The end of developers

Computer Assisted Programing (CAP) will soon bring a rapid fall in the hiring process of low level developers and radically change the digital industry.

When AI Frameworks do the coding

So easy for an AI to learn coding

Expert from IBM, Yahoo, Google, Amazon have spent the last decades teaching machines to understand natural language: human language. It is a very hard task as human languages are used with so many unexplicit rules.

Guess how much easier it is to teach an AI agent the basics of python, javascript, php or C#? The syntax is fairly rigid and accessible.

Who has the data, wins the prize

To teach an AI, you need data.

For natural language processing, you need a lot of exchanges with your customers (written in products reviews or in phone exchanges). From these fairly short samples you can start doing some predictions about the way a customer expresses her/his like or dislike of a product.

For coding, it is even simpler. If you are the owner of a large github like repository, you just have to plug your learning agent on the millions of sources already at hand. That’s a pretty nice data bank.

Already parsed and tested

And even better. Human data had to be tagged by hand (this product comment is “neutral”, “enthusiast”, “displeased”) before you could use it to teach a machine to recognize the emotion. To teach a machine coding, you would just have to look at the thousands of unit testing results to see what a “good” code is, and what is a bad code.

After that, don’t tell me you really need a real developer to do your setters and getters. You can reverse engineer any app on the market and teach the machine to create the most efficient code.

Next framework will be self creating (created)

Although I am still interested by the old frameworks (phonegap) or the newer ones (Facebook’s react or react-native), I believe that the next generation of framework will be based on an assisting coding agent. The best libraries in the future market of coding platform will be based on interfaces to efficiently write the path from a user case to the code.

Developers, as any manufacturing workforce already discovered it, can be replaced. Well, part of them. This means that among the different degrees offered on the market you would have to make the right choice for your children. This is of course absolutely not visible in the hiring statistics trends of James Manyika from McKinsey… just a nice low and discret signal.

When the machine designs its database

Today’s algorithms for managing data are almost made by computers. I mean, the best performing data management system is really created by the machine. Not quite alone, but it is on its way. Humans are still launching the calculations.

But next generation of data systems will be created from a framework where the machine tests billions of solutions and produces the best system. Check how amazon AWS works and you will understand how everything is self regulated. Next step is from regulation to design.

Rethinking the orientation trends

Although not appearing in Bill Briggs and Craig Hodgetts’ report from Deloitte their seems to be a possibility that the future (not immediate) of the digital age faces a turn.

Two signs of the fall of developers to expect in the next 5 years (well usually, I am always wrong with the timing):

1- the AI coding frameworks: sets of libraries that writes the code or make it way more efficient

2- the UX change to communicate and code: far from codepen, a complete design change is coming up, with a totally new set of interfaces to develop software. Not by coding directly, but by letting the machine code for you


For your next five years of study or work, learn UX to create new ways of working with machines, and learn how AI is coming up to mutate any job actually existing. If you are a developer, anticipate the change. Get higher in the food chain by understanding how AI is going to code for you.