Ease and Speed of Development is Critical to Natural Language Success
Back in the summer last year, a start-up company thought it would try and raise 100k to fund a project to build a social robot. $2.3m later and Jibo is well on its way. To me it encapsulates all the amazing ways natural language applications can change our lives and judging by the way many people immediately poured money into the crowd-funder, I’m not the only one.
It won’t make you tea in the morning, but it demonstrates how robots can do so much to make our lives easier just through conversation. Critics point out that your phone or tablet probably does most of what Jibo can do in the promotional video, but this is where I think the guys at Jibo have been clever. They haven’t just built a cute robot, they’re building it as a platform.
Like us at Artificial Solutions, they’ve understood that you can’t build a device with all the specialist knowledge that one person might require. For a start what I might want a social robot to do is probably completely different to the next person. Not to mention that if you wanted to learn a new language for example, then you would want to learn from the best. An app by Babbel or Rosetta Stone is going to have far more appeal, than “the software that came with the box”.
It’s one of the reasons we developed Teneo to be used by non-specialist linguistics. If, like many of our competitors, we expected our customers to have the skill set required to develop even the most basic of natural language applications, then that would immediately make it cost prohibitive for the next layer of content providers. The developers that build the myriad of apps that can be found on sites such as Google Play or the App Store.
We recognised early on that without an easy way to build natural language devices or applications, and for others to collaborate to build the content, then the delay to projects could be anything up to two years. Or in some cases never started.
The reason behind this is complex. Partly it’s due to the problems of finding the right skillset. We’ve been building natural language solutions for over 15 years, we’ve had time to employ the very best computational linguists in the world. It’s their knowledge that is built into Teneo. Another part is being able to analyse the data that we have built up over all these years on how people really talk to technology. And we have it in 21 languages.
It’s enabled us to accelerate the development of natural language projects tenfold and more, allowing many companies to take advantage of the benefits of being at the forefront of technology that is disrupting business models.
Jibo is expected to be shipping developer units later this year to enable the ecosystem to be populated with apps when the production units ship. In July 2008, when Apple first launched the App Store there was 500 apps available. 3 days later there were 800 and over 10m downloads. How easy and cost effective it is for developers to build natural language applications and integrate them into Jibo will be critical to its success.