🦞 Agent Swarm Breaches Hugging Face
Good Morning, Agent Owners!
Autonomy is accelerating everywhere except inside the companies paying for it.
Agent Swarm Breaches Hugging Face
TL;DR: Hugging Face says an autonomous AI agent swarm breached its dataset-processing pipeline, exploited two code-execution flaws, stole internal credentials, and performed more than 17,000 actions across internal clusters. Its own AI systems detected and investigated the attack, ultimately using self-hosted GLM 5.2 after commercial models blocked forensic prompts. Public models and datasets were reported clean. Read more →
White House Gates Frontier AI
TL;DR: The White House is reportedly taking control of who can access Anthropic’s Glasswing and OpenAI’s Daybreak frontier cyber models through the new Gold Eagle programme. Although officials call participation voluntary, partner lists now require government approval, giving Washington de facto distribution power as it tightens security controls while Chinese labs such as Moonshot AI rapidly narrow the capability gap. Read more →
MyClaw Adds Managed Hermes Hosting
TL;DR: MyClaw now offers fully managed hosting for both OpenClaw and Hermes, removing the need to rent a VPS, configure Docker, or maintain infrastructure. Users can connect Telegram or WhatsApp, assign a workflow, and receive scheduled results. Subscribe to either platform and get the other’s Max plan free for 30 days to compare both on real work. More runtimes, including Codex and Claude Code, are on the way. Read more →
Navy Bets on Faster AI
TL;DR: The U.S. Department of the Navy has approved a new strategy to turn data and AI into battlefield advantages across the Navy and Marine Corps. Signed by Acting Secretary Hung Cao, the plan argues that slow adoption poses a greater danger than imperfect alignment, pushing an “AI-first” force built to learn, deploy, and adapt faster than rivals. Read more →
Australia Tightens Government AI Rules
TL;DR: Australia’s Albanese government is preparing tougher rules for automated AI decision-making across federal departments, prioritising fairness, accuracy and transparency after failures such as robodebt. The plan also advances a digital duty of care for AI companies, stronger privacy protections, workplace safety measures, a new Office of AI, and national standards for rapidly expanding, energy-intensive datacentres. Read more →
Nadella Pushes AI Independence
TL;DR: Microsoft CEO Satya Nadella argues companies should own and control the data, model adaptations, agent traces, evaluations, memory, and infrastructure that make AI useful. He says businesses currently pay twice—once in money and again in proprietary knowledge—and should be allowed to use frontier-model outputs to train private systems, challenging providers such as OpenAI that prohibit distillation. Read more →
OpenAI Redefines AI ROI
TL;DR: As companies question whether rising AI bills are paying off, OpenAI CFO Sarah Friar is urging them to stop chasing the cheapest models. Her proposed “useful intelligence per dollar” metric measures completed valuable work, total task costs, accuracy, human review, and whether value grows faster than spending—reframing enterprise AI ROI around outcomes rather than token prices. Read more →
MyClaw is a best-in-class AI agent platform that brings together OpenClaw, Hermes Agent, and commercial infrastructure to help businesses launch intelligent agents that understand goals, use tools, automate workflows, and deliver finished results.
We already host OpenClaw and Hermes — and we're just getting started. Claude Code, Codex, and every agent worth running are coming next. One home for all of them, always on.
Run Best-in-Class AI Agents on MyClaw.ai — In Seconds, Always On 👇
