Amazon’s AGI Lab Reveals Its First Work: Advanced AI Agents

Amazon is still seen as a little laggard in the race to develop advanced artificial intelligenceHowever, when it silently comes to the AI ​​performance, he created a lab that put notes. Amazo’s Agi SF LaboratoryLocated in San Francisco and dedicated to the building Artificial general explorationAI or AI, which surpasses people’s capabilities, revealed the first fruits of today: a new AI model that is able to strengthen some of the most advanced AI agents available everywhere.

Amazon says that AI is a new model, called Amazon Nova Acti, Openai and anthropic, Openai and anthropical, Openai and anthropic, AI, AI is better performed in the surface website and screenshots of Amazon Nova ACT, KLOD 3.7 SONET and Openai computer use agent. A large part of the Amazon’s plan to compete in the AI ​​market is to focus on construction agents and reflects the efforts to create a means that can measure the best of the new model.

“I believe that the main atomic unit in the future intends to call a giant [AI] Agent, “Amazon’s AGI SF Laboratory was a starting engineering vice-president when he was previously launched on Openai and later AI agents, E-commerce giant.

Most of the leading AI laboratories are already in the spotlight Growing AI agents increasingly. Promises to master the independent actions of the EU, as well as to conversate and make technology more useful and valuable. It is still an ongoing work to move the conversation.

In the last six months, Openai, anthropic, Google and others demonstrated Web agents traveling A quarter is acting in response. However, in most cases, these agents are still invalid and can easily be easily attached by open surveys.

Luan says Amazon’s goal is to build AI agents, which are dependentless, not an endless. What agents returned, “The problem of interesting opportunities working for 60 percent of the time,” says the problem of the road, “he says that it needs to be used to engage in unusual external work with the help of self-driving cars.

Many so-called agents are designed to prevent a large language models themselves, but also fragile their behavior. Amazon Nova ACT is a version of the company’s most powerful homegrown model Amazon Nova This has received additional training to decide what actions and when it will be. In general, Luan says that the AI ​​models are struggling to decide to intervene in a position.

Nova’s agident uses Amazon to improve their abilities Learning reinforcementa method that helps others AI models follow the justification better.