By Jurica Dujmovic, MarketWatch
Courtesy Everett Collection
Approximately 40% of the global population has access to the internet. The global economy depends on it, and millions of companies and their employees design websites, coding applications and providing SaaS ( Software as a Service ).
However, the job market is changing, and the golden age for computer programmers may soon be over. In the near future, programmers could find tough competition from the very tools and services they designed. Here are some of them:
Drag-and-drop page builders
It makes less and less sense for the owner of a small business to pay a lot for a custom website nowadays. Instead of coughing up a few thousand dollars for development and then more money for maintenance, frugal business owners could spend $4-$50 per month and design their own website, using one of many drag-and-drop webpage builders that don’t require any prior coding knowledge. The website can be built from scratch, or by choosing one of a myriad of professionally designed templates.
All the entrepreneur has to do is to replace default text and images with his own content, and he’s ready to go. Instead of weeks or months of going back and forth with a development team, he could have the site he’s always wanted in just a few hours or days. True, he would need to do it all on his own, but just look at the savings!
And if he decides he doesn’t want the site after all, the amount of money lost is equal to the price of a subscription fee, which is usually much cheaper than thousands of dollars otherwise paid to professional developers.
Another option: let artificial intelligence build your website. Firedrop “uses artificial intelligence and machine learning to automatically create uniquely designed websites around your content. No templates, just bespoke design created in minutes.”
An increasing number of drag-and-drop builders are jumping on the AI bandwagon, with WIX.com Ltd.’s /zigman2/quotes/209382272/composite WIX -4.68% ADI being just one of such builders that could change the face of the World Wide Web forever.
Small businesses and clients who need simpler layouts for their websites will find automated website builders enticing. That in turn could easily cause a mass exodus of web developers from that market segment.
All is not lost, though. There will always be a need for robust, secure, custom and complex solutions that require specialized knowledge only a human coder has. In these situations, a drag-and-drop approach simply won’t do, and AI page builders aren’t quite there yet. Extensive coding and design knowledge are the key here, and a demanding client understands that and pays well for this.
Software developers, beware
It isn’t just web developers who have walls closing in on them. Software developers have reasons to worry as well. Their nemesis could, again, be AI. It all started with AlphaGo , a computer program developed by Google DeepMind and designed for a singular purpose: to win the board game Go. In March 2016, AlphaGo defeated Lee Sedol, a 9-dan professional Go player, winning the match by taking four out of five games.
While developers don’t necessarily play this game, they do something quite similar: write code. Much like a player in the game of Go, who must consider his past moves as well as possible future ones before acting, a programmer needs to be aware of both the old and yet-to-be written lines of code. Otherwise, his code may conflict with the previously stated conditions or may not solve unpredicted problems down the line.
AlphaGo shows it can find the optimal solution for a very specific problem –— how to win a game by having the most playing pieces on the board. The method it uses, called “ Monte Carlo tree search ”, enables it to find the most promising moves by performing and analyzing numerous playouts, recording the most successful outcomes for each stage of the game, and then applying those steps in further gameplay, improving its odds and proficiency.
Mciura, Creative Commons on Wikipedia
Although the algorithm isn’t without its flaws and disadvantages, it’s easy to imagine how it could be employed to solve certain type of programming challenges. The AI would first use deep learning to acquire enough information to run its simulations and analysis of various problem-solving approaches. Then it would run that data through the Monte Carlo tree search and arrive at the most optimal solution for a specific problem. In time, the AI would be able to come up with a solution faster than a human developer ever could. Its suggestions would be neat, without errors in syntax or execution, and it would be a much cheaper approach.
If you think all of this is still a ways off, consider Indian software consultancy and outsourcer Wipro Ltd. /zigman2/quotes/206450547/composite WIT +0.48% /zigman2/quotes/205309648/delayed IN:507685 +0.85% . Since last June, the future of 3,000 of the company’s engineers has become “uncertain” after Holmes (Heuristics And Ontology-based Learning Machines and Experiential Systems), an AI built for automating simple software maintenance, has taken over routine maintenance. Indiatimes.com reports the move will save the company $46.5 million.
After reading all of this, you may think I’m concerned about the future of developers. Not really. I still have the same attitude I expressed in my article about robots taking over our jobs: Although some simpler professions and tasks will be taken over by tools (I consider both the AI and “user-friendly” apps to be tools), there will still be enough room for those willing to improve and take on new and unknown challenges.
These tinkerers of the new age will be able to concentrate on solving more complex and challenging problems, unfettered by the drudge of simplistic, repeatable tasks that will be handled by AI systems. As the next generation of programmers grows and becomes more proficient, they will develop new tools that will then take over yet another portion of their tasks, forcing those performing them to grow and adapt, and the cycle will continue.
What do you think of AI writing code? How about AI running city services? Education systems? Please let me know in the comment section below.