AI Agents — Latest News and Frameworks

SpaceX's $60 billion Cursor deal would add Anysphere's AI coding agent to xAI's developer-tool push, with key revenue, closing and regulatory caveats.

OpenAI and Visa are building Visa-backed payment infrastructure for AI agents, with tokenized credentials, user controls and rollout details still undisclosed.

Huawei Cloud has introduced Agentic Infra as a new paradigm for unified AI infrastructure, with its upcoming Ascend 950DT expected to launch in August.

Anthropic has launched Claude Fable 5, bringing Mythos-class AI to regular Claude users with safety routing, a discounted June 22 access window, and usage-credit pricing.

Google is upgrading NotebookLM with Gemini 3.5, Antigravity, source discovery and exports, but first access stays limited to AI Ultra and Workspace plans.

OpenAI's planned ChatGPT super app is expected to put agents and Codex inside one hub, turning free prompts into paid tool paths as rivals push competing agent platforms.

A June 7 meeting frames Doosan's NVIDIA-linked AI push across robotics, power systems and data center materials, with a 2028 laminate target shaping plans.

Perplexity's Search as Code lets AI agents generate Python search workflows, but claimed token savings and benchmark gains still need outside validation.

Microsoft has released VS Code 1.123 with portable AI session history, /chronicle commands, and an extension-update delay for safer daily agent workflows.

Sakana AI has opened a Recursive Self-Improvement Lab to test whether AI can cut compute dependence.

Meta's rumored $200/month Hatch AI agent would have to prove whether personal automation can justify premium pricing beyond business tools and AI power-user plans.

OpenAI has expanded its genomics and drug discovery model GPT-Rosalind with life-sciences plugins and controlled access.

Alibaba's Qwen3.7-Plus targets screen, coding, and cloud-console automation as computer-use AI rivals push beyond browsers into app and terminal tasks.

Anthropic says Claude now authors over 80% of Anthropic production code, shifting risk from writing software to reviewing AI-made changes before they ship inside live systems.

OpenAI Codex helped security researchers expose HTTP/2 Bomb, a decades old server-memory attack; nginx, Apache, and Envoy already have fixes but IIS and Pingora stay unresolved.

Meta Business Agent adds paid customer-service AI to WhatsApp, Messenger and Instagram, as Meta tests subscriptions and usage-based billing beyond ads.

Microsoft has added the Surface RTX Spark Dev Box to NVIDIA's local AI stack, linking Windows PCs, Azure and OpenShell controls for agent testing ahead of rollout.

Tencent is reportedly developing a WeChat AI agent that would use mini programs to complete in-app tasks, with review and external tests still ahead.

GitLab has announced plans to cut 14% of staff and exit 22 countries, shifting savings to AI-agent infrastructure while reliability risks grow for customers.

Microsoft's new Agent Control Specification gives AI agent teams runtime policy checks across tools, model calls, and approvals, adding auditable controls without replacing runtimes.

Google has released Gemma 4 12B, a local multimodal AI model for laptops that tests whether audio, images, code, and tool calls fit in 16GB memory locally.

Perplexity's local-cloud AI router decides when work stays on a PC or moves to cloud models, making task classification the launch's privacy test.

Researchers built a contained AI powered malware worm that adapts attacks across lab hosts, exposing how local open-weight models complicate malware containment.

Microsoft´s new Intelligent Terminal adds a docked AI agent pane for failed shell commands, giving developers a separate Windows Terminal fork to test fixes.

Microsoft has released ASSERT, an open-source framework that turns AI-agent policies into executable tests so developers can catch failures before deployment.

Microsoft Scout brings an OpenClaw powered AI coworker into Teams, with permissions, audit logs, and a human approvals central.

Microsoft is developing Project Solara as an Android-based platform for devices organized around AI agents rather than traditional apps.

Nvidia says AI demand will keep supply constrained through 2027 even as its revenue outlook keeps climbing.

Anthropic has disclosed a 31.5% prompt-injection success rate for Claude's browser agent before safeguards, showing how hostile web instructions can reach live tools.

Mistral has rebranded its Le Chat AI assistant as Vibe, folding work automation and remote coding into one AI agent with cloud sandboxes, connectors and tiered pricing.