“Do You Hear The People Type”?
The Need for an AI Revolution
First comes technology. Then comes utility. Then comes integration. Then comes revolution.
This is the pattern behind the transition of most new technologies into general adoption. And I believe it’s the key to us taking artificial intelligence to the next level. However, before we can apply this to our current challenges with AI, let’s see how it’s worked in the past.
There are quite a few names that are ubiquitous within the computing world, but few might be as recognizable as a singular fruit: Apple. Over the last 40 years, they’ve gone from a garage project to the most profitable company in the world that has a overall worth of over $860 billion. What makes them so successful? What’s their lightning in a bottle?
First comes technology. Then comes utility. Then comes integration. Then comes revolution.
First Comes Technology
Whenever a new technology emerges, there first appears to be a “big bang” moment. What I mean by that is that regardless of whatever person or group of people is responsible for the emergence of a new technology, there quickly appears an arms race to duplicate and improve that technology. We’ve seen it with microprocessors. We’ve seen it with operating systems. We’ve seen it with coding languages.
And for Apple with the Apple I, it was with the 8-bit processor chip. The first advantage Apple gained in the personal computer race was to provide high-quality tech. Between Tandy, Commodore, and Altaire, Apple didn’t have to be the best — they needed to be one of the best. They just had to be in the game. However, the Apple I wasn’t the product that brought the mind-melting revenue that shot Apple to the forefront.
That’s because an 8-bit processor chip, by itself doesn’t accomplish much. Technology without a use case will die in the water.
Then Comes Utility
At the end of the day, a solution without a problem isn’t a product — it’s art. It’s something to be seen and appreciated, but unless it is actually useful, it won’t impact people’s lives in a significant way. That’s why the key to stepping out of the “big bang”, with its raw energy and new technology, is utility. The first company or person that is able to take those raw materials and make them into something useful will win the day.
Apple did this with the Graphical User Interface (GUI). In today’s world of incredibly visual experiences, it’s sometimes easy to forget that the idea of driving computational power through a visual interface was a rather novel idea. Code and command line used to be the only way to get that machine to fire those processors.
And while the true genesis of the GUI for computing is attributed by most to Xerox, Apple can be attributed with applying that paradigm to personal computing. In doing so, these incredibly powerful (for the time) 8-bit processors went from being a science experiment to being something that could actually accomplish tasks.
Then Comes Integration
In Apple’s case, and most cases, integration is the real turning point. For Apple it came in the form of Visicalc. The most important aspect of Visicalc is that it was not developed by Apple. Visicalc was a spreadsheet application that was developed by Visicorp, and it’s release on the Apple II was the catalyst for widespread adoption of microcomputers in the business and professional markets.
At the end of the day, no company, no matter how wealthy or powerful or successful, is able to accomplish everything. So when it comes to converting emerging tech into a revolution, the key is in integrations. Systems have to be built in such a way that third party developers can take advantage of the work that’s already been done.
Then Comes the Revolution
Finally, all of these developments in personal micro-computing led to a singular event: the 1984 Apple Mac. It was the culmination of a good technology, connected with the utility of a GUI operating system, and packaged with applications that would set the stage for Apple’s overall success 40 years later. Through that revolution came personal computers and all of the associated conveniences and innovations.
AI State of Affairs
Artificial Intelligence has tentatively started down this path. But we are soon arriving, or have already arrived, at a point where it needs a good aggressive shove.
First comes technology, and this is where we are right now. With Artificial Intelligence (AI), Machine Learning (ML), and Natural Language Processing (NLP) we are in an arms race right now. Here at Adobe we have Sensei. Salesforce has Einstein. IBM has Watson. Microsoft has an AI platform and the LUIS framework. The technology is out there and it seems everyone has a flavor of it.
But in the AI world, we’ve seen a few companies starting to turn this raw computing power into utility. By and large, it’s still mostly just a service for businesses. We haven’t seen a lot of AI making it’s way into the consumer world yet. The one exception to that rule might be the recent increase in consumer facing NLP with services such as Siri, Alexa, and Ok Google, but it’s important to remember that NLP is just one small slice of the entire AI pie.
The integrations step for AI is an even bleaker landscape than the utility. Again in the AI world, the only example we can see of this currently is in the world of voice. Alexa’s skills are the perfect example. But where is the development and experiential framework for deploying AI visually? Who owns the chatbot framework that has enterprise scalability and inter-modality?
We’ve got a long way to go before AI, ML, and NLP is anywhere close to a revolution. The question isn’t if we’ll get there — I know we will. The question is how.
The AI Revolution
As previously stated, we’ve landed squarely in the technology stage. What we need to do next as engineers, makers, thinkers, and designers is bring utility to the technology. What does AI really look like for a user of Adobe Photoshop? How is machine learning delivered to a sales rep in Salesforce? What functions can your grandma accomplish using natural language in Gmail?
Once we have a firm grasp of use cases like these, the next big step will be the delivery format. Voice-enabled AI has come a long way in the last two years, but there are still huge gaps in every other area. How does AI manifest itself in a product UI? How does it manifest itself through a phone app?
When those use cases and that delivery platform exists, we can then stitch these experiences together: voice, in-product, message delivery, in-app, and passive interactions. Then — and only then — will we be truly prepared for integrations, and that’s when the true magic will happen.
We don’t need more stand alone chatbots.
We need networks of chatbots that communicate within various environments. We need AI platforms that can deliver content using various technologies. We need a system that can be extensible enough for integrations.
We have the technology. We can find the utility. With the utility we’ll encourage integrations. And then we’ll have the AI future all of us are looking forward to.
