Artificial intelligence models have many uses. But one area where these models excel is computer coding. And at Google, engineers already rely heavily on AI to code. Indeed, during the presentation of third quarter results, Google boss Sundar Pichai revealed that 25% of the company’s code is already generated using AI.
“We also use AI internally to improve our coding processes, which drives productivity and efficiency. Today, more than a quarter of Google’s new code is generated by AI, then reviewed and accepted by engineers. This allows our engineers to do more and move forward faster”he indicated.
Developers have already widely adopted generative AI
In any case, Google engineers are not the only ones to rely on generative artificial intelligence to code. In a study by the Stackoverflow forum (published in 2024), 76% of respondents indicated that they already use or intend to use AI to code. And 72% of developers surveyed indicated that they are very supportive of the use of AI. In 2023, the GitHub platform surveyed American developers: 92% already use AI-based tools.
Gemini lands on GitHub
In any case, Sundar Pichai’s revelation is good publicity for Google’s Gemini AI, which is arriving on the GitHub coding platform. As a reminder, this platform belongs to Microsoft. And it already offered an assistant for developers called Copilot which is based on OpenAI technologies. But recently, GitHub announced the possibility for developers to use other models, such as Claude 3.5 Sonnet or Gemini 1.5 Pro in place of OpenAI’s AI models.
For what ? “Over the past year, we have seen a boom in high-quality language models, large and small, that individually excel at different programming tasks. It is clear that the next phase of AI code generation will not only be defined by multi-model functionality, but also by multi-model choice,” explained the Microsoft platform. In other words, instead of restricting GitHub users to OpenAI technologies, the platform wants to allow developers to take advantage of the advantages of other models, for specific tasks.
AI boosts Google Cloud revenue
Otherwise, it should be noted that Google’s investments in the field of generative AI are already bearing fruit. Indeed, if advertising remains its main source of revenue, the cloud branch experienced exceptional growth in the third quarter, thanks to new products intended for companies wishing to use generative AI. For example, Google Cloud services partly power Snapchat’s My AI assistant. Google’s cloud and AI models are also used by Volkswagen America to provide a virtual assistant to its customers.
- During Google’s quarterly results presentation, Sundar Pichai revealed that 25% of Google’s code is already generated by AI
- Code generated with artificial intelligence must nevertheless be validated by human developers
- Studies show that generative AI-based tools are already widely used by developers
- And soon, developers who use GitHub will be able to use Gemini thanks to an integration between the two products
