• Interview Rafael Yuste, ideologist of the Brain project: "Humanity will climb on the chepa of artificial intelligence"
  • Innovation The 'cobra effect', or why the obsession with rankings and metrics generates corruption

Millennials were the future until the Zetas came asking what Tuenti was. USB sticks and DVDs were modern and functional supports for information until the cloud appeared. And the same path followed the fax, the mp3 player and the low cost digital cameras.

The speed of transformation of technology has accelerated so much in recent years, that less and less time passes since it is sold to us as the great disruption that will change our lives until it disappears completely from the map. Vertiginous changes that also occur at the economic, political and industrial level. To the point that they can even catch the President of the Government himself with the changed foot.

On May 5, Pedro Sánchez announced in Alicante that the Executive would approve an item of 560 million to improve the quality and equity of the Spanish education system. Almost 300 of those millions would be used to train schoolchildren in robotics, programming and new technologies. Mainly, incorporating the necessary equipment in schools and in giving support to teachers to provide this new type of training. The goal, Sánchez said, is for 5.5 million students to "learn the language of machines," programming and robotics, because it's the world they're going to live in.

It seems undoubted that today's children are going to develop in a world of robots, but... Is it really going to be that important in their lives to learn programming? More and more experts believe that, precisely because we are entering the world of artificial intelligence (AI), knowing how to write lines of computer code in languages such as Javascript, Python or C ++ may not be a great advantage in terms of job opportunities.

THE END OF THE HUMAN CODE: IN 2040 OR IN 2024?

"Programming trends suggest that software development will undergo a radical change in the future: the combination of machine learning, AI, natural language processing and code generation technologies will improve in such a way that it will be machines, and not humans, who write their own code in 2040," said a study by the US research center Oak Ridge National Laboratory in 2017.

But 2040 seems too distant a horizon when ChatGPT and generative language models based on AI are already a reality within reach of any smartphone. To the increasing automation of functions that developers have been incorporating in recent years have been added more recently AI tools capable of completing or generating complex code strings themselves.

For example, Github Copilot, emerged from the collaboration between OpenAI and Microsoft and that many developers around the world are already using to replace the most tedious parts of their work. Or AlphaCode, from Google, which competed in a competition in which it had to respond to a series of code challenges and its results corresponded to that of the best 54.3% of humans.

"They are very powerful and they will improve for sure, this is not a sector that is paralyzed by an atavistic fear of the consequences of technology, it is the one that creates them," says Javier G. Algarra, academic director at the U-tad University Center and an expert in AI.

And at that speed of change and refinement of technology, the functions of human programmers could be blurred long before 2040. "There will be no programmers for five years," said Emad Mostaque, CEO of Stability AI, the creator of a popular AI engine to generate images and music from text. He said this during his speech at a conference organized by Goldman Sachs in which he also stressed that "41% of the new GitHub code is already generated by AI."

AI WRITES 55% FASTER CODE

Mostaque's estimates might be somewhat exaggerated because he runs a company with interests in the sector, but what if they are true and we have the end of human programming just around the corner? What if it is no longer necessary to teach our young people to speak the language of machines, as Pedro Sánchez proposes?

"I wish it had been done five or 10 years ago, but it's not too late as long as you focus well," says Senén Barro, professor of Computer Science and AI at the University of Santiago de Compostela. " My generation had to learn to speak the language of machines, but if ChatGPT is revolutionary it is precisely because it is machines that have learned to speak ours," he adds.

That is, it could no longer make sense to spend years mastering languages full of commands, symbols and logical levels if we can ask the AI to write lines of code from some indications, written or verbal, in our language. "I tell my computer science students: 'You're going to live your whole life of computational thinking and applying it to solve problems, not writing code,'" Barro explains.

In this need to pilot the AI, raise needs and monitor its possible errors is where the hopes for the future for programmers would lie. "AI tools are a great help for the productivity of developers, who can write better code faster, but in no case will they replace them," says G.. Algarra from the most optimistic vision.

A study conducted by Github among programmers who used Copilot revealed that they generated code 55% faster and that 88% of them felt more productive and 74% said they were happy to be able to focus on less routine work. And that is positive, no doubt, but if today, with AI tools that can be improved, code is written 55% faster, it is most likely that at some point at least half of the programmers will survive.

"It is often said that there will always be tasks that will not be automatable and that people will have to do," argues Barro. "And it is quite likely that we will not reach 100% automation in decades, but if 80% is reached, it is clear that fewer people will be needed to dedicate themselves to programming," he concludes.

COMPUTATIONAL THINKING

According to estimates by the consultancy Evans Data, in 2021 there were 26.9 million software developers in the world. That half or two-thirds of them suddenly find themselves unemployed could be a real employment drama.

In that context, should we encourage our young people to focus on learning programming? "If the question is whether it is good for young students to learn to program to open professional doors, the answer is 'no'", ditches Senén Barro.

However, both he and G. Algarra believe that measures such as the one proposed by the Government are still necessary, although reoriented towards computational thinking. That is, not to teach schoolchildren how machines speak, but how they think, because that is where the job opportunities will be.

The importance of moving towards school 4.0

J.B.

The measure announced by the Government to promote robotics and programming among schoolchildren is part of the program that has been baptized as School Code 4.0. The initiative arose in line with the visit of the CEO of the Code.org platform, Hadi Partovi, to Pedro Sánchez in June 2022. This NGO already teaches programming to one in three students in the US and campaigns to promote learning about programming around the world. Personalities such as Bill Gates, Barack Obama and Mark Zuckerberg have supported them. In Spain, they already have a program with the Junta de Andalucía and have been supported by Queen Leticia and Ana Botín, president of Banco Santander, among other authorities.

According to the criteria of The Trust Project

Learn more