Google's Project Astra, aimed at creating advanced AI applications and agents for real-time, multimodal comprehension, won’t be released until at least next year. CEO Sundar Pichai discussed this timeline during Google’s Q3 earnings call, saying, “[Google is] developing experiences where AI can perceive and interpret the world around you.” He noted that Project Astra is just a preview of this future, with releases expected as early as 2025.
Initially unveiled at Google’s I/O developer conference in May 2024, Project Astra includes various technologies, such as smartphone applications capable of recognizing surroundings and answering context-based questions, along with AI assistants that can take actions on behalf of users. During a prerecorded I/O demonstration, a prototype from Project Astra showcased its ability to respond to questions about objects visible through a smartphone camera, like identifying a neighborhood or a part on a damaged bicycle.
Earlier this month, The Information reported Google’s plan to release a consumer-friendly AI assistant by December, equipped to handle tasks like making purchases and booking flights. However, this seems unlikely now unless it is separate from Project Astra.
Anthropic has recently pioneered a large generative AI model capable of managing applications and web browsing on a PC, but its struggles highlight the complexities of developing effective AI agents.
Post a Comment