Google’s Secret Weapon: “Goose” AI to Boost Coding

Business Insider reports that Google has internally launched a substantial natural language model named “Goose,” aimed at assisting its employees in swiftly generating programming content and accelerating the development and efficiency of new products.

Inside Google, a large-scale natural language model known as “Goose” is facilitating the rapid development of internal products and enhancing work efficiency, crafted with the engineering expertise the Google team has amassed over the past 25 years.

Internal documents reveal that “Goose” was developed through a collaboration between Google Brain, DeepMind, and Google’s internal teams before the merger into Google DeepMind. Currently, it is primarily used within Google.

“Goose” draws on the engineering prowess the Google team has developed over the past 25 years. It can answer Google-related technical questions and interact through natural language, thus aiding employees in quicker programming and coding, thereby hastening the launch of new Google products and services.

However, Google has not issued any response regarding this matter.

Google has already made available to the consumer market and developer community a large-scale natural language model named “Gemini.” Recently, it announced the more extensive Gemini 1.0 Ultra model and introduced the new Gemini 1.5 Pro model, which boasts computational results nearly identical to those of the Gemini 1.0 Ultra but requires less computing power and offers higher efficiency. Moreover, the Gemini 1.5 Pro supports text, images, videos, audio, and code generation across multiple modes, enabling natural language interactions within a single dialogue box.