To help improve programming efficiency, leaked documents show that Google has developed an internal AI model called "Goose"

GoogleCurrently accelerating the AI As the war over artificial intelligence intensifies, Google has also begun to use AI to "empower" its own development work to improve employee efficiency by integrating it into its major products.

To help improve programming efficiency, leaked documents show that Google has developed an internal AI model called "Goose"

According to Business Insider on Thursday (15th), Google has developed a robot called "goose" for its internal team.Large Language Model, exclusively for employee use.Internal documents show that the purpose of the "goose" is to assist in the development of new products.

"Goose" is described in the document as a "descendant of Gemini," a new type of large language model trained based on Google's "25 years of engineering expertise."

The internal summary also mentioned that "Goose" can answer questions related to Google's specific technology, and the code is written by the internal technology stack.And support new features such as editing code based on natural language prompts.

The document states that Goose is part of a plan to "bring AI into every stage of the product development process" and is a collaboration between Google Brain, DeepMind, and Google's internal infrastructure teams. Currently, Goose is available to at least some employees.Having a 28k context token window (referring to the amount of text, numbers, and other information a language model can consider when processing a query) is "particularly valuable" for development tasks.

statement:The content is collected from various media platforms such as public websites. If the included content infringes on your rights, please contact us by email and we will deal with it as soon as possible.
Information

SoftBank CEO Masayoshi Son plans to raise $100 billion to establish an AI chip company

2024-2-17 8:41:09

Information

Google open-sources Magika: millisecond-level content type recognition, with an accuracy rate of over 99% in a million file test

2024-2-17 8:46:48

Search