Amazon has announced the establishment of a new R&D lab in San Francisco, dubbed the Amazon AGI SF Lab, which will focus on building "foundational" capabilities for AI agents. The lab, led by David Luan, co-founder of AI startup Adept, aims to create agents that can "take actions in the digital and physical worlds" and "handle complex workflows" using computers, web browsers, and code interpreters.

The Amazon AGI SF Lab will build on the work of Amazon's broader AGI team, with Pieter Abbeel, a robotics research lead who joined Amazon through the acquisition of AI robotics startup Covariant, working closely with Luan and the lab. According to a joint post by Luan and Abbeel, the lab's initial focus will be on several key research bets that will enable AI agents to perform real-world actions, learn from human feedback, self-course-correct, and infer goals.

The lab will be seeded by Adept employees, and Amazon is looking to hire a few dozen additional researchers in fields such as physics, math, and quantitative finance. This move comes after Adept agreed to license its tech to Amazon in June, with Luan and other co-founders joining the e-commerce giant. Luan will continue to work under Rohit Prasad, the former Alexa head who's leading an AGI team focused on building large language models.

The quasi-acquisition of Adept has drawn regulatory scrutiny, with policymakers seeking to determine whether tech giants like Amazon and Microsoft are quashing AI competition. Microsoft struck a similar deal with AI company Inflection in May. Adept, founded two years ago, aimed to create an AI model that can perform actions on any software tool using natural language, essentially creating an "AI teammate" trained to use a wide variety of different software tools and APIs.

The "agentic" AI sector is expected to be worth $31 billion by the end of the year, according to Emergen Research. Several startups, including Orby, Emergence, and Rabbit, are working on agent products to complete tasks largely autonomously. OpenAI and other major AI players are also developing agent products, with OpenAI rival Anthropic releasing its take on the tech earlier this year. Google is reportedly working on AI agents that can make purchases, such as booking flights and hotels.

Amazon has dabbled in the agent space, but has yet to make a serious play. In July, the company announced conversational agents for its Bedrock AI development platform, and just last week, it brought agents to its Amazon Q Business assistant platform for business customers and devs. Amazon CEO Andy Jassy has hinted at a more agentic Alexa, capable of not only responding to questions but taking actions. The establishment of the AGI SF Lab marks a significant step towards realizing this vision.

The implications of Amazon's move into the agentic AI space are far-reaching. As the company continues to develop its capabilities, it could potentially disrupt industries such as customer service, software development, and e-commerce. With its vast resources and expertise, Amazon is well-positioned to make a significant impact in the AI agent space. As the sector continues to grow, it will be interesting to see how Amazon's efforts shape the future of AI.

In conclusion, Amazon's establishment of the AGI SF Lab marks a significant milestone in the development of agentic AI. As the company continues to push the boundaries of what is possible with AI agents, it could have a profound impact on various industries and the tech landscape as a whole.