The near future of programming is in creating products based on artificial intelligence technologies, managing them and implementing them in other industries.
From programming to program management
There is an important difference in the process of solving problems using “classical” programming and using AI technologies. In the first case functionality and logic of a program is clear and under control of the developer, because it is fully determined by the code structure created by him. In the case of neural networks, as in the example of Corey Becker’s program, the algorithm that is used to perform the task is not available and often cannot be interpreted by a human, even after “disassembling”.
This changes the essence of the programmer’s job, as a transition is made from the producer to the controller of the AI product. Such a person is required to write less code, but to work more with:
- evaluating and analyzing the results of the product,
- managing it,
- planning its development.
Today, the business, so to speak, keeps one high-level system architect who describes the tasks and hands them over to a dozen coders to translate the tasks into the syntax of a given language. This is a paradoxical situation: with high-level languages, the more complex the tasks are, the easier it is to hire a bunch of coders in Bangalore to solve them.
“Industrialization” in programming
The arrival of artificial intelligence methods in the profession is not an apocalypse in the spirit of Terminator movies, it is about money. Progress and a growing economy require a division of labor and the emergence of narrow specialists. As the trend rises, the demand for specialists grows, which is what we are seeing today.
The more the market becomes saturated and more specialists are involved in “production”, the more the need arises to reduce costs, including by cutting personnel costs.
Programming will also undergo a similar “industrialization,” and this is normal. Artificial Intelligence will become the guy who will perform routine operations and help to test hypotheses and make decisions. The number of specialists solving problems will be reduced. Roughly speaking, instead of 10 employees, you will need two or three to:
- set tasks for artificial intelligence,
- evaluate and analyze its results,
- develop new approaches.
The ability to write code will not be a unique skill that can be used to make money. The ability to think creatively, the ability to develop new methods and algorithms will remain in value. In this sense, the programmer profession will always be in demand and there will be a necessity to hire ai developers.
Money speeds up technology
Developments in the field of artificial intelligence are currently shifting from the scientific to the business world, so the development of technology will accelerate. One of the priorities in the industry is to bring the analysis of natural language queries to a new level. Over the past 20 years, we have managed to reduce the percentage of incorrectly recognized words from 43% to 6-7%. Today we need to focus on teaching AI how to respond appropriately to queries in cases of linguistic uncertainty.
Industry giants such as Microsoft, Google, Amazon, IBM and many others are investing in projects related to AI and natural language analysis. It is predicted that by 2025, the market for products based on artificial intelligence methods will grow six-fold to approximately $36 billion. In 10 years, the cash infusion and market demands will force AI, which today writes primitive code, to “close” more complex areas of work:
– error search and correction,
– algorithm analysis and optimization.