As Artificial Intelligence (AI) is developing rapidly, technical leaders are divided on its impact. While some software predicts a complete overhaul of development, others doubt about the promotion around it. Dario Amodi, the founder and CEO of Anthropic, recently made a bold prediction that AI will generate 90 to 100 percent of all codes by the end of 2025. This aligns with an ongoing trend where software development is rapidly becoming automatic, reducing the number of engineers required for the manufacture of applications and platforms.

However, not all technical leaders are sure that AI’s influence on coding is as revolutionary as it seems. Linux manufacturer Linus Torwalds has dismissed the AI ​​industry as “90 percent of marketing and 10 percent reality”. In an interview with TFIR, Torwald admitted that “AI is interesting” but states that he is planning to “ignore the technology” until publicity dies. He predicts that it will actually take at least five years for meaningful AI applications to emerge. Despite his doubts, he admitted that AI devices such as chat and graphic design have already proved useful.

The skepticism was also recently echoed by Narayan Murthy, the founder of Infosys, who criticized the growing AI tendency in India. Speaking at Tie Con Mumbai 2025, Murthy claimed that many so -called AI applications are only “simple programs”, which is being labeled as AI due to its popularity. He says that the fundamental principles of AI distinguish between machine learning, which identifies large -scale correlations in data, and intensive learning, which mimic the functions of the human brain. According to the idol, unwanted learning models operated by the nerve network have more ability for meaningful AI progress.

Despite these concerns, major technical companies keep integrating AI into software development. Google CEO Sundar Pichai revealed that more than 25 percent of Google’s new code has already been generated by AI, in which human engineers have reviewed and refined it.

The AI ​​writing code marks a large innings in the industry, as regular coding works are being rapidly automated, allowing engineers to focus on high-level problems and innovation. However, it also increases concerns about the future demand for entry-level programming jobs.

At the same time, Meta’s chief AI scientist, Yan Lakeun, has predicted that the current wave of generative AI and large language model (LLM) will be obsolete within five years. Speaking at the World Economic Forum in Davos, Laken argued that today’s AI lacks essential abilities like physical awareness, continuous memory, logic and complex plan. He believes that the capabilities of the current AI programs are limited and that a new AI paradigm will emerge over the next few years, possibly in the “decade of robotics”, where AI is more originally integrated with physical automation.