Google expands its Gemini API Managed Agents with powerful new features including background execution for long-running tasks, remote MCP server support, custom function calling, and mid-session credential refresh that preserves sandbox state — making AI agents significantly more reliable and flexible for production use.
A new study reveals that using JSON mode for AI agents causes smaller models to fail up to 60% of the time and costs up to 11x more than traditional CLI arguments due to shell escaping failures, with developers urged to avoid rewriting CLIs for agents until running measurable experiments.
Shepherd launches an alpha runtime that makes AI agent executions reversible using Git-like forking, delivering 5x faster copy-on-write performance than Docker commit, 95% KV-cache reuse on replay, and OS-level sandboxed permissions that block unauthorized writes before they happen.
Guidance, an open-source Python library with over 21,600 GitHub stars, gives developers precise control over AI model outputs using regex, context-free grammars, and JSON schema validation, while boosting efficiency by reducing unnecessary model processing across backends like OpenAI and llama.cpp.
European AI startup UMA unveils a humanoid robot powered by Real-Time Learning architecture that watches and learns from human demonstrations, positioning Europe as a serious contender in the global robotics race against the US and China.
AI law startup Norm reaches unicorn status after raising $120 million in a Series C round led by Khosla Ventures, pushing its valuation to $1.2 billion as it disrupts traditional legal billing with AI agents supervised by human attorneys serving enterprise clients.
A new DigiCert survey reveals nearly 80% of enterprises have experienced AI-related security incidents in the past six months, as rapid deployment of AI systems dramatically outpaces governance, leaving companies unable to track what AI is running or who is accountable when failures occur.
Microsoft is ditching OpenAI and Anthropic AI models in favor of cheaper, in-house MAI models across products like Excel and Outlook, while simultaneously cutting 4,800 jobs and doubling down on AI expansion — a bold cost-cutting gamble that risks alienating enterprises demanding top-tier AI performance.
DoorDash launches DashBench, a rigorous AI code review benchmark that replays historical pull requests to surface real developer findings, revealing that multi-stage architectures dramatically outperform single-pass models and that trustworthy AI evaluation demands triangulating multiple imperfect signals rather than relying on any one metric.
Microsoft is ditching OpenAI and Anthropic AI models in favor of its own proprietary MAI models inside Excel and Outlook, completing tens of thousands of prompts weekly as the tech giant slashes costs and bets big on in-house artificial intelligence.