By Agis Veroutis*
In software companies around the world, Artificial Intelligence has begun to take over the jobs for which until yesterday they were hiring developers.
With a connection to the Codex programming assistant on top of Microsoft's Visual Studio Code, which many developers use, a developer's daily code output of about 80 lines can skyrocket to 1,000 lines and more. Of course, this code must work properly and someone has to check it.
According to Jensen Huang, CEO of Nvidia, which makes the Artificial Intelligence computing hardware, the one that enters data centers, its developers are turning to building AI agents that write the code. He likens writing code to typing, but anyone who has written even 50 lines in Python or even C++ knows that much more than typing commands, what matters is the architecture and the flow of data within each program.
Finally, when the AI agents that will program other AI agents are created, will the developers be left without a job?
Developer Recruitment
Already, in several Greek software companies, the recruitment of developers has been frozen, while those who leave are not replaced. With the huge leap in productivity created by the use of AI for existing companies, it won't be surprising to me if in less than two years these companies are operating with less than 15%-20% of the developers they employ today.
Already in the large software companies abroad, the number of hires of apprentices or junior employees has decreased to 1/9 according to some approaches of people working in the field. This raises the question "where will the new entrants learn the job", so that the day after tomorrow they can staff the needs of companies for jobs with more experience? Today's junior is left over, while tomorrow's experienced one is expected to appear ready, trained at someone else's expense.
This does not necessarily mean that the developers and software engineers who leave them will be completely out of work. Many may decide to take advantage of the enormous productivity offered by AI and build new software companies, with a much smaller number of employees and equal production to the companies they left from. Their former employers will have reduced payroll and gained some new competitors.
Multiplication of productivity
This medley of programming talent can fuel a rapid acceleration in the creation of new AI agents, which will multiply the productivity of other professions. The number of employees needed for the same workload will be reduced or the productive result will be multiplied by fewer people.
Who will receive this extra productivity is the crucial issue. Will we see it in lower prices, in higher wages, in fewer working hours or mainly in the profits of those who own the systems? AI can produce more. The distribution will be decided by people who have interests.
Of course, in the Second Industrial Revolution, which began in the second half of the 19th century and brought electrification and mass production, new industries and professions were created. The increase in productivity has allowed many more people to engage in education, health, research and services that previously few could afford.
This, of course, was little consolation to the worker who delivered ice columns to homes and lost his job due to electric refrigerators, or to the stables cleaner when the horses disappeared from the streets and were replaced by cars.
Similarly, email reduced the need for letter delivery, while parcels gained their own huge market. New jobs could be found elsewhere and require knowledge that the laid-off worker did not have.
Not even drivers?
It is to be expected, with the advent of self-propelled vehicles in cities initially and everywhere afterwards, that many positions for taxi drivers, truck drivers and all kinds of drivers, such as garbage trucks, will gradually disappear. Tesla has already taken the Cybercab out for use as a taxi, without steering wheel or accelerator and brake pedals, not even a driver's seat.
A Tesla self-driving taxi in Austin, Texas will pay for the vehicle's depreciation, maintenance, insurance, cleaning, support, and the electricity that will charge its batteries. The driver's payroll will have been spared. Do you think it costs more or less than a current taxi in the US? And if it costs less than half the money per kilometer, do you think there are many who will prefer the taxi with a human driver in 3-4 years from now?
With the rapid advancements in the physical form of AI, i.e. robots, there are already automated production lines, such as Xiaomi's, that can work without lights. The machines have sensors and do not need the lighting that we need. The old joke about the factory of the future is getting dangerously close: a man and a dog. Man to feed the dog and the dog to not let man tamper with machinery.
Of the 60 million. employees of call centers and telephone operators of offices and clinics, etc. worldwide, less than 10 million. will retain their positions in the next 5 years, as AI will be able to answer calls that until now were answered by humans, with greater speed, accuracy and clarity. And, above all, at less than half the cost.
Some of this will happen in the near future, while others are already happening. The difference for the employee is that he can lose his job much faster than it will take to find out the next one. Of course, as long as he knows what it will be.
The two extremes
There are two extremes in front of us. At one extreme, more than half of the people remain unemployed and live on the verge of survival with a Guaranteed Minimum Social Income by eating legumes or souvlaki every day, with all that this entails for possible social unrest.
At the other extreme, the remaining labour is shared among more, working hours are reduced, and the standard of living goes up, as AI lowers the cost of products and services. At the same time, new specialties and jobs are being created, even in needs that we currently leave uncovered, such as companionship and care for the elderly who live alone.
Between the two there are many intermediate stops. In developed countries, automation can cover some of the worker shortages caused by low birth rates, while ageing increases care needs. But this does little to help the man who was fired today, from a profession in which he spent half his life.
On the other hand, the same AI that will reduce the need for human labour promises to speed up medical research. If the expectations of extending healthy life with the Yamanaka factors by several decades are justified, the person who will be left over in his fifties may have more than that much life ahead of him. The allowance will pay for survival. For the feeling that it is useful somewhere, however, there is no fund.
What does all this mean politically?
Will Marx become fashionable again, as in the two previous industrial revolutions in 1900 and 1968? Or will we expect AI to create the utopia of the carefree life without work, of the nobility of northern Europe at the beginning of the last century?
The question of the day is who will want to live longer, in a society that considers them unnecessary?
* Since capital.gr, collaboration with Phileleftheros Group
