Smartisan TNT's Revolutional Product Interaction Exploration
Smartisan Technology released their new product TNT workstation today in the morning. Major Chinese media are very disappointed or even cynics about this new product. But after watching the product launch video, I felt very excited about the revolutional interaction possibilities that Smartisan Technology brought to us：
Focusing on more specific context helps with speech recognition. For example, if we ask Siri "What's weather today?" Siri needs to classify the request into the weather domain before being able to search for today's weather information. But if we've given specific context like "weather" to Siri before asking "What's weather today", Siri will perform better since it is limited to a very small scene and specific context. With limited hypothesis space, the speed and accuracy of recognition can increase dramatically, which is a very good direction for creating voice products with better user experience.
Next step forecast
In Smartisan TNT's use case, entering a sentence in a slide will be followed up by a list of recommend sentences that you may like to enter next. This feature is similar to the email product released by Google in Google I/O 2018. Smartisan Technology have done a better job: On the one hand, it can not only recommend the next sentence when typing, but also can associate images with inputs for richer usage scenarios. On the other hand, they make full use of large screens giving a selectable list rather than a single candidate. In 2017, some products with this feature have begun to emerge, such as Facebook's M in Messenger. In 2018, I think the "guess the next step you want" feature will virally spread across all kinds of products gracefully embedding AI into products.
Large screen parallel
Searching across multiple APPs at the same time is just a very natural and necessary thing, but it has never been done before. Smartisan Technology has done such a natural thing showing how we can parallel messenging, searching, and viewing information across platform in parallel. Why don't we get this feature on Mac? Maybe the overdose of mobile web experience has limited our imagination.