★ Top Story
The VergeOn June 12th, the Trump administration issued an export control directive ordering Anthropic to suspend foreign access to its newly released Mythos 5 and Fable 5 models — models Anthropic had described as exceeding any previously released AI in capability. The company spent the weekend in an emergency standoff with the White House, forced to pull its most powerful models offline with little warning. The episode marks one of the most dramatic instances of direct government intervention in a frontier AI company's operations to date.
Read the original →
The VergeA Semafor report indicates that the White House's export control order against Anthropic's Mythos 5 and Fable 5 was triggered in part by intelligence suggesting a China-linked group had already accessed the models. If confirmed, the breach would represent a significant national security incident involving cutting-edge frontier AI. The revelation reframes the ban from a policy dispute into a potential espionage crisis.
Read more →
TechCrunchTechCrunch's analysis argues that the Trump administration's forced takedown of Anthropic's cybersecurity-focused models was less about a specific technical vulnerability and more about a broader assertion of government authority over the AI industry. Whether the move was reactionary, retaliatory, or both, it sends an unmistakable signal that no AI company — however well-connected — is beyond the reach of federal intervention. The incident is being closely watched as a precedent-setting moment for AI governance.
Read more →
The VergeThe forced offline of Anthropic's newest models — including blocking access for the company's own foreign-national employees — has alarmed governments and enterprises worldwide that had been relying on US-based AI providers. The incident is being cited internationally as a concrete argument for investing in sovereign AI capabilities rather than depending on American frontier models. Analysts say it could accelerate AI development efforts in the EU, India, and elsewhere.
Read more →
TechCrunchDozens of prominent cybersecurity professionals have signed an open letter urging the White House to lift export restrictions on Anthropic's Fable 5 and Mythos 5, warning that the ban actively harms defensive security operations more than it protects against adversaries. The experts argue that restricting access to state-of-the-art AI models puts US infrastructure at greater risk by handicapping the defenders who rely on them. The protest adds a national security counterargument to what had been framed as a national security justification.
Read more →
The VergeTech industry lobbyists are making a final push for comprehensive federal AI legislation that would preempt the growing patchwork of state-level AI regulations — a goal that has eluded them for years. The effort is being complicated by attempts to attach AI preemption language to child safety bills, a tactic that has drawn sharp criticism from consumer advocates. With dozens of state AI laws already on the books or in progress, the window for a clean federal override is narrowing fast.
Read more →
Ars TechnicaNvidia is tapping debt markets for the first time in five years, seeking more than $25 billion in a bond offering that will test how much appetite investors still have for AI-sector exposure amid a wave of corporate borrowing. The fundraise signals that even the most cash-generative company in the AI supply chain sees value in locking in capital at current rates. The deal's reception will be closely watched as a barometer of institutional confidence in the AI infrastructure buildout.
Read more →
TechCrunchMeta is launching an "AI Mode" within Facebook search that generates AI-powered answers by drawing on publicly available content from across its family of apps, including posts, Marketplace listings, and more. The feature is part of a broader wave of AI integrations Meta is deploying as it works to close the gap with rivals and deepen user engagement on its platforms. Privacy advocates are likely to scrutinize how broadly "public" content is defined and how it feeds the underlying models.
Read more →
TechCrunchSalesforce has agreed to acquire Fin, an AI-native customer service platform, for $3.6 billion in one of the larger enterprise AI deals of the year. The company plans to integrate Fin's technology and team into Agentforce, its platform for building and deploying custom AI agents across business workflows. The acquisition underscores how aggressively legacy enterprise software vendors are spending to stay competitive as AI-first challengers reshape the CRM and support markets.
Read more →
TechCrunchTens of thousands of workers are being displaced by AI-driven automation at the very moment a small circle of AI insiders is accumulating wealth at an almost incomprehensible scale, creating a volatile social and political dynamic. TechCrunch examines how the widening gap between AI's winners and its casualties is fueling public anger that could reshape the regulatory and labor landscape. The piece argues that without deliberate policy intervention, the backlash could prove destabilizing for the broader tech industry.
Read more →
TechCrunchBengaluru-based AI startup Sarvam has reached unicorn status after closing a $234 million funding round anchored by a $150 million commitment from Indian IT giant HCLTech. Sarvam focuses on building AI models and infrastructure tailored to India's linguistic and cultural diversity, positioning itself as a homegrown alternative to Western frontier AI providers. The deal reflects India's accelerating push to develop sovereign AI capabilities amid growing geopolitical uncertainty around US-based model access.
Read more →
TechCrunchStartup NewCore has launched out of stealth with $66 million in funding, targeting what it sees as the next frontier of enterprise security: managing the identities and access privileges of AI agents operating inside corporate systems. As AI agents increasingly take on roles that were previously held by human employees — with access to sensitive data and the ability to take consequential actions — NewCore argues that existing identity management tools are wholly inadequate. The company's emergence signals a new category forming at the intersection of AI deployment and cybersecurity.
Read more →
TechCrunchMalaysia-based Respond.io has raised $62.5 million to scale its AI agent-driven customer messaging platform, which handles high volumes of customer inquiries and bills per conversation rather than per seat. The company is eyeing acquisitions in North America and Europe as it looks to compete with established players in the enterprise customer communications market. Its per-conversation pricing model reflects a broader industry shift toward outcome-based AI billing.
Read more →
MIT Technology ReviewCasey Harrell, who has ALS and is fully paralyzed, has now logged thousands of hours using a brain-computer interface to communicate — making him the most experienced long-term user of speech-enabling BCI technology in the world. His sustained use over nearly three years is providing researchers with an unprecedented dataset on how these implants perform and evolve outside of controlled lab conditions. The case is being hailed as a milestone in demonstrating that BCIs can function as durable, everyday assistive tools rather than short-term experiments.
Read more →
MIT Technology ReviewSouth Korea has emerged as one of the world's most AI-saturated societies, with AI-powered systems embedded in immigration, transit, healthcare, and daily consumer life in ways that would feel striking to most Western visitors. MIT Technology Review explores the cultural, historical, and policy factors — including a strong state-led technology push and a population accustomed to rapid digital adoption — that have made South Korea a living laboratory for AI integration. The country's experience offers both a preview of where other nations may be headed and a set of cautionary lessons about speed outpacing deliberation.
Read more →No stories from this source today.