I had the opportunity to sit down with Ajay Juneja, the founder of Mountain View based Speak With Me, at Tag Camp this weekend. I had a half hour demo in his car and was completely blown away by what he’s created. You can see pictures of Ajay and his car at my flickr account.
What’s so special about his car? He’s layered a proprietary dialog manager, which semantically parses and analyzes data, on top of off-the-shelf speech recognition software, to create one kick ass computer system that is controlled with an array microphone. It allows you to control your car via speech. It was jaw-droppingly impressive.
For the demo, Ajay controlled his stereo system verbally, changing songs, picking tracks, controlling volume, etc. The commands he used to do this were in absolutely plain English and varied considerably in structure. For instance, Ajay at one time said “go to track 3”. Another time he said “play for me another one bites the dust”. Later, “I would like to hear wonderful tonight by eric clapton. “louder”. “quieter”.
Everything worked. No errors. Perfect. He repeatedly tried to trick the system by asking for a song with the wrong artist, etc. The system simply asked him, in its charmingly computerish voice, whether he wanted to hear the song, or a song by the artist.
Ajay then showed me their navigation product, which will integrate into car’s existing nav systems. Drivers will be able to control many aspects of the driving experience verbally once this product is commercially available.
I didn’t want to get out of the car. Frankly, I wanted to find a way to steal his car.
Robert Scoble saw Ajay’s car months ago at BarCamp. He agreed (verbally) to a non disclosure agreement and couldn’t talk much about it. Now, Ajay’s company has advanced enough that he is talking about many aspects of the product, and giving demo’s, without NDAs in place.
Ajay graduated from the computer science program, with a minor in robotics, at Carnegie Mellon University. He also took graduate courses at the language technologies institute (see also the speech to speech product being discussed all over the web today, which is has also been developed at Carnegie Mellon), where this software has been under development for ten years. The original Carnegie Mellon researcher on the technology behind speakwithme is Matthias Denecke, who is now an advisor to Ajay.
Speak With Me will not be commercially available until at least late 2007, when it will be built directly into cars (they are negotiating deals now) and available at high end car stereo shops. The car stereo product will be around $400-$600, and the navigation system will be in the $2,500 range.
They will also offer additional services. One example – the car’s voice can be a celebrity. Ajay asked me how I’d like to have Paris Hilton, or whoever, interact with me to control my stereo and navigation system. Yep, I’d be up for that.