Back when I was a kid in the late ’80s and early ’90s, the best way to get stuff done on a computer was with the Microsoft DOS command prompt — a text-based operating system for computers that looks positive antiquated today.
It looked like this:
This so-called “command line” interface has a surprising amount of flexibility, letting you do just about anything on your computer in just a few keystrokes. In fact, programmers in particular often still use the command line to get work done. Plus, you feel like a super-cool hacker from “The Matrix” when you successfully navigate it.
But they also have a ridiculously steep learning curve, requiring you to memorise a dictionary’s worth of commands to get anything done. To move from folder to folder, for instance, you might type “cd c:/users/matt/businessinsider/articles.” It’s far from impossible, but it’s complex and intimidating.
The command line gave way to graphical user interfaces like Microsoft Windows, which eventually evolved to include touchscreens and the web browser. Now, even Microsoft thinks that chat interfaces like Amazon Echo, Facebook Messenger bots, or its own Cortana are going to be the next big thing.
DOS, part 2
The promise of voice assistants and chatbots, Microsoft CEO Satya Nadella said in early 2016, is that instead of learning how to talk to computers on their terms — like memorising keyboard shortcuts, web addresses, or long lists of DOS commands — we can finally make computers understand plain speech, the same way humans have communicated naturally for millennia.
It’s a great idea. But so far, that hasn’t really materialised. I recently got my hands on an Amazon Echo Dot, my first-ever voice assistant, and I’m struck by how much it’s like those early days of DOS. There are a very finite number of things Alexa understands, and you basically have to memorise all of them to get anything done.
That’s mostly fine for simple tasks like setting a timer, adding things to a shopping list, or playing a song. But Amazon Alexa is getting new “skills” all the time. Some of them are useful, some of them are fun, some of them are just silly, but there are literally thousands of them. And because there’s no screen, there’s also no reminder of all the things that you’ve enabled Alexa to do.
I’m not the first one to notice this, either: “If it takes you more time to work out what you can ask the AI assistant than to drag the meeting to a new slot on your calendar, you’re doing it wrong,” Andreessen Horowitz’s Benedict Evans wrote in 2016.
What happens next
So, sure, I can order an Uber with Alexa, or play 20 Questions. But I literally have to have all of it memorized, and know the exact right sequence of words. Thanks to an integration with IFTTT, a service that lets you hook apps up to each other, I can have Alexa ring my phone — but only if I remember the magic words “Alexa, trigger Find my Phone.”
That effect seems to be having negative effects on the growing Alexa ecosystem in general. A study reported on by Recode indicates that 69% of the 7,000-plus Alexa skills out there have zero or one customer reviews, indicating that nobody is using them. And even if they use those skills, the study says, they drop them after a week or two. A lot of Alexa skills are created and then abandoned, it seems.
As with DOS in the ’80s, the current market for voice assistants is immensely powerful. But also as with the earliest days of the PC, it’s hard for anybody but the most dedicated power user to get the absolute most out of these devices.
This doesn’t mean that voice assistants are doomed — quite the opposite, in fact. If this super-early version of the voice assistant concept is performing as well as we know it is, it actually proves that there’s something really big happening. If some people like Alexa now, with these oddities, quirks, and roadblocks, imagine how much they will like whatever the next version looks like.
What that next version looks like, who knows. Advances in artificial intelligence mean that Alexa, Cortana, and Google Assistant could get way better at knowing what you mean and trigger the appropriate app automatically; Amazon is said to be working on some kind of Echo device with a screen, indicating a new kind of interface.
It also means that there’s room for someone else to come in and beat Alexa, despite Amazon’s early lead. And with Apple still quietly revving up its voice assistant strategy to combat the rise of the Amazon Echo, maybe the Cupertino company can come in, build the Windows to Amazon’s DOS, and spark a repeat of the PC and smartphone revolutions that changed the world.