Meet Siri's Little Brother, Trapit
waderoush writes "Virtually overnight, Siri, the personal assistant technology in Apple's new iPhone 4S, has brought state-of-the-art AI to the consumer mainstream. Well, it turns out there's more where that came from. Trapit, a second spinoff of SRI International's groundbreaking CALO project (Cognitive Assistant that Learns and Organizes), is preparing for a public beta launch this fall. The Web-based news aggregator lets users set up persistent 'traps' or filters on specific topics. Over time, the traps learn to include more articles that match users' interests and exclude those that don't. Philosophically, it's the exact opposite of social-curation news apps like Flipboard or Pulse, since it uses adaptive learning and sense-making technologies to learn what users like, not what their friends like. 'Just as Siri is revolutionizing the human-computer interaction on the mobile device, Trapit will revolutionize Web search as we know it today,' the company asserts."
While a program that fetches more things you are interested in is great, you should realize the consequences of such a program. In particular you should realize the concept of a filter bubble. Namely that by only picking out things you are already interested in, you exclude things that you could be interested in or things that are too important to exclude.
There's been a TED talk about this, I suggest you watch it so that you can take active steps (when needed) to step out of your comfort zone now and then:
http://www.thefilterbubble.com/ted-talk
Apple's Siri is not necessarily 'State of the Art', but like just about everything Apple does... It just works. Siri is causing a splash because ... unlike Android. It works properly. I don't use voice on my Android because it is worthless to me. I say 'Call my wife' It says. 'Calling Lowes Home Center'. It NEVER EVER gets it right. I have several friends with Androids and only one friend with that perfect voice that can get it to understand him, and even he often has to ask it twice . My wife HATES my Android and never bothered with a Smart phone before because she did not really like them. Too big and bulky. Her phone finally broke and she bought the 4S.
Like everything else Apple does. It just works. She talks to it. It understands every word. I talk to it ... It understands every word. .. and it ALWAYS seems to say something appropriate in response. True that the Android voice can do more than Siri. But I would rather have a voice that can do less properly than one that can do lots of stuff wrong. The only thing I find the Android voice useful for is a good laugh. I fire it up occasionally and ask it something and get a chuckle with just how wrong it gets my request. When she got Siri, we had a house full of people that evening and we passed my Android around playing with the voice. It did not once get anything right anyone said. 7 different voices asking it stuff and not once was it even close. Siri understood everyone perfectly.
So the Android voice is useless. Siri is useful. Therein lies the difference.
But, just go on hating in ignorance, it's so much easier.
When I tested it out, it did much better than my Android, with no "training". Try Android voice with a Boston accent. I tell it to call my favorite bar and it calls a sheep.....
One of the people who worked on Watson, the computer mind put to the test on Jeopardy, is my former brother in law. When BrotherInLaw -1 began on computer AI there was, at the time, no one more advanced than he to challenge his thesis. The stuff we're seeing now in Siri is very much like what Watson did and projects BIL -1 has been working on for over 10 years, only put to "commercial / consumer" use; something inevitable. I doubt anyone involved with the first missions to the moon were all up in arms saying "What? Velcro? *ththt* That's been out for ages." Remember, to much of the media and your average user, this IS bleeding edge!
This is what happens with technology. It gets invented, it gets used in science and technology circles for a while then, if it's got commercial appeal, it ends up in the hands of Joe 6GB.To those lambasting Apple, while I assure you is something I enjoy, is sort of shooting fish in a barrel.
All that said, I use Android for one very simple reason: Apple's Ap Store policy makes me rage. Their puritanical requirements on nudity, "obscenity", etc as well as their tight fisted control over interface is preposterous and reprehensible. When I'd heard they forced a German news agency change their iPhone ap due to a few boobies was when I decided I would never, ever own one. Many of my users have them, they're bought by my employer, I've been offered a new iPhone each year, but for the last two years I've very much enjoyed my Android. The voice command blows, no argument. The screen pivot is comical. But all the aps I have, I enjoy. I can play around with whatever aps I want and not brick the device. To me, that's a fair cop; One programs functionality (Siri) does not out weigh freedom to do as I wish with my devices.
I just choked on my cup of tea reading that. It's voice recognition feed into some search engines, Wolfram Alpha, Yelp and some snippets from Wikipedia and the result plays through text to speech, mashed up with voice commands. If you call such a remix of off-the-shelf tech and existing services state-of-the art AI then you must be joking.
If it's so obvious and easy to do, why haven't you done it? From reports from actual users, it seems to me that for the first time we have a voice recognition system that can do more than respond to a small number of precisely-defined words. If that's not state-of-the-art, I want to know what world you live in, and can I have some of the futuristic tech you must be using?
Indeed voice commands have been in many phones for a while,
Including older iPhones - but here's the problem: They barely work. I use it very occasionally for simple things, like getting the time in winter when the phone is somewhere in an inside pocket.
From all I've seen, Siri works. That right there is the entire secret. It doesn't have 25613 features, but it works.
What's worse is Apple probably managed to get a patent or two on Siri.
They bought it. If there were any patents, they certainly now own them, but it's not Apple's fault or decision. Siri was almost complete when it got bought up.
It is so obvious that a bunch of coders at a hackathon could put something similar together in a few hours and have a demo of the same thing. Oh... wait... they've done exactly that, it's called Iris Alpha from a firm called, and it took eight hours.
Allegedly. Plust quite frankly, this nice video here:
https://market.android.com/details?id=com.dexetra.iris
has me minus-convinced. Funny how there is always a cut between the question and the answer...
Point is, while Apple's idea is clever, the polish and packaging good and the marketing cleverest, but it is absolutely not start of the art artificial intelligence,
Agreed. It is, however, the state of the art of the personal assistant. It is precisely the polish, integration and Steve's obsession with perfection that makes it a success. I'm sure there's at least a hundred prototype projects around that can do more, have more advanced AI, etc. etc. etc. - but none of them are in a state where you could put them out into a mass market.
And that's why Apple is making more money than they know what to do with, and the Iris Alpha coders are playing "look ma" in the Android market place.
Personal disclaimer: Don't get this wrong as a lack of respect. The same reason is why a friend of mine makes a living with computer games, while I have the better game ideas but barely make what I spend on engine licenses, etc. and consider it a hobby - when I think a game is done, he starts the polishing process, the other 50% of development.
And Apple is a master of that part.
Assorted stuff I do sometimes: Lemuria.org