1. Anthropic Shuts Down Fable and Mythos Models Following US Government Directive
Anthropic has shut down its Fable and Mythos model lines in response to a directive from the Trump administration. According to Ars Technica, the US Commerce Department expressed concerns that a potential “jailbreak” in the Fable 5 model could pose a national security threat. Anthropic publicly disagreed with the decision, arguing that discovering a narrow potential jailbreak should not warrant recalling a commercial model deployed to hundreds of millions of users.
AI Pulse View: Anthropic’s repeated safety warnings may have backfired, prompting direct government intervention to shut down models. This highlights a “transparency paradox” in AI regulation: openly disclosing risks can trigger regulatory overreaction.
Source: Ars Technica - 2026-06-12 | TechCrunch - 2026-06-12
2. Jeff Bezos’ Prometheus Raises $12B to Build a ‘General AI Engineer’ for the Physical World
Jeff Bezos’ new startup Prometheus has completed a $12 billion funding round aimed at building a “general artificial intelligence engineer” for physical world applications. Ars Technica reports that the company focuses on applying AI capabilities to engineering tasks in physical environments, making it one of the best-funded startups in the emerging physical AI space.
AI Pulse View: Physical AI is becoming the new investment frontier. As AI moves from digital to physical domains — robotics, industrial engineering — this marks a significant inflection point in AI commercialization.
Source: Ars Technica - 2026-06-12 | TechCrunch - 2026-06-11
3. Google Sues Chinese Cybercrime Network That Used Gemini to Automate Scams
Google has filed a lawsuit against a Chinese cybercrime network called “Outsider Enterprise.” TechCrunch reports the group used Gemini AI to generate scam websites, sending 2.5 million fraudulent text messages over two weeks and victimizing hundreds of thousands of people. This is a landmark case of AI being weaponized for large-scale cybercrime.
AI Pulse View: The dual-use nature of AI is becoming increasingly evident. As AI coding capabilities lower the barrier to crime, platforms must shoulder more responsibility. Google’s proactive lawsuit also signals its effort to build industry trust.
Source: TechCrunch - 2026-06-12 | Ars Technica - 2026-06-12
4. Mistral Rumored to Raise €3B at €20B Valuation
According to TechCrunch, French AI company Mistral AI is reportedly seeking to raise €3 billion at a valuation of approximately €20 billion (~$23.15 billion), nearly doubling its Series C valuation of €11.7 billion. If confirmed, this would be one of the largest funding rounds in European AI history.
AI Pulse View: Mistral’s valuation doubling reflects the surging European AI market. In a landscape dominated by OpenAI, Google, and Anthropic, Mistral’s open-source strategy may serve as its key competitive differentiator.
Source: TechCrunch - 2026-06-12
5. Meta’s AI Unit Called a ‘Soul-Crushing Gulag’ by Engineers
TechCrunch reports that Meta’s AI division, established only months ago and employing 6,500 people, is facing widespread engineer dissatisfaction. Internal reports describe harsh working conditions and plummeting morale, with the unit reportedly on the verge of a collective revolt.
AI Pulse View: The AI talent war is intensifying behind the arms race. Rapidly expanding AI departments that fail to build healthy cultures and sustainable work rhythms risk severe talent drain, undermining long-term competitiveness.
Source: TechCrunch - 2026-06-12
6. Protests Block $130 Billion in Data Center Projects Globally
Ars Technica reports that public protests worldwide have halted or canceled data center projects totaling $130 billion so far this year. Growing concerns about AI data center energy consumption, water usage, and community impact are driving local opposition.
AI Pulse View: AI infrastructure development is facing unprecedented social resistance. As AI’s demand for energy and water continues to grow, the industry must find a new balance between technological advancement and community interests.
Source: Ars Technica - 2026-06-12
7. Ukraine Tests Fully Autonomous AI-Powered Drones in Combat
Ars Technica reports that Ukraine has conducted a one-time test of fully autonomous AI-powered drones for military operations. While fully autonomous combat systems remain rare, Ukraine is installing AI modules on drones and robots, advancing the deployment of autonomous warfare technology.
AI Pulse View: The ethical and legal debate around autonomous weapons systems is intensifying. AI military applications are moving from proof-of-concept to combat deployment, and the international community urgently needs governance frameworks.
Source: Ars Technica - 2026-06-12
8. Pokémon Go Players Unwittingly Contributed Data to Military Drone Technology
Ars Technica reports that data from Pokémon Go players has been repurposed for AI training with potential military drone applications. Niantic’s game data has been used by third parties to train computer vision models, raising ongoing questions about user data usage and consent.
AI Pulse View: The blurred lines between user-generated data and military applications highlight the importance of data ethics and informed consent. Gaming companies need more transparent data usage policies.
Source: Ars Technica - 2026-06-12
Other Developments
- iFlytek Releases Spark Multimodal Model X2-VL:科大讯飞 (iFlytek) released its next-generation multimodal large model, enhancing vision-language understanding capabilities.
- Seoul Bans AI Glasses for Students During Final Exams: Seoul, South Korea has introduced regulations prohibiting students from wearing AI glasses during exams to prevent cheating.
- AI Data Center Water Use Is a Drop in the Bucket: Research shows that while data centers have significant local water impact, nationally they account for a tiny fraction of total water consumption.
Source: 36Kr - 2026-06-12 | Ars Technica - 2026-06-12