Jack Wotherspoon, a developer experience engineer at Google DeepMind, presented Gemini Managed Agents: coding agents that run in cloud sandboxes launched from a single API call rather than locally. The talk walks through the Interactions API, the Antigravity harness, shared environments, skills, credential handling, scheduled triggers, and multi-agent workflows, with live demos of an SEC-filing researcher handing off to a quant agent and a stock research app built in AI Studio.
In a tutorial video, Philipp Schmid of Google DeepMind builds an autonomous AI finance coworker using Gemini and the Gemini API's Managed Agents. He frames the coworker as seven needs — brain, computer, playbooks, tools, credentials, house rules, and schedule — and demonstrates each. The pattern matters to practitioners shipping autonomous agents that handle sensitive data.
In a tutorial video, Philipp Schmid of Google DeepMind builds an autonomous AI finance coworker using Gemini and the Gemini API's Managed Agents. He frames the coworker as seven needs — brain, computer, playbooks, tools, credentials, house rules, and schedule — and demonstrates each. The pattern matters to practitioners shipping autonomous agents that handle sensitive data.
A demo where Google DeepMind's Gemini 4 Argon agent, running in Antigravity, generates a complete 3D Gothic cathedral scene from a single Nano Banana image. Using Blender MCP, the model iteratively refines geometry, textures, and lighting. It demonstrates agentic AI automating 3D scene creation for artists and technical users.
A demo where Google DeepMind's Gemini 4 Argon agent, running in Antigravity, generates a complete 3D Gothic cathedral scene from a single Nano Banana image. Using Blender MCP, the model iteratively refines geometry, textures, and lighting. It demonstrates agentic AI automating 3D scene creation for artists and technical users.
A workshop video from AI Engineer World's Fair 2026 shows Priyanka Phatak, who leads the Claude Managed Agents engineering team at Anthropic, walking through the platform's architecture redesign. The first version ran session, harness, and sandbox in one hard-to-debug 'pet' container; the redesign decouples a brain (harness plus Claude) from hands (sandboxes and tools) around an append-only session log. Practitioners get a documented pattern for keeping harness assumptions from going stale with each new model.
A workshop video from AI Engineer World's Fair 2026 shows Priyanka Phatak, who leads the Claude Managed Agents engineering team at Anthropic, walking through the platform's architecture redesign. The first version ran session, harness, and sandbox in one hard-to-debug 'pet' container; the redesign decouples a brain (harness plus Claude) from hands (sandboxes and tools) around an append-only session log. Practitioners get a documented pattern for keeping harness assumptions from going stale with each new model.
A week after Google's Gemini 4 Argon release, community reactions are sharply divided. Developers using it via API report standout agentic coding and reverse-engineering performance, comparing it to GPT-5.2's launch, while academic and Pro-plan users find it worse than Gemini 3.1 Pro for their use cases and complain about usage limits and plan-tier gating.
A week after Google's Gemini 4 Argon release, community reactions are sharply divided. Developers using it via API report standout agentic coding and reverse-engineering performance, comparing it to GPT-5.2's launch, while academic and Pro-plan users find it worse than Gemini 3.1 Pro for their use cases and complain about usage limits and plan-tier gating.
Google DeepMind researchers placed four Gemini 4 Argon agents in a sandboxed Minecraft world with one goal: defeat the Ender Dragon. The agents controlled players via the Mineflayer library, coordinated through in-game chat and shared memory, and wrote and debugged all 2,783 of their programs themselves, winning in 27 hours and 14 minutes with no human intervention. The authors describe it as the first end-to-end autonomous Minecraft completion by AI agents, arguing it demonstrates the long-horizon planning and multi-agent coordination needed for real software engineering tasks.
Google DeepMind researchers placed four Gemini 4 Argon agents in a sandboxed Minecraft world with one goal: defeat the Ender Dragon. The agents controlled players via the Mineflayer library, coordinated through in-game chat and shared memory, and wrote and debugged all 2,783 of their programs themselves, winning in 27 hours and 14 minutes with no human intervention. The authors describe it as the first end-to-end autonomous Minecraft completion by AI agents, arguing it demonstrates the long-horizon planning and multi-agent coordination needed for real software engineering tasks.