Try to imagine a world where everyday computer tasks are solved by themselves, only with the help of artificial intelligence. It might sound like science fiction, but according to Anthropic, the company behind the Claude artificial intelligence, that dream might not be so far off. It recently introduced a new function called “Computer use”.
As it stands in official statementartificial intelligence can use a computer much like a human. He can move the cursor, click on places on the screen, write text using the virtual keyboard or work with the software according to the given instructions. If an error occurs, Claude should be able to recognize it and try to correct it. Automation using artificial intelligence could theoretically make simple and routine computer activities easier.
Example of the Computer use function in practice:
Source: YouTube.com
However, the developers themselves admit that the function is at the very beginning of development and is very buggy. In internal testing, for example, Claude stopped herself in the middle of assigning a coding task and started opening photos of Yellowstone National Park. Although this one is I guess smiling, it’s still a reminder that even the best AI models aren’t flawless.
The “Computer use” feature is already in public beta testing. So, if you are among the technology enthusiasts, you can try this feature yourself. Apart from that, there was also the introduction of an improved version of the Claude 3.5 Sonet, bringing across-the-board improvements across multimodal tasks.
Source: www.cnews.cz
