Anthropic Unleashes AI That Controls Your Computer
The world of artificial intelligence is buzzing with rapid innovation, and this past week has seen a flurry of exciting developments. Leading the charge is Anthropic, a company that has been consistently releasing new features for its AI assistant, Claude. In a recent five-week period, Anthropic rolled out an impressive 74 updates, showcasing a relentless pace of improvement.
Claude Gains Computer Control and Project Management
Among the most impactful recent releases is a feature allowing Claude to directly interact with your computer. Enabled through Claude’s desktop app on paid plans (starting at $20 per month), this ‘computer use’ capability lets Claude control your mouse and keyboard. Imagine being away from your desk and texting your computer to perform a task; Claude can now execute those instructions by clicking and navigating your system. While this feature is currently slow – taking several minutes for simple actions like opening an application – its potential is significant, especially when combined with Claude’s ‘dispatch’ feature for remote operation.
Claude has also enhanced its organizational tools with a new ‘projects’ feature within Claude Co-work. Users can now create distinct projects, assign them custom instructions, and even attach files for added context. This allows Claude to adopt specific personas or follow detailed guidelines for each project. For example, a user could instruct Claude to ‘always respond like a gym bro and call me dude and bro constantly,’ demonstrating the system’s ability to adapt its communication style.
For coders, Anthropic introduced ‘auto mode’ for Claude Code. This update streamlines the coding process by allowing Claude to perform more actions, like web searches or running certain terminal commands, without requiring explicit permission for each step. This eliminates the frustrating back-and-forth of repeated permission requests, letting developers focus on their work.
Google Enhances Gemini and Music Generation
Google has also been busy, launching updates to its Gemini AI. The new Gemini 3.1 Flash Live model is now integrated across various platforms, including its API, enterprise services, search, and the Gemini Live application. This multimodal AI can process webcam feeds and screen shares, enabling interactive conversations where Gemini can guide users through complex software like OBS Studio.
Google is also improving its music generation capabilities with Lia 3 Pro. Previously limited to 30-second tracks, this advanced version can now generate songs up to three minutes long. Users can prompt for specific song structures like intros, verses, and choruses. This technology is being rolled out across Google’s AI products, including Vertex AI and Google AI Studio.
In a move to attract users from competing platforms, Google now allows users to migrate their chat history, memories, and preferences from other AI applications, such as ChatGPT and Claude, into Gemini. This makes switching to Google’s AI offerings more seamless.
New AI Tools Emerge for Creativity and Voice
Beyond the major tech players, new AI tools are making waves. GenSpark offers an all-in-one AI workspace that integrates various models for tasks like creating presentations, analyzing data, and writing reports. Notably, their paid plan, priced at $20 per month, offers unlimited AI chat and image generation until the end of 2026, a significant value compared to paying for individual model access.
For those interested in AI-generated audio, Smallest.ai has released a new text-to-speech model designed for conversational AI. It aims to sound more natural, mimicking human thought and listening processes, and can be trained on a user’s own voice for a personalized touch. Mistral AI has also introduced an open-weights text-to-speech model that can be run locally on a user’s computer, offering competitive quality with options like 11 Labs and the ability to adapt to custom voices with just a few seconds of reference audio.
Love Art AI has launched a feature called ‘Move Object,’ which allows users to select a part of an image and reposition it, creating new compositions. This can then be used to generate short animations by creating start and end frames for video generators.
OpenAI Shifts Focus, Cuts Projects
In a surprising move, OpenAI has announced it is discontinuing its Sora video generation model, along with its API and app. The company is refocusing its efforts on its core chat and coding models, citing the significant computational resources required for video generation. This decision led to the dissolution of a partnership with Disney, which had been slated to access OpenAI’s technology.
Additionally, OpenAI has reportedly shelved plans for an ‘adult mode’ for ChatGPT. These strategic shifts suggest a move away from experimental or niche features towards core AI functionalities. The company is also facing challenges with its advertising platform within ChatGPT, as early advertisers report a lack of clear data on ad performance and measurable business outcomes. Despite these hurdles, OpenAI is enhancing its shopping experience within ChatGPT, making it easier for users to discover and compare products, potentially paving the way for new e-commerce monetization strategies.
Why This Matters
The rapid pace of AI development, as seen with Anthropic’s continuous updates and Google’s Gemini enhancements, signals a maturing AI landscape. Features like Claude’s computer control, though currently slow, point towards a future where AI assistants can perform complex tasks autonomously, acting as true digital helpers. The improvements in AI music and voice generation democratize creative tools, allowing individuals to produce professional-quality content more easily. While OpenAI’s strategic pivot away from video generation highlights the economic realities of AI development, its continued focus on core chat and coding models, alongside new shopping integrations, suggests an ongoing effort to find viable business models. The emergence of new competitors and open-weight models also fosters greater accessibility and innovation within the AI community.
Source: AI News: Anthropic Went Crazy This Week! (YouTube)
