Voice applications and natural language are bigger than the kitchen.

by Bryan Knouse, 29 Oct 2018

Voice is the interface, not the application.

We are the first generation since the dawn of the spoken language to have more nonverbal conversations than spoken ones. We leverage language through our eyes more today than our mouths and ears. As a result, screens and keyboards have transformed nearly every aspect of the workplace and daily life…

…and then came Alexa.

Approaching human parity in natural language processing through digital electronics is an inflection point of our experience and communication with technology. As a “hard” technology, commercial realization is not quite there yet, but it’s close. However, this hasn’t stopped the public from going crazy over voice-user interface products.

The best-selling product on Amazon last year, on Amazon Prime Day, and this past holiday season? You got it — the Echo (Alexa) device. Amazon leads the smart speaker market with the Echo line, and has introduced video screens with the Echo Show. Microsoft, Google, Facebook and others have followed suit, offering their own line of voice-user interface products as well. Apple most recently released the Siri-powered HomePod. These products offer unique advantages in ease and efficiency that just aren’t possible with standard user interfaces. 

Voice -natural language is the fundamental means of communication available to us.

Sure, these products make you look super cool when you have guests over — telling your house robot to change the channel — but there’s a problem with these snazzy new gizmos. .  They just don’t seem to do much. This comes down to understanding a simple dichotomy: The application and the interface are two separate entities.

You may have heard the saying “we don’t need a quarter-inch drill bit; we need a quarter-inch hole.” The product we want is a helpful customer support experience, completed by interacting with a machine. We want to manage HR processes by working with the system. We want to draw insights and predict business expectations from data by asking logical questions. We want to build web applications by telling the machine what to create and how it should look. Natural language is the most effective interface for execution.

When chatting with the enterprise community, I might say, “voice will change your office” to a cascade of criticisms: “Do you really think people will sit around talking to their computers all day?”  It’s not about the talking, it’s about the language you’re using for it. 

Natural Language is Bigger Than Your Kitchen

There are quite a few new technology advancements in the digital space, but generally speaking, I see four big efforts right now — AR/VR, decentralization/crypto, natural language and more generalized AI (a fuzzy amalgam of neural networks and data science).  Conversational technology is striking among its peers in its current capability and breadth of application. If we think about which of these has “crossed the chasm” to mainstream adoption, natural language technology is the most advanced in its developmental life cycle.

Natural language technology has been under development almost as long as the computer itself. Today, Alexa and the gang are everywhere — you can even talk to your microwave!  The direction we are heading is inspiring, but one cannot help feeling that Alexa isn’t all that ‘smart’. (That goes for Google Home, Cortana and Siri as well.)

Despite the need for better intellect, we’re still able to communicate naturally with our machines. That sounds cool, with obvious implementations in hands-free situations, like cooking in the kitchen or operating a vehicle. With conversational tech, the kitchen is an obvious first implementation. However, it will neither be the most useful application or final form.

Malleable Technology

As a developer, it might be appropriate to say that we “speak” computer. Many times I have felt that I don’t have a tech job or an engineering role; rather, I’m a translator. I take ideas made of human language and turn them into similar ideas made of computer language. A client might provide visual mockups (pictures and layouts) to a UI developer and say, “Turn this into a computer format with HTML and CSS.” Ours the process of turning this visual idea into something the computer understands and is able to utilize in another environment. This labor is complex and tedious because the end result requires deep domain knowledge as well as fine-tuned execution. In other words, a UI developer needs to understand how to write serverless applications in Javascript while catching the extra semicolon on line 1009.

The strongest advantage of natural language is that we may continue to offload this translator labor to the machine. Between the two of us (the computer and the human), endeavors such as the translation of human idea into code is moving further and further into the domain of the computer; the machine is becoming more responsible for the work everyday. And we are released more to simply tell the machine what to do. Similarly with programming languages, each newer and simpler language is a further abstraction out from machine code and thus easier and faster to leverage as a tool. Until something like a neural lace becomes possible, the furthest abstraction and most fundamental interface there can be is natural human language. In the process of doing this, our experience with technology is becoming less brittle, less prone to minute errors causing major fractures. This is the result of many things such as scalable infrastructure, more powerful computation and stronger interconnectivity. This growth has inspired software and technology that is beginning to feel malleable — much like language itself.

This pioneering change in how we leverage computers will be profound. Today we’re in the kitchen. Tomorrow; data analytics, web development, management, and customer experience. Beyond that — who knows? Alexa certainly doesn’t. As long as we continue to develop the most natural interface available to the human-machine duo, the number and impact of applications made possible will flourish. Even the ones that make us cringe.

 

Bryan is a software developer and entrepreneur based in New York City.  Over the last several years, he has focused on natural language technology as a means to improve the speed and simplicity of enterprise software and data analytics.  Want to talk software, data, or natural language technology?  Reach him at bryan.knouse@gmail.com

Want to learn more?

Contact us!