Artificial intelligence is changing the economy, society – and also the craft of programming. But how exactly and what are the consequences?
How is AI changing the way we write code for software? The main strength of AI lies in the automation of processes that follow established processes. In other words, AI can very well “predict” things that always happen the same or very similarly and implement them faster than any human – this is also the case with programming. Here, the tools are becoming very popular with young and old, with beginners as well as with long-standing experts.
An expert voice on this is that of Andrej Karpathy, former director of autopilot development at Elon Musk’s company Tesla and co-founder of OpenAI. The latter company has become known worldwide primarily for its chatbot ChatGPT.
He addressed the public with the above post on X and explains how his way of working with programming has changed thanks to AI. With the help of tools such as the AI tool Cursor (software homepage), he says that a so-called “semi-programming” has developed.
There are over 400 USB sticks sticking out of walls in Germany – What is the “Dead Drops” project?
More videos
Autoplay
More speaking than coding
Does the software engineer no longer program at all? Yes, but he becomes more of a kind of monitor of the process he initiates. In other words: According to Andrej Karpathy, he or she begins by writing a few lines to lay the foundation for the program. Then he comments on the lines and explains in a narrative way what the goal should be. From here on, to put it simply, the AI goes and completes where something is missing and adds anything else. The programmers monitor what comes out of it and usually just confirm the AI’s suggestions.
This is not only much faster, but also relieves humans of very monotonous tasks such as allocating hardware resources – so-called scheduling. OpenAI wants to outsource up to 95 percent of repetitive coding tasks to AI in the long term.
There is more time for creative approaches to program code.
However, not only the OpenAI co-founder, but also other experts in the field, as Gründer.de also discusses, are concerned about this development despite all its positive facets.
What is the danger of using AI as a programming aid? The problem lies in human nature: sooner or later we forget what we no longer do. That is why experts in the field fear that in the long term, important basic knowledge will no longer be available among programmers. Older people will lose practice and younger people may not even have learned it in enough depth to be able to practice it and be familiar with it.
Aside from its practical benefits, AI currently provides one thing above all else: great opportunities for companies and private individuals to make money. In this sense, it is exciting to look back to the beginnings of the Corona pandemic, because anyone who had the right instinct on the stock market could have won big with a company we all know: Anyone who had invested just 100 euros in Nvidia shares at the beginning of Covid-19 could buy the best graphics card for more than 2,000 euros today.