Anthropic is expanding its Claude models with capabilities to perform tasks independently on a PC. The feature is being tested by developers.
Do you get the hips from filling out digital forms? Wouldn’t it be easier if AI could do it for you? It’s not a distant future. Anthropic announces new enhancements for Claude 3.5 Sonnet, one of the latest versions of the company’s LLM. Claude 3.5 Sonnet should now be able to work independently with a PC.
In a YouTube video, Anthropic demonstrates what is possible. We see Claude in the video processing a fill-in form completely independently. The AI model searches for the requested data itself in various applications. It takes screenshots of your screen to search for information and enters it in the appropriate field.
Cumbersome and error-prone
Anthropic is not putting this technology in the hands of a wide audience for now, but is having it tested by a limited circle of developers. The company itself indicates that Claude’s computing skills are still highly experimental and “at times cumbersome and error-prone.
For example, there are still many actions you routinely do on a PC that Anthropic’s model is not yet capable of, such as dragging and zooming. Anthropic has also deliberately introduced some limitations. Claude does not get access to your social media accounts to avoid AI writing messages on your behalf, and government agency websites are also marked as off limits.