Anthropic’s new AI tool acts on your behalf
Google is reportedly developing a ‘computer-using agent’ AI
Anthropic’s new AI tool acts on your behalf
Google is reportedly developing a ‘computer-using agent’ AI
Microsoft bets on Copilot Agents
The fastest way to build AI apps
Writer Framework: build Python apps with drag-and-drop UI
API and SDKs to integrate into your codebase
Intuitive no-code tools for business users
***
Anthropic’s new AI tool acts on your behalf
Anthropic, an artificial intelligence startup, is releasing a new tool that can understand and interpret what's happening on a user's computer screen. This tool, called "computer use," can complete a range of online tasks for the user, such as browsing the web, clicking buttons, and typing, with the user's permission.
The company is releasing a beta version of this tool to developers using its Claude technology. This technology has been tested with a limited set of enterprise customers in recent weeks.
Anthropic's approach to agent tools is different from other companies. Instead of integrating with various applications on the backend, its technology can process what's happening on a user's computer screen in real time. This method is said to create a more intuitive experience. According to Jared Kaplan, co-founder and chief science officer at Anthropic, this technology will be the first model to be able to use a computer the way people do However, the technology faces significant limitations and safety considerations. The company said the system struggles with everyday computer actions like scrolling, dragging and zooming. Getting AI agents to work well will take a while. Meanwhile GenAI is ready to go using Writer….
***
Google is reportedly developing a ‘computer-using agent’ AI
Google is reportedly working on a new AI-powered tool, codenamed "Project Jarvis," that can carry out tasks for users, including gathering research, making purchases, and booking flights. According to sources familiar with the project, Jarvis is designed to automate everyday web-based tasks by taking and interpreting screenshots, and then performing actions such as clicking buttons or entering text.
The tool is said to be powered by a future version of Google's Gemini AI model and is specifically tuned to work with the Chrome web browser. While details are still scarce, the project is part of a larger trend among major tech companies to develop AI models that can interact with and perform tasks on behalf of users.
Microsoft, Apple, Anthropic, and OpenAI are all working on similar projects, with Microsoft's Copilot Vision allowing users to interact with webpages, Apple Intelligence expected to enable cross-app functionality, and Anthropic and OpenAI developing tools that can use a computer on behalf of the user.
However, it's worth noting that Google's plans to unveil Jarvis in December are subject to change….
***
Microsoft bets on Copilot Agents
In a significant expansion of its artificial intelligence capabilities, Microsoft is launching autonomous agents for its Copilot AI assistant software. These agents, which range from simple prompt-and-response to fully autonomous, can perform a variety of tasks, including sending emails, handling employee onboarding, and working independently as part of a team.
According to Microsoft, the agents are designed to function as "the new apps for an AI-powered world," and are trained using large language models (LLMs) and workplace data specific to each client. This allows the agents to learn and adapt to the unique needs and workflows of each organization.
To get customers started, Microsoft is also releasing 10 pre-designed agents, each tailored to a specific business function. These include agents for prioritizing high-potential sales leads, communicating with suppliers, and managing customer service interactions. The move is seen as a major step forward in the development of AI-powered automation, and is expected to have a significant impact on the way businesses operate….