Artificial Intelligence and the Death of Programming
Don’t let that cute little logo fool you. That’s the grim reaper of programmers, and he’s coming for you.
Github Copilot is Microsofts AI tool for autocompleting code. It was released for technical preview last year, and is now available as a subscription-based service. As you can see from the below image, it works, and it works really well.
So the day that many programmers have feared is here. We are now obsolete. The app market is dead, offshore coders are numerous and cheap, and now on top of everything, Microsoft has created a robot that writes better code than your best programmer.
Of course, none of that is true. As always the truth is far more nuanced than that.
Let’s take it back to the 1960s for a second. Programming was nowhere near the sexy and glamorized career it is today. Computers were expensive, so companies could only afford a couple to be shared between their employees. Back in the day, a programmer would type up his code on a sheet of paper with no assistance from autocomplete or ahead-of-time compilation, nothing to aid him other than the brain between his two earlobes. He would then take this fragile piece of paper to a laundromat style batch processor, where he’d have to wait his turn with the other programmers to punch in his code. Once his code is entered into the compiler, he’d grab a cup of coffee and wait a couple of hours for his “Hello World” to be outputted (this is all hilariously represented in this Stanford short film). It was a mind-numbing and painful process, reserved for the select few who could bear it.
Cut to the 21st century where computers are now prevalent. These days you have massively powerful IDE’s (software for writing software) that make writing an application so easy any mathematically inclined cave man can do it. No more waiting in line to run your program, no more writing chunks of untested code (unless you’re a maniac and you still use Notepad). You don’t even have to be in the same building to collaborate on a project. What’s the effect of all these luxuries on the industry? What happens to the field of computer science when you make programming exponentially easier?
In Q1 of 2022, technology still holds 5 out of the 10 most in-demand jobs according to LinkedIn. As the barrier of entry decreases for computer programming, the demand for apps increases. It’s now easier than ever to implement technology into your business, and everybody wants to do it. Having a technological edge is imperative to compete in todays market. Hospitals are building their own apps that track patients medical data, law enforcement is relying increasingly more on software to track criminal data, and every school and university now has a digital infrastructure to track student data. This explosion of data gave birth to new highly successful fields like Big Data and Data Analytics, and these giant wells of data also need to be safeguarded, hence the dawn of Cybersecurity. Very few people could’ve predicted the advent of all these fields. Making programming easier only proliferated it more. Humanity is building a new technological future, and we’re not even close to done yet.
And now we return to Github Copilot. My honest review is that it’s totally amazing. It’s cut down on a ton of remedial, boring tasks that programmers have to typically slog through (typing up repetitive stylings, finding the right sort for your list). If I’m being completely frank, what Copilot really does is save you a lot of trips to Google, because now instead of looking up the solution and copy-pasting it, Copilot will spit it out right in your IDE. So it might make StackOverflow obsolete (definitely not, StackOverflow is our king) but programmers will be alive and well.
There are a few notable weaknesses to Copilot. The most obvious one is that it’s wrong…a lot. In that sense it reminds me of Siri. Even though Siri is a great technology, the constant mishearing and mishaps are endlessly frustrating, and potentially catastrophic when you end up texting your boss instead of your wife. Copilot have the same unintended consequences, and it’s a heck of a lot harder to debug production code that was largely written by an AI. So there are quirks and bugs like with any new technology, but once it becomes a little more stable I think it will make programming a joy, and I’d probably utilize it everyday. It takes away a lot of the frustration that comes along with coding in the form of repetitive tasks and grunt work, and that’s unequivocally a great thing.
I feel like people get the wrong idea about automation. They think that all the jobs will vanish, and there will be nothing else for humanity to do.
I believe it will be quite the contrary. All the remedial jobs, the boring work, and the repetitive tasks that are beneath the true capacity of human beings will be gone. AI will give us more free time, and out of that free time will spawn an explosion of creativity and innovation unlike anything ever seen. Automation will free mankind up for a higher purpose, something machines will never be capable of.
But that’s way into the future. No one can see that far. The nearer future will look more like this: automation will spawn an explosion of new jobs, just like the advent of the computer spawned an explosion of new computer related jobs. Computer Programmers, Data Analysts, and Cloud Engineers now dominate the market, but these are jobs that were unthinkable before computers. It’s the same for automation. We can’t imagine the work involved in maintaining AI, developing new features and fixing old ones, but I guarantee you that work will be there. It’s foolish to think that this will only destroy jobs. It will destroy old and useless jobs, but it will also give birth to a massive amount of new and important ones. Just like the cotton gin, just like the steam engine, and just like the Macintosh before it.
It’s just evolution. This has been the case for every technological innovation since the dawn of time, and it will continue to be the case.
As always, there’s nothing to fear ✌️
