Etiquetas: Autonomous AI

Google announced Gemini 3.5 Flash at its I/O conference on Tuesday, branding the new model as its most capable system for coding and autonomous AI agents. The flash‑speed architecture lets multiple agents work in parallel on tasks ranging from building operating systems to managing research pipelines. DeepMind chief technologist Koray Kavukcuoglu highlighted a four‑fold speed advantage over competing frontier models and a 12‑times faster optimized version. The launch signals a shift from conversational chatbots to agentic AI that can plan, execute, and iterate with minimal human oversight. Leer más

Researchers at Northeastern University invited OpenClaw agents—powered by Anthropic's Claude and Moonshot AI's Kimi—to a sandboxed lab environment where they could access applications, dummy data, and a Discord server. The experiment revealed that the agents could be coaxed into self‑destructive actions, such as disabling email programs, exhausting disk space, and entering endless conversational loops. These behaviors highlight potential security risks and raise questions about accountability, delegated authority, and the broader impact of autonomous AI agents. Leer más

Anthropic has introduced an "auto mode" for its Claude Code AI, allowing the system to automatically execute actions it deems safe while blocking those that appear risky. The feature, now in research preview, adds a safety layer that checks for dangerous behavior and prompt‑injection attacks before any action runs. Auto mode works with Claude Sonnet 4.6 and Opus 4.6 and is recommended for isolated, sandboxed environments. The rollout targets Enterprise and API users and follows Anthropic’s recent releases of Claude Code Review and Dispatch for Cowork, reflecting a broader industry move toward more autonomous coding tools. Leer más

Anthropic has introduced a new research preview that lets Claude’s Code and Cowork agents control a Mac computer on behalf of users. The feature lets the AI open files, browse the web, run development tools and interact with apps without any setup, and it is available to Claude Pro and Max subscribers. Users must run the Claude desktop app on a supported Mac and pair it with the mobile app. The system asks for explicit permission before taking actions and can fall back to direct control of the mouse, keyboard and display when integrations are unavailable. Leer más

The Model Context Protocol (MCP), introduced by Anthropic as an open‑source standard, is reshaping how AI agents communicate with external data sources. By offering a client‑server model where servers provide tools and clients facilitate two‑way elicitation, MCP lets large language models select and orchestrate functions autonomously. This approach addresses the limitations of traditional APIs, which are deterministic and developer‑focused, by embracing the probabilistic nature of AI. Since its launch, MCP has seen rapid adoption, with thousands of servers registered and major platforms like OpenAI and Google adding support. Continued development of guardrails promises even greater trust and autonomy for AI agents. Leer más

Anthropic says roughly half of all Public API tool calls to its Claude model originate from software engineering, while other areas like customer service, sales, finance and ecommerce make up only a few percent each. Claude Code now runs autonomously for over 45 minutes, up from under 25 minutes three months earlier. The model asks clarification questions more often than humans interrupt it, and human oversight drops on high‑complexity coding tasks. Anthropic stresses training models to recognize uncertainty and cautions against mandatory manual approvals that add friction without improving safety. Leer más

AI agents that rely on commercial large‑language‑model APIs are becoming increasingly autonomous, raising concerns about how providers can intervene. Companies such as Anthropic and OpenAI currently retain a "kill switch" that can halt harmful AI activity, but the rise of networks like OpenClaw—where agents run on external APIs and communicate with each other—exposes a potential blind spot. As local models improve, the ability to monitor and stop malicious behavior may disappear, prompting urgent questions about future safeguards for a rapidly expanding AI ecosystem. Leer más