AI News for June 13, 2026 — 7min.ai

[NEWS] US government forces Anthropic to disable Claude Fable 5 and Mythos 5 worldwide: Anthropic shut down access to Claude Fable 5 and Mythos 5 late Friday after the US Commerce Department issued an export control directive at 5:21 PM ET barring any foreign national from using the models, including non-citizen Anthropic employees.
[NEWS] SpaceX closes first trading day at $2.2T, makes Elon Musk world's first trillionaire: SpaceX surged 19% on its Nasdaq debut Friday under ticker SPCX, closing at $160.95 after opening at $150 and briefly hitting $176.52.
[NEWS] Mistral in talks to raise €3B at €20B valuation, nearly double September mark: French AI lab Mistral is in early discussions to raise around €3B ($3.5B) at a roughly €20B valuation, almost double the €11.7B mark set when ASML led its Series C last September.
[TOOLS] Apple's redesigned Siri AI earns positive early reviews in iOS 27 beta; Craig Federighi explains anti-sycophancy design: Early testing of Apple's redesigned Siri AI in iOS 27 beta 1 drew positive coverage from The Verge and 9to5Mac, with reviewers calling it 'good enough at most things' — a marked shift from Siri's decade-long reputation.
[NEWS] Google sues Chinese AI scam network in first joint action with FBI and carriers: Google filed suit against Outsider Enterprise, an alleged Chinese cybercrime group that used Gemini to spin up 131 software kits generating thousands of fake websites impersonating Google, YouTube, USPS, and New York's E-ZPass.
[RESEARCH] Google Research introduces 'faithful uncertainty' to replace hallucinations with hedged guesses: Google researchers published a paper on "faithful uncertainty," a metacognitive technique that aligns LLM responses with internal model confidence so models say "my best guess is" instead of either confidently hallucinating or refusing to answer.
[RESEARCH] PixelRAG ditches HTML parsers, indexes screenshots, beats text RAG by up to 18.1%: Researchers from UC Berkeley, Princeton, EPFL, and Databricks released PixelRAG, a retrieval system that skips text parsing entirely.
[NEWS] Ukraine confirms fully autonomous drones killed Russian soldiers in 2024 test: Aero Center's CEO told New Scientist that fully autonomous drones killed Russian soldiers during a battlefield test roughly two years ago.
[NEWS] OpenAI probed by coalition of state attorneys general on multiple issues: A coalition of state attorneys general has opened an investigation into OpenAI, requesting information across a wide range of topics.
[NEWS] Meta's three-month-old AI unit described internally as a 'soul-crushing gulag': Wired reports Meta's three-month-old Applied AI team of roughly 6,500 engineers and PMs is on the verge of revolt.

About 7min.ai · Subscribe to newsletter