Demonstrating the extent to which AI can be used for the development of softwarecomes the curious revelation from Google CEO, Sundar Pichai. During the third quarter 2024 earnings call, Pichai acknowledged that “more than a quarter of the ecosystem’s new code is generated by artificial intelligence”. The use of state-of-the-art systems can effectively help speed up times and take care of a good part of the work, but it is still unable to completely replace the human personnel who must always be present in reviewing and controlling the results. “The AI does it and then it is reviewed and accepted by the engineers,” he added.
Less time, more functions
Pichai suggests that AI features in Google Search are transforming the way users interact with the platform: “In Search, our new AI features are expanding what people can search for and how they search for it. In In the case of Cloud, our generative solutions are helping to drive greater product adoption among existing customers, attract new ones and win important deals.”
Likewise, he referred to the latest data from Alphabet, Google’s parent company, which obtained revenues of $88.3 billion in the third quarter of 2024. For its part, its cloud service that provides AI infrastructure for other companies registered $11.4 billion, with a notable year-on-year growth of 35%. According to the CEO, these results are a testimony to the role of AI as an engine of financial growth for the company.
The new Google proposal is known internally as ‘Project Jarvis’ and is defined as “an agent capable of using computers.”
The AI does it and the human checks it
A recent portal survey GitHub revealed that 92% of American developers already use AI-based tools to program; the results of another survey applied by the programmers site Stack Overflowindicates that 75% of professionals already use or intend to apply AI tools in their daily work.
Pichai calls AI help a “ratio of increase for productivity and efficiency, to do more and faster.” Conceptually similar to autonomous vehicle driving, which requires a human to be present and ready to act in the event of any unforeseen event. A machine cannot do its thing, it is always necessary to check the procedure and check the results for errors. AI works very similar to that of a human, with flaws included; If today the support of highly qualified personnel is already crucial to identify problems and solve them, there will be more and more AI tools not only for humans, but by humans with AI, which will bring with it critical aspects that could increase exponentially. A detail that will require attention in an industry that is already having a profound impact on flesh-and-blood personnel.
Article originally published in WIRED Italy. Adapted by Alondra Flores.
#generated #quarter #Googles #code
