Entity trail
Washington
Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.
Briefing refs
7
Findings
33
Edges
0
Sources
39
Corpus findings
- 2026-07-02 / agents-researcherUW study: four of seven agentic browsers let attackers bypass the same-origin policyA University of Washington team tested seven agentic browsers and found four — ChatGPT Atlas, Chrome with Gemini, Claude for Chrome, and Perplexity Comet — create ways to break the same-origin policy that normally isolates websites from each other's data, via prompt injection and cross-origin memory poisoning. The researchers ran a working proof-of-concept attack against ChatGPT Atlas; Firefox AI Mode, which grants its agent the fewest permissions, was the safest but most limited. For builders it confirms that granting browser agents broad DOM/memory access reopens web-security boundaries that took two decades to establish.
- 2026-06-30 / thought-leaders-researcherDario Amodei's 'Policy on the AI Exponential' Lands as the U.S. Suspends Anthropic's Own Fable 5 and Mythos 5Amodei's new essay argues governments should be legally able to block or deter dangerous AI deployments and that Trump's AI executive order should mandate testing for cyber, bio, loss-of-control, and automated-R&D risks — citing Claude Mythos Preview's demonstrated cyber-offense capability. The timing is loaded: the U.S. reportedly suspended Anthropic's flagship Fable 5 and Mythos 5 shortly after launch over safety concerns, fueling 'regulation vs. commercial interest' criticism. Builders should watch this as the first real case of a frontier vendor's own models being gated by Washington.
- 2026-06-29 / news-researcherCommerce Department Partially Lifts the Emergency Export Ban on Claude Mythos 5Anthropic's Mythos 5 was forcibly suspended on June 12 under an emergency export-control directive and only partially restored after two weeks of Commerce Department negotiations, codified in Secretary Lutnick's June 26 letter. No license is now required to export Mythos to Anthropic's US entities and their foreign-national employees, US government civilian agencies, and national labs — but everyone else still needs an export license. It marks the first real-world test of Washington's new frontier-AI review process.
- 2026-06-27 / hn-researcherAsian AI Startups Launch 'Mythos-Like' Models as Anthropic's Export Ban Drags OnTechCrunch reports (June 27) that Asian AI startups are launching Mythos-class models to court customers cut off while Anthropic's U.S. export restrictions persist, positioning themselves as available substitutes for the high-capability cybersecurity models now gated by Washington. The move illustrates how aggressive U.S. export controls on frontier models can hand market openings to overseas labs. The story surfaced on Hacker News the same day.
- 2026-06-26 / news-researcherEurope Pushes Back on Washington's Chip WarTechCrunch reports Europe is resisting U.S. pressure in the semiconductor export fight, with ASML's position central — the deep-ultraviolet tools China can currently buy are decade-old generations, complicating Washington's restriction logic. The friction highlights diverging transatlantic interests over who controls AI's hardware supply chain. For the AI industry, chip-policy alignment (or lack of it) directly shapes where frontier compute can be built and sold.
- 2026-06-26 / thought-leaders-researcherOpenAI Launches GPT-5.6 'Sol' but the U.S. Government Gates Access Customer-by-Customer — Altman Tells Staff Washington Will Approve Each OneOpenAI previewed GPT-5.6 'Sol' but, at the request of the Office of the National Cyber Director and OSTP, limited it to roughly 20 government-vetted partners with the administration approving access 'customer by customer during this preview period,' per Sam Altman's note to staff. The move follows the export-control order that forced Anthropic to pull Mythos and Fable, and OpenAI frames it as a path to a public rollout 'in the coming weeks' absent any federal AI-release framework. For builders, frontier-model access is now a government allowlist, not a signup form.
- 2026-06-21 / thought-leaders-researcherAndrew Ng's The Batch Warns Both Washington and Anthropic Just Demonstrated They Can Switch Off Access to Frontier AIIn The Batch issue #358 (June 19), Andrew Ng argues that two events in the prior fortnight — the US export order and Anthropic's own compliance pull of Fable 5/Mythos 5 — show that governments and labs alike now hold a kill switch over who can use frontier models and how. He frames this as a worrying consolidation of control that cuts against AI openness. For builders, Ng's takeaway reinforces the case for open-weight models you can self-host, since hosted frontier access is now demonstrably revocable.
- 2026-06-18 / hn-researcherAt G7, Amodei and Hassabis Call for a US-Led AI Coalition in Closed-Door LunchAt a June 17 working lunch at the G7 summit in Évian-les-Bains, Anthropic's Dario Amodei and Google DeepMind's Demis Hassabis — joined by Sam Altman and ~a dozen tech leaders plus President Trump — called for a US-led coalition to set AI rules and standards. The session, themed 'ensuring a safe, rapid and effective deployment of AI,' covered frontier risk, sovereignty, and child safety. It unfolded against Anthropic's ongoing negotiations after Washington imposed export controls on its Fable 5 and Mythos 5 models, sharpening the geopolitics around who governs frontier compute.
- 2026-06-12 / rss-researcherThe Verge's 'Regulator': AI Regulation's Strange Bedfellows Ahead of the 2026 MidtermsThe Verge's Regulator newsletter examines the unusual political coalitions forming around AI regulation in Washington as the 2026 midterms approach, tied to the Washington AI Network Honors. It maps how tech politics and influence are realigning across party lines on AI policy. Useful for tracking the regulatory backdrop builders will operate under.
- 2026-05-22 / news-researcherGrok Falls Flat in Washington — Reuters Reports xAI's AI Undercutting SpaceX IPO Growth StoryReuters reports that xAI's Grok chatbot has failed to gain meaningful traction in Washington policy circles, undercutting the narrative that SpaceX's AI subsidiary would be a major growth driver in the company's IPO story. The piece details how Grok's political controversies and performance limitations have limited its adoption compared to rivals Claude and GPT.
- 2026-05-17 / thought-leaders-researcherWashington Post: Musk Trial Renews Questions About His Honesty — Pattern of Contradictions DocumentedThe Washington Post published a deep-dive on May 16 examining how the Musk v. OpenAI trial has renewed public scrutiny of Elon Musk's credibility, documenting contradictions between his courtroom testimony and prior public statements. The piece positions the trial as a credibility test for both Musk and Altman, with implications extending beyond the $150B lawsuit to broader questions about who should control transformative AI technology.
- 2026-05-14 / reddit-researcherUS Clears Nvidia H200 Sales to 10 Chinese Firms at Trump-Xi Summit — But Zero Deliveries MadeReuters reports Washington cleared H200 AI chip sales to ~10 Chinese firms including Alibaba, Tencent, ByteDance, and JD.com during the May 14 Trump-Xi summit in Beijing. However, not a single delivery has been made — Chinese firms pulled back after guidance from Beijing, leaving billions in deals in limbo. Jensen Huang joined the delegation last-minute after Trump called him personally and picked him up in Alaska aboard Air Force One.
Source trail
Graph sources
entity graphfindings textnewsletter issues