerful that you can ask nearly any question and it brings back results, or at least possibilities with links to those possibilities.
“Watson-enhanced Siri” is exactly what excites me about the IBM/Apple hookup to create intelligent enterprise apps for Healthcare, Marketing, Sales and Services etc.
The spin-offs and hand-me-downs are bound to produce a lot of excitingly useful stuff for the consumer market, in much the same way as the Aerospace Industry gave us anti-skid braking, turbo-charging, fuel injection, satellite navigation and radar in the consumer automotive market.
Hurry tomorrow, we need you NOW!
That’s a good point, I never thought about that when I heard about the Apple-IBM partnership. It’s entirely possible Apple has a lot more planned than just enterprise there… Please enhance Siri with Watson!!!
Sori has been an interesting experiment and it has got a lot of people talking, even if only to mock Siri.
OK Google is much more accurate. And you don’t even have to push a button.
It’s definitely not much more accurate, as I’ve used both. Siri has Hey, Siri in iOS 8 by the way.
Cortana has surpassed Siri greatly and utilizes Neural Net technology, which Apple has only recently invested in.
Cortana it’s far more personal, reliable and her fundamental structure allows her to learn and to grow.
Siri is pretty much just a voice activated assistant.
Cortana is more context aware and capable of pristinely supplying information. Apple has many strong points. Siri is not one of them. Cortana, which just entered the market as a developer preview beta product is already surpassing Siri after only just under 4 months, against a product with a three year head start.
Microsoft has it’s weak points Cortana isn’t one of them.
Wow, spell check! pristinely = proactively 😳
Question: What about the technology behind WolframAlpha? That’s the only puplic available service that IS capable of answering comlex requests instantly. Would be great if they would team up with some voice-recognition guys to develop a true assistant service!
Siri uses Wolfram Alpha, but only to answer questions rather than carry out tasks AFAIK.
One more thing, that I’ll probably never understand…
How do these voice interpreters work and why is the user always bond to one single language?
I’m from Switzerland. Siri works well for me in English but I can’t dictate mails / messages in German without changing the settings first. Isn’t there the language independent PHONETIC ALPHABET that could accept any input???
Why do I have to pronounce a French street name (e.g. “Rue de la gare”) like a retard to get it into the system???
I mean I see, that a company like Apple doesn’t care about us 8m Swissies with 4 languages and multiple dialects but I hardly can’t believe the multi-cultural USA does not have a need for multi-language requests…!?!
I read a lot of sic-fi about expert systems and I was really surprised how far from reality they are.
Was the same for voice-recognition. I first tried it in the 1980s, when it was hopeless, and again every few years. Then suddenly it was fantastic.