The AI ​​tool that changed my life as a developer

Author, Jason Norwood-Young

I have a confession do: I haven’t done my job since the beginning of the year, but I still get paid for it.

A big news site pays me money to do software development that I don’t do. The software is still in production, which has so far avoided uncomfortable conversations about my continued employment, but I’m not doing the work; well not all anyway.

Instead, my artificial intelligence boyfriend, Github co-pilotis at the forefront, typing lines of code while I supervise, adding a bit of guidance here and there, and counting my money.

(I’m exaggerating a bit for effect – I think Copilot writes about a third to half of my code, depending on the task. Most of the problems we’re trying to solve in code have already been solved, and being the largest code repository in the world, Github has this solution at your fingertips.)

The promises of AI have been as varied as the prophecies about what will one day be our digital overlords: they’re supposed to drive cars, replace all lawyers and accountants, fly autonomous drones, and identify enemy fighters to shoot down.

They will either create ultimate labor freedom for humanity, destroy the economy and put us all out of work, or simply turn the planet and everything on it into microchips, according to the science fiction or futuristic book that you listen.

start to feel

In practice, however, they’re still pretty bad at driving, and unfortunately they’ve completely failed to eradicate lawyers or accountants. Frankly, AI was starting to smell as dodgy as bitcoin, the metaverse and Brexit customs checks – a shared delusion that a technology was capable of so much more than it really is.

Then I met my friend Copilot. I put it on my code editor, Visual Studio Code, a development environment created by Microsoft, which, not by chance, bought Github for US$7.5 billion in 2018. At the time, I was coding in PHP, not what I use every day, so I was a bit rusty. I thought this new coding AI might remind me to put in missing semicolons, which PHP is obsessed with. If you remember to put a semicolon at the end of each line, you already know most of what you need to know to code in PHP.

It didn’t remind me to put semicolons at the end of the line. Instead, I wrote a function name, pressed enter, and that produced all the code the function should do. Immediately. My jaw hit the ground.

I tried another function I needed. Again, he filled in all the details. Then I tried to write a plain English comment about what I wanted my code to do. He wrote the code. After a while, he also started writing the reviews.

To explain why I found this so incredible, let me digress a bit from 18th century Eastern Europe. Wolfgang von Kempelen, a Hungarian civil servant, polymath and part-time inventor, presents Austrian Empress Maria Theresa with an automated chess game capable of playing a very good game of chess against a human. Named The Turk because it prominently displayed a life-size model of a “Turkish wizard”, this machine circled Europe for 84 years, beating Napoleon and Benjamin Franklin, among others.

The article continues below…

From the book which attempted to explain the illusions behind the Kempelen chess-playing automaton (known as The Turk) after making reconstructions of the device. Author: Joseph Racknitz. Source: Humboldt University Library

Keep in mind that we didn’t have chess computers until the 1950s, and they sucked until the 1980s. It really was a marvel of its time. And, of course, a complete sham. Cleverly hidden within the mechanism was a very cramped chess master, which moved the pieces under the board via a magnet.

Until Copilot, most AIs I worked with were something like this – either it was unnecessary, or a smartly written traditional system that used a tower of logic and processing power to simulate the AI, or literally humans pretending to be machines, providing answers to questions on the internet. Amazon even has a product called Mechanical Turk (after our chess engine from the 1700s), which allows you to send computer-like tasks to humans for meager returns.

That’s why I was both so impressed and suspicious of Copilot, and to be honest, wondered where they were hiding the little chess master.

So what does this mean for accountants, lawyers, fighter pilots and, of course, software developers? Are we all going to lose our jobs and turn to basic income benefits to survive? Copilot, I believe (and hope), shows us another way: man and machine working together, to be more than either can be alone.

Copilot, I believe (and hope), shows us another way: man and machine working together, to be more than either can be alone

If you think a software developer’s job is to produce code, then Copilot is definitely a threat to software developers. If you think a software developer’s job is to produce solutions to problems, then that’s an incredible leap forward.

I can free my mind from the minutiae of syntax and semicolons to think on a new level, that of creating solutions to problems. I code twice as fast, and what I output is significantly better. And I think it will be the same for other industries targeted by AI: it is a tool, not an existential threat.

Copilot has been in closed beta since the start of the year, but on June 21, they opened the doors and are taking subscriptions. It costs $10/month, and honestly, I can’t give them my money fast enough. Because, really, it’s not my money, it’s my boss’s.

  • Jason Norwood-Young is a journalist turned software developer. He creates software that helps news publishers continue their work of creating news, profitably. He is currently founding a startup to provide WordPress-based news publishers with a suite of tools for publishing, marketing and subscriptions. Originally from South Africa, he lives in Utrecht, the Netherlands

Source link


Comments are closed.