AI governance moved onto two fronts overnight. A bipartisan coalition of 26 attorneys general urged Congress to establish federal safety testing, incident reporting, enforcement, and international coordination without preempting state law, while OpenAI and Anthropic leaders called for global oversight at the UN Security Council.
The agent economy is also expanding into enterprise operations and infrastructure. Ema raised $77 million to scale coordinated “AI employee” workflows, Bird secured $450 million in debt financing while opening communications infrastructure to agents, and Perplexity launched a faster search service designed for agentic workloads.
OpenClaw founder Peter Steinberger said the project removed roughly 400,000 lines of low-value tests with little change in code coverage, highlighting how agent-generated code can create maintenance overhead even when conventional metrics look healthy.
Minnesota Attorney General Keith Ellison joined a bipartisan coalition asking Congress to establish federal safety testing, transparent incident response, qualified safety leadership, and international coordination. The letter also asks Congress not to preempt state AI laws and to give state officials authority to enforce federal protections.
OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei told the UN Security Council that advanced AI requires global oversight and democratic accountability. The meeting exposed a policy divide: company leaders and several governments supported common frameworks, while the U.S. representative rejected centralized international governance.
The three labs reportedly aim to launch an industry standards organization by late 2026 or early 2027. Its proposed work includes supporting independent model testing, defining incident-reporting practices, documenting voluntary commitments, and setting qualifications for auditors, though critics warn that the structure could disadvantage open-source developers.
Ema raised a $77 million Series B led by Creaegis, bringing total funding to $140 million. The company says its multi-agent platform automates workflows across existing enterprise software for more than one million active users and prices work by completed tasks and outcomes rather than seats or tokens.
Bird secured a $400 million term loan and a $50 million revolving credit facility led by J.P. Morgan with other lenders. Its new Agentic Harness lets agents send messages, place calls, manage email, and obtain eSIM plans, extending enterprise communications infrastructure to software acting for people and businesses.
Selected posts from the edition, with links to the originals.
“The more time I spend working with coding agents, the more convinced I am that they make software engineering even harder. We can do amazing things with them, but unlocking their full potential requires extraordinary discipline and knowledge.”
View post on X ↗“Web search in Hermes Agent is now fast and free. @perplexity_ai built Fast Search for agents, and it is free for all Nous Portal tiers.”
View post on X ↗“With @GoogleDeepMind, @emblebi, and research partners, we’re making AI-predicted protein complex structures for 2,800+ viruses openly available. This gives scientists a head start in preparing for potential outbreaks.”
View post on X ↗“I think the difficulty of software engineering is essentially constant no matter what abstraction level you move to, because human cognition adapts to new tools until it can fully utilize itself. Tools are only affordances, not a magic wand that makes work disappear.”
View post on X ↗“Today we’re launching our leaderboard for Terminal-Bench-Science 0.1, an agentic benchmark for scientific research work. GPT-6 Astra (max) and Claude Opus 5.5 (xhigh) currently top it at 63% and 62%.”
View post on X ↗“Interesting new paper from Google DeepMind studies agents that follow bad advice from users. Users often suggest a fix that sounds right and is wrong. XYEval adds one confident, misleading hint to tasks from tau2-bench, SWE-bench, Terminal-Bench, HLE and MCP-Atlas.”
View post on X ↗“Introducing Fast Search in the Perplexity Search API. Fast Search runs on Photon, our new Rust-based retrieval and ranking service that we built with a small team of engineers and hundreds of agents. It returns 95% of search results in 230 ms or less.”
View post on X ↗“GPT-6 Astra helps @harvey turn stacks of documents into structured legal drafts, so lawyers can focus more on strategy.”
View post on X ↗“jev is the top model at 1k-10k context on openrouter!”
View post on X ↗“OpenClaw deleted around 400k LOC of its own tests without much change in code coverage. Modern models just love writing tests for every tiny change, even if they aren't useful. This skill helped.”
View post on X ↗