IT companies replace programmers with AI, they run out of money

Has the dark future of science fiction books, where artificial intelligence and machines make humans unnecessary, come about?? Employers from the IT industry began to abandon the services of human programmers in favor of artificial intelligence GPT-4. They think, AI is capable of creating quality code much faster without getting paid. Thus, companies save time and money, and programmers are already at risk of losing their jobs. This is just one group of all at risk – many other professions are also under the crosshairs AI.

According to a post by Vice, one by one small startups are starting to ditch the services of real human programmers and instead use GPT-4, OpenAI’s latest neural network. This approach to writing software is beneficial for companies in almost all aspects – they save money, get quality code and high speed of writing software. This new situation is a wake-up call for many programmers, especially beginners.

By the end of March, GPT-4 does know how to write programs, but not the most complex ones yet. The advantage of this solution is that it can generate code based on several user requests expressed in a simple conversational language. At the same time, AI does not require salary, leave and contributions to the pension fund, and the employer does not pay taxes for it.

Pioneers in the exchange of people with GPT-4

One of the first to apply GPT-4 knowledge to write software code and tell the world about it is entrepreneur Joe Perkins, founder of Landscape, a startup specializing in the development of software tools for venture capital funds. He shared on Twitter that using GPT-4 he wrote five microservices for his company’s new product in just one night.

According to Perkins, an experienced programmer would devote two weeks of his time to the same job and ask for £5,000, while the GPT-4 did it in three hours. Perkins, on the other hand, only spent $0.11, without specifying what exactly the money was given for. Under these circumstances, programmers will soon start losing their jobs, or at least the good salary, and the programming profession will lose its charm.

GPT-4 not only wrote the code, but also provided step-by-step instructions for running it, Perkins said. “I didn’t understand part of the code, so I asked for comments to be added to key parts of it. Very simple,” says the entrepreneur.

“I think we’re probably still a long way from when a Google engineer will write their code with OpenAI.” But at the same time we are a startup business, the quality of our product is very important. We have limited resources. So the ability for me to just go into GPT-4 and use the code it generates saves a lot of time and other resources, including money,” Perkins says.

As a virtual programmer, the GPT-4 neural network will be incredibly useful in the field of education as it will help users write, test and explain code. People, including those with limited technical knowledge, will easily be able to learn code, says Landscape’s founder.

As an example, Perkins cites a friend of his who does not have a technical background. With the support of the new neural network, he was able to create in just 15 minutes a Chrome extension that allows users to translate selected text on a webpage into any language.

Substitute or assistant?

The examples of Joe Perkins and his friend clearly demonstrate the incredible potential of neural networks in programming. It may seem that the days of the programmer as a profession are numbered, but not everyone is sure of this.

Without denying all the advantages of GPT-4 in terms of writing code, David Joyner, executive director of online learning at the Georgia Institute of Technology, believes that neural networks will not replace live programmers. According to him, they are more likely to become their virtual colleagues and assistants and can take over some of the work. In theory, this approach will also speed up the software development process.

Programming is in danger

However, everything indicates that programming as such in the modern world is in great danger. Even if we do not take into account the classic platforms with little code and those without code, where if programming knowledge is needed to create applications, they are only the most basic, there are services that with the help of AI will write the necessary code in no time .

Microsoft, one of the biggest investors in OpenAI, the creator of GPT-4, is taking big steps in this direction. Already in the summer of 2021, the company launched a virtual assistant programmer Copilot, which is embedded in the GitHub repository and takes code from real developers.

This month saw the release of a new generation of Copilot, built with high probability on top of GPT-4, which was integrated into the Microsoft Power Platform for writing business applications. The virtual assistant understands spoken language and can, with two or three requests, generate ready-made software or a piece of code.