90% of our code is written by artificial intelligence

According to Dario Amodi, the CEO of Entropic, 90% of the code written in many of the company’s teams is currently generated by Claude artificial intelligence. Of course, he emphasized that this does not mean replacing engineers, but has fundamentally redefined their role.

“Why aren’t engineers fired even though 90% of the code is written by artificial intelligence?” This was an interesting question that Salesforce CEO Marc Benioff asked Dario Amudi at the Dreamforce conference. Amudi also responded that the idea that artificial intelligence will replace engineers is a “misunderstanding”.

According to him, even when artificial intelligence writes 90% of the routine and repetitive code, the same number of engineers are needed, maybe even more. The reason is that engineers now have more “leverage”. They are freed from tedious daily tasks and can focus on more important areas; These important sections are: Writing 10% of the remaining code which are the hardest and most complicated parts. Editing and optimization Codes generated by artificial intelligence and finally Supervising a group of artificial intelligence models and their guidance.

“What ends up happening is you’re 10 times more productive,” Amody says. He emphasized that this process means “redefining the role” of engineers, not “replacing” them.

Writing entropic codes by artificial intelligence

According to Insider, Entropic is not the only company that has turned to artificial intelligence to write its codes. The CEO of Y Combinator has also announced that about a quarter of the founders of startups in the company’s winter 2025 period have produced up to 95% of their code using artificial intelligence. These statistics show that Amodi’s claim is not an exaggeration and we should accept the new reality of the software development world.

Although the position of experienced engineers seems to have become more secure and even more important in this new world, a new study from Stanford University is sounding the alarm for the next generation of developers.

The rise of artificial intelligence coding tools has had a major negative impact on entry-level software engineering jobs, researchers’ study shows. According to the research, hiring of young developers between the ages of 22 and 25 has dropped by nearly 20% since ChatGPT launched in November 2022. Thus, if new engineers cannot find their first job and gain experience, we will face a shortage of experienced engineers in the future.