Artificial intelligence capable of controlling a computer independently

Technology news continues to evolve at an amazing pace, and one of the most notable recent developments is Anthropic introduces new features for the Claude 3.5 language model. This model, in its version called “Sonnet,” now allows artificial intelligence to control a computer autonomously.

Claude 3.5 New Sonata capabilities in detail

Claude 3.5 Sonnet update introduces a revolutionary feature: The ability of artificial intelligence to use a computer like a human. This includes actions such as moving the mouse, clicking buttons, and typing text. This function can be accessed via Anthropic's API, Amazon Bedrock, and Google Cloud's Vertex AI.

According to Anthropic, rather than designing specific tools to accomplish individual tasks, the approach is to: Teaching general computer skills to ClaudeThis makes this AI compatible with a wide range of standard tools and software. Many partners like Asana, Canva, and The Browser Company have already started exploiting this model. These first uses show great promise for CLOUD's capabilities to automate common computer tasks.

Potential impact on developers and users

For developers, The introduction of the Claude 3.5 Sonnet could mean increased productivity and simplification of complex processes. Artificial intelligence is capable of performing tasks that require multiple steps while evaluating and adjusting its actions in real time. For example, if Claude encounters an obstacle, he can correct himself and try again until he succeeds, thanks to what Anthropics calls the “moving book nature” of screen vision. This means that Claude takes successive screenshots to reconstruct the continuous visual flow.

In terms of performance, OSWorld gave Cloud 3.5 a score of 14.9% for tasks that only involve screenshots, and 22% for tasks that require multiple steps. Although this ability currently exists Public betaIt already shows great potential for various future applications.

Safety and reliability of the Claude 3.5 Sonata

The critical aspect of this new technology is security. Anthropic specifies that while training the model, Claude did not have access to the Internet, reducing the risk of sensitive data being compromised. In addition, initial tests showed that the model can handle simple errors without human intervention, highlighting its strength in the face of unexpected scenarios.

Additionally, organizations like GitLab have seen clear benefits from using Claude for DevSecOps tasks. The model allows for better logical reasoning without adding additional latency, making complex projects easier to manage.

Claude 3.5 Haiku: Another innovation from Anthropy

At the same time, Anthropic also unveiled the Claude 3.5 Haiku, a faster and more compact iteration of the Claude 3. While the Claude 3.5 Haiku focuses more on programming efficiency and careful following of instructions, it also offers less reaction time. According to Anthropic, Claude 3.5 Haiku is designed for user-facing products, specialized sub-agent tasks, and generating personalized experiences from large amounts of data such as purchase history or inventory records.

This new version, which combines the use of Anthropic's API, Amazon Bedrock infrastructure, and Google Cloud's Vertex AI, ensures that these models remain accessible and integratable across professional contexts.

Since its launch, Claude 3.5 Sonnet has received enthusiastic feedback from its first-time users. Companies that have tested this AI have highlighted its effectiveness in automating workflows. For example, DoorDash uses Claude 3.5 Sonnet to automate some internal processes, indicating a noticeable improvement in productivity.

With this major development, Claude's newly acquired skill 3.5 Sonnet has the potential to redefine the current standards of human-computer interaction.. Although currently intended for general experimental use, this innovation opens the way to countless potential applications for both individuals and businesses.