On February 6, 2023, Sundar Pichai, the CEO of Google and Alphabet, made an exciting announcement regarding the company’s progress in the field of artificial intelligence (AI). Google introduced Bard, an experimental conversational AI service powered by their Language Model for Dialogue Applications (LaMDA). This development is a result of Google’s dedicated efforts in AI research over the past six years, which have led to significant advancements in AI computation and generative large language models.
Google, in collaboration with DeepMind, has played a crucial role in pushing the boundaries of AI technology. The scale of AI computations has been doubling every six months, surpassing the well-known Moore’s Law. Google’s Transformer research project and breakthroughs in diffusion models have served as the building blocks for many generative AI applications seen today.
Bard, the latest AI service under development, aims to combine the vast knowledge available in the world with the intelligence and creativity of large language models. By drawing on information from the web, Bard offers users fresh and high-quality responses. The service aims to simplify complex topics, inspire creativity, and nurture curiosity, catering to a wide range of interests and age groups.
The initial release of Bard will feature a lightweight version of LaMDA, requiring significantly less computing power. This approach enables a broader user base and more opportunities for feedback. Google plans to gather feedback from both external sources and internal testing to ensure that Bard’s responses meet high standards in terms of quality, safety, and accuracy of real-world information. As the testing phase progresses, Google aims to enhance the quality and speed of Bard’s responses.
Google has a longstanding history of integrating AI into its everyday products, benefiting billions of people worldwide. AI technologies like BERT, MUM, LaMDA, PaLM, Imagen, and MusicLM have greatly improved user interactions with different types of information, including text, images, video, and audio.
Furthermore, Google is committed to making it simple, safe, and scalable for developers to leverage their AI advancements by building on top of their best models. Starting next month, Google will onboard individual developers, creators, and enterprises to try their Generative Language API, initially powered by LaMDA, with plans to introduce more models in the future.
The introduction of Bard and Google’s ongoing AI research signify their dedication to advancing the field of AI, pushing the boundaries of what is possible, and bringing the benefits of AI to users and developers alike.