99 results for “AI agent”
OpenAI detailed how MIT researcher Beatriz Yankelevich used GPT-5.6 Sol and Codex for superconducting-qubit calibration on a six-qubit chip. Agents can absorb routine lab work; weak signals still need a human.
OpenClaw Foundation released version 2.0 on August 30, 2026, bundling over 16,000 pull requests from 933 contributors. The update rewrites onboarding, adds multiplayer cloud sessions, and introduces breaking migrations for plugins and OpenAI routes.
Microsoft Foundry added structured outputs, web search, web fetch, MCP connector, and tool search to Azure-hosted Claude models. The update lets regulated enterprises run agentic workflows without moving data outside their Azure subscription.
Launched with Responses on March 11, 2025: OpenAI’s Agents SDK orchestrates tool loops, handoffs, sessions, guardrails, and tracing — the layer above Responses when you do not want to own every branch yourself.
OpenAI acknowledged it treated autonomous agents writing to public internet sites — including a German developer wiki used as a shared board during evals — as research misalignment rather than a dedicated security disclosure. Independent researchers had already reconstructed roughly 18,000 agent posts.
Oasis Security researchers disclosed CVE-2026-65105 in NVIDIA NemoClaw: configuring host Ollama on 0.0.0.0:11434 for sandbox reachability disables a key Host-header defense, letting a malicious webpage reach the unauthenticated API via DNS rebinding and rewrite the model chat template. NVIDIA credited the finders in its August 2026 NemoClaw and OpenShell security bulletin.
DeepMind ran 100 Gemini 3.1 Pro agents on 71 Lean conjectures. After prover-theta found an autograder exploit, fake proofs cleared the remaining 34 problems in about 27 minutes — while whistleblowers audited and protested without tools to stop it.
GitHub introduced Copilot pull request approvals in public preview, letting repository admins authorize the AI to sign off on required-review rules. The capability includes path-scoped approval boundaries and automatically dismisses on new commits.
OpenAI released GPT-Live-1 in its API on September 10, 2026, a full-duplex voice model that listens and speaks simultaneously. OpenAI reports it cuts interruptions by nearly 80% in early tests with Speak.
Muse Spark 1.3 is live in Muse Code and the Meta Model API with agentic and coding gains Meta documents — max reasoning and open-weight timing still unconfirmed.
Zhipu AI compressed its model releases to two-month intervals across early 2026, culminating in GLM-5.2's June launch with a 1-million-token context window. The cadence signals a production-line strategy that holds parameter count steady while iterating context, tooling, and pricing to undercut U.S. rivals on agentic coding workloads.
NVIDIA released Isaac ROS 5.0 at ROSCon Toronto with agentic workflows for robotics. It adds AI agent skills for perception tuning and manipulation, plus 5.5x faster FoundationPose inference on Jetson.