Software Creation is Liquifying
Photo courtesy of Griffin Keller
This post originally appeared on Medium. I wanted to share it here on my blog as well because I think it’s important that we understand the huge shifts happening in software creation which will affect all of the readers of this blog, particularly those who are not developers. It’s a bit long, but stay with it.
Liquify (v) — become liquid or fluid when heated
In his new book, The Inevitable, Kevin Kelly talks about how technology has liquified previously illiquid mediums such as photography and video.
In the past 100 years, the process and knowledge required to take great pictures or make a movie has gone from something you had to study for years and invest many thousands of dollars in to something we all do every day without even thinking twice.
Software, and more specifically software creation, is going to follow the same arc as photography, video and other previously “experts only” mediums for three reasons:
- The advent of ubiquitous tiny brains
- Free cost of creation
- The workers are few
The Onset of AI and Tiny Minds
What I’ve come to realize about AI is that while everyone’s dreaming of (or perhaps having nightmares about) a generalized intelligence that rivals the human brain, there’s actually something far more important happening right here, right now.
We’re able to build any number of what I call tiny minds to take over the parts of our lives where machines can do it better. This extends, in my opinion, to the creation of software in the form of highly personalized APIs and logic.
I’ve had an Amazon Echo in my house for the last few months, and within a few minutes of setting it up I had one of those moments where I knew I was tasting the future. Even more so than Siri, the Echo is a taste (at least for me) of what AI does best right now. It doesn’t do everything, but what it does do it does very well. It is invisible, which is in my mind how you know technology has done it’s job.
So what at all does this have to do with creating software?
If you watch this video by Dag Kittlaus of Viv, you’ll see a glimpse of we’re heading.
Imagine being able to talk to a machine not only to have it do stuff for you (as in this example with Viv or Echo) but beyond that being able to train it just for you.
Instead of using AI and deep learning to store things like images and results, we are going to use it to store logic to execute, and it will be just for us. And it will be just for a specific purpose.
We’ll have API endpoints that only we can use that have logic only for us. We might have a weather endpoint that gives us the weather just the way we want it.
- “I’d like you to tell me the chance of precipitation before you tell me the temperature please.”
- “Can you also tell include the aggregate temperature over the last 30 days in addition to today’s temperature?”
Or perhaps a stock trading API endpoint that we train by being asked questions that combine our preferences with the best algorithms. Imagine Wealthfront allowing me to add my specialized logic to their already tremendous algorithms not by creating a ruleset but by learning over time just for me.
I have a friend who buys and sells cryptocurrency all through a trading algorithm that he’s written specific to his tastes and risk profile. Why shouldn’t everyone be able to do that, and train the program just by talking to it?
The idea of personal, ephemoral code spaces where we create our own little programmable universe is within sight.
The beauty of this new universe is that, just like we don’t care about the know-how or technology required to take great pictures anymore, we won’t care about how to write great code.
Our tiny minds will do all of the work for us.
Free Cost of Creation
You might not take as good of pictures as Vivian Maier, necessarily, but they’ll be pretty good. Same with video. Your phone plus some simple editing apps and voila! Instant classic.
With software, however, there’s still a massive cost to create something that is truly ours. Companies build us applications that we can personalize to some extent, but they’re still not really ours. Developers get involved to build even the simplest of websites or programs.
That is changing, however, right underneath our noses. We’ve gotDropsource, Zapier, IFTTT and others that allow you to write code without writing code. They’re the latest generations of software and business process automation tools.
If we have our own code spaces with our own tiny minds, we’ll need some tools to create and manage our space.
We’re going to use our voices a lot. That’s what I’ve learned from Echo. It’s so easy to do most of the work just by talking to the machine. Most people will say “Yeah, but speech recognition is still not great.” True, but remember my speech recognition engine only has to be able to understand me, not everyone who speaks English, and I’ll be able to train it on the go.
When we don’t use our voices, we’re going to have tools as easy to use as Instagram right on our phones. Drag some stuff around, configure some flows and save it to the cloud for our personal APIs.
Or, perhaps even better yet, we don’t really ever have code that we’re “writing.” All we’re really doing is providing some guidance and our code is created and executed on the fly.
The Workers Are Few
I won’t belabor this point, but there’s simply no way that we can train enough software developers to create all of the software that we’re going to want to create over the next few decades.
As billions of new devices and sensors onboard to the internet, along with the massive data they’ll produce, we simply won’t have enough to humans do all the work. We’ll have to rely on machines to take over 90–95% of the work, and leave the humans to do the parts only we can do.
As we’ve done in the past, we will invent our way out of this shortage.
From Users to Creators
We’re getting ready to enter the era of truly just in time software. For the first time in history, it’s possible to conceive of, build and run code on the fly across the network in less than a second.
We’re seeing it with Viv already, and we know from other digital mediums that, in the end, everyone will create whatever it is they dream of from the comfort of their own palm.