nthropic made courageous statements last year in its pitch to investors. The AI company said it intended to create a virtual assistant that could conduct
research, respond to calls, and even handle other back-office tasks. Even more so, at the investors' pitch, Anthropic called this artificial intelligence theory was working on a “next-gen algorithm for AI self-teaching”.
After research and intensive work towards its development, Anthropic’s artificial intelligence is slowly starting to surface in the media. Anthropic’s AI is believed to have someday the capability to automate parts of the economy.
This Tuesday, Anthropic made public the news upgrade from its older versions of the Claude 3.5 Sonnet model that has the ability to interact with any desktop app. With the newest “Computer Use” API that is now available in open beta, the latest update can imitate keystrokes, mouse gestures, and clicks, reproducing a real person responding and delivering tasks on a PC.
Subscribe to our newsletter
Anthropic wrote in a blog post “We trained Claude to see what’s happening on a screen and then use the software tools available to carry out tasks,” and that “When a developer tasks Claude with using a piece of computer software and gives it the necessary access, Claude looks at screenshots of what’s visible to the user, then counts how many pixels vertically or horizontally it needs to move a cursor in order to click in the correct place.”.
Even more so, for developers, the Computer Use can be tried out via Anthropic’s API, Google Cloud’s Vertex AI platform, or Amazon Bedrock. The news model of the AI software, 3.5 sonnet will also be released for Claude apps, yet it will not support Computer Use.
Since the emergence of technology, many companies have tried over and over again to create an AI that would be able to automate and deliver tasks from a PC. Automat, Introduced AI or RPA vendors have tried to create the so-longed-for “AI agent” to help us be more productive.
In this competition, the market only became oversaturated, and the automation of software is still a field that needs to be researched and developed. In an interview for TechCrunch, Anthropic said “Humans remain in control by providing specific prompts that direct Claude’s actions, like ‘use data from my computer and online to fill out this form” and “People enable access and limit access as needed. Claude breaks down the user’s prompts into computer commands (e.g. moving the cursor, clicking, typing) to accomplish that specific task.”.
A more budget-friendly alternative will also be updated. Anthropic said that they will also update Haiku, the cheapest model Anthropic has available. “With low latency, improved instruction following, and more accurate tool use, Claude 3.5 Haiku is well suited for user-facing products, specialized sub-agent tasks, and generating personalized experiences from huge volumes of data–like purchase history, pricing, or inventory data,” said Anthropic about the model.
Yet, the model still needs further development and research so start with something small. “Claude’s Computer Use remains slow and often error-prone,” wrote Anthropic, and “We encourage developers to begin exploration with low-risk tasks.”.