Edition: June 17, 2026

The Trump administration has imposed export controls on Anthropic's AI technology due to concerns over its access to South Korean telecom giant SK Telecom, which has ties to China. Anthropic has disabled access to its models entirely rather than restricting it based on nationality. The move has led to a standoff between the White House and Anthropic over bringing the technology back online.

World leaders, including French President Emmanuel Macron and Indian Prime Minister Narendra Modi, have expressed concerns over the US potentially cutting off access to American AI models, highlighting the need for digital sovereignty and a 'trusted partners' scheme. The issue was discussed at the G7 Summit, where leaders also talked about the creation of a network to bypass US restrictions. The goal is to maintain open trade and access to advanced AI models while ensuring national security.

Anthropic's newest AI models, Fable 5 and Mythos 5, were taken offline after the Trump administration ordered the company to cut access for all foreign nationals, citing national security concerns. The move has raised questions about the use of export controls to regulate AI access. The legal basis for the order remains unclear.

In Shenzhen, China, a startup called IO-AI Tech is developing technology that allows workers to remotely control humanoid robots using VR headsets and motion-tracking gear, with the goal of automating tasks in factories and convenience stores. The company is working with local manufacturers to refine its prototypes and train robots to perform specific tasks. This technology has the potential to unlock more capable and general AI models in the future.

Social media platforms such as Threads, Instagram, and TikTok are introducing tools that allow users to personalize their algorithms, giving them more control over the content they see in their feeds. This shift reflects an evolution in recommendation systems, moving away from a one-size-fits-all approach and towards a more tailored experience. Users can now train their own algorithms and influence what appears in their feeds, with benefits including a more engaging and relevant experience.

Odyssey, a world model AI startup, has raised $310 million in a Series B round, reaching a $1.45B valuation with backing from Amazon and other major investors. The company's technology simulates the physical world with accurate physics, with applications in areas like video-game creation and robotics. Odyssey will now optimize its models to run on Amazon's AWS Trainium chips.

A recent Pew Research poll found that 63% of Americans believe AI is advancing too quickly, despite a significant increase in chatbot usage since 2024. Younger generations are more likely to use AI, but also tend to have a more pessimistic view of its impact. Only 16% of respondents think AI will have a positive impact on society.

A new Pew Research study reveals that only 16% of Americans believe AI will have a positive impact on society, with 40% expecting a negative impact and 67% doubting government regulation. Despite skepticism, many Americans use AI daily, with 44% using ChatGPT. The study highlights a generational and gender divide in AI perceptions and usage.

The White House is requiring Anthropic to prevent jailbreaks on its AI model, Claude Fable 5, before it can be rereleased, but experts question whether this is technically possible. The administration wants Anthropic to proactively test for potential jailbreaks and report them to the government. However, cybersecurity experts believe that guardrails on AI models are only a temporary solution and can be bypassed by skilled users.

Google has introduced a new smart home speaker powered by its Gemini AI, allowing for natural language requests and multistep commands. The $99.99 device offers advanced features, including two-way conversations and nuanced question-answering. Some features will require a $10/month Google Home Premium subscription.

Researchers have introduced MolmoMotion, a language-guided 3D motion forecasting model that predicts the future motion of objects in 3D space, outperforming existing methods. The model uses a sparse set of surface points to describe motion and can be used in downstream applications such as robotics and video generation. MolmoMotion is being released openly, along with a large dataset and benchmark, to facilitate further research and development.

XDOF, a startup emerging from stealth, has raised $70 million to build data pipelines and collection tools for robotics companies, aiming to address the lack of training data for robots to interact with the physical world. The company is already working with several top AI labs and has partnered with UC Berkeley's AI Research lab to release a large collection of high-quality robot training data. This move is seen as a crucial step in advancing the development of capable robots.

New research published in Nature showcases the capabilities of AMIE, a medical AI system, in managing health conditions by tracking symptoms and fine-tuning medications. AMIE was found to match clinicians in overall management reasoning and scored higher in plan preciseness and guideline alignment. This breakthrough could potentially support medical care and give physicians more time to spend with patients.

Pramaana Labs has raised $27 million in seed funding to develop formal verification for AI, focusing on high-stakes industries like law and drug discovery. The startup combines conventional large language models with deterministic verification, using tools like the LEAN programming language. This approach aims to bring reliability and accuracy to AI systems in sensitive fields.

Google has released a new Home Speaker powered by its Gemini assistant, offering improved natural conversation understanding and features like Continued Conversation and Gemini Live. The speaker is available for preorder starting June 17 and costs $100. It comes in four colors and offers a range of features, including integration with security cameras and smart home devices.

Canada Pension Plan Investment Board has committed up to $741 million to Indian data center operator CtrlS, betting on India's growing role in the global buildout of cloud and AI infrastructure. The investment will help CtrlS expand capacity and build infrastructure tailored for AI workloads. This move is part of a broader trend of global investors funding India's data center boom.

DeepL, a German translation startup, has acquired Mixhalo, a real-time audio startup, to enhance its translation suite for live events. The acquisition will allow DeepL to expand into the live event space with its voice-to-voice translation suite. This move aims to improve translation experiences at conferences and other events.

Researchers are working on solar geoengineering to counteract global warming, but the project faces significant engineering challenges and investment requirements. Meanwhile, scientists are making progress in understanding interoception, the sense of how we feel inside, with implications for treating conditions like obesity and anxiety. SpaceX has surpassed Amazon in market value, reaching $2.659 trillion.

Pinterest has launched an experimental AI shopping app called 'Ask Pinterest' that allows users to ask questions using natural language to get personalized recommendations and inspiration. The app is designed to expand the visual discovery experience beyond the main app to a conversational, chatbot-like interface. The move is part of Pinterest's efforts to explore a more conversational approach to shopping and product discovery.

The UK government is deploying Google Cloud generative AI to automate council planning operations, aiming to reduce application decision timelines by 50%. The 'Extract' application and 'Augmented Planning Decisions' prototype will help process unstructured data and alleviate administrative backlogs. This move is expected to support the UK's target of constructing 1.5 million new homes by 2029.

The Strands Robots SDK integrates with LeRobot to enable a seamless workflow from the Hugging Face Hub to robot hardware, allowing for simulation, training, and deployment of robot tasks with minimal code changes. This walkthrough demonstrates a single agent loop that records demonstrations, trains policies, and deploys to physical robots. The integration enables coordination of multiple robots through a peer mesh.

The next generation of humanoid robots may not resemble humans, with designs focusing on functionality rather than appearance, as seen in Genesis AI's new robot Eno. Eno is designed for general-purpose use, with human-like hands but a non-traditional body. This shift in design philosophy could lead to more versatile and efficient robots.

Researchers introduce GLM-5.2, a new AI model designed for long-horizon tasks, boasting a solid 1M-token context and improved coding capabilities. The model outperforms its predecessor and other open-source models in various benchmarks, showcasing its potential for sustained engineering work. GLM-5.2 is released under an MIT open-source license, making it accessible to developers worldwide.

Scientists are exploring the concept of solar geoengineering, which involves artificially cooling the planet by releasing materials into the stratosphere to reflect sunlight. Researchers are moving beyond computer simulations to focus on the practical engineering challenges, including designing high-altitude aircraft and mastering the chemistry of dispersing materials. However, critics argue that pursuing this research could make it more likely to be deployed despite the potential dangers and unknown consequences.

The Agentic Resource Discovery specification is a new open standard that enables agents to search for tools, skills, and other agents at runtime, rather than relying on pre-installed capabilities. Developed by contributors from Microsoft, Google, and others, ARD aims to simplify the discovery and integration of agent capabilities. The Hugging Face Hub has implemented ARD, providing search access to thousands of skills, ML applications, and MCP servers.

The 'slowtech' revolution is gaining momentum as people seek to reduce their phone addiction and regain control over their attention span, with some opting for older technologies like iPods and wired headphones. This movement is driven by a desire for a more mindful approach to technology use. Investors and companies are taking notice, with some developing products and apps aimed at reducing screen time.

Roelof Botha, former Sequoia Capital managing partner, has joined SpaceX's board of directors, bringing extensive public company experience and a deep audit committee background. Botha has a long-standing relationship with CEO Elon Musk, having worked together at PayPal in 2000. He will serve on the board's audit committee until SpaceX's next annual shareholder meeting.

Snap's stock plummeted over 5% after unveiling its new AR glasses, Specs, which come with a hefty price tag of nearly $2,200. The cost has raised concerns about the product's profitability, particularly among its core teenage user demographic. The company's CEO, Evan Spiegel, justified the price by comparing Specs to high-end computers or laptops.

NEA partner Tiffany Luck discusses the current state of AI adoption in enterprises, highlighting the tension between hype and return on investment. Many companies are still figuring out how to measure the effectiveness of their AI spending. Luck shares her insights on the future of AI and its potential for creating 'magic moments' in consumer business.

Anthropic, an AI startup, has joined the Frontier carbon removal coalition, contributing to a new $915 million funding tranche and marking its arrival as the first AI company to join the group. This move signals a potential shift in Anthropic's approach to energy and sustainability. The funding brings the total pledged to Frontier to $1.8 billion, supporting the removal of 1.8 million tons of carbon.

NEA partner Tiffany Luck discusses the shift from 'tokenmaxxing' to measuring return on AI investment, the future of personal agents, and the AI IPO landscape. Luck highlights how startups are helping enterprises track AI spend and the value being created across the AI stack. The conversation explores the tension between AI hype and ROI in the tech industry.

A feud between the White House and Anthropic, an AI company, is unfolding, with CEO Dario Amodei attending a G7 Summit on innovation and AI. The details of the feud are still emerging, but it may have significant implications for the development of frontier AI. The situation is complex, with multiple factions and interests at play.

Snap has debuted its new $2,195 Specs glasses, a device aimed at bringing computing into the world and making it more human. The glasses are designed to help people stay connected to their surroundings without looking at their phones. Snap CEO Evan Spiegel showcased the Specs in a recent interview, highlighting their potential to change how we interact with technology.

Meta has introduced a new AI Mode in Facebook's search bar, designed to tackle complex queries by drawing on public posts across Meta apps. The feature aims to provide more personalized and informative results, but its reliability is still uncertain. As with any AI-powered tool, there are concerns about its potential to spread misinformation.

Google's first new smart speaker in six years, the Google Home Speaker, will start shipping on June 25th with preorders opening today. The $99 speaker features a round design, touch-capacitive buttons, and a light ring, available in four colors. It was built for Gemini for Home and has seen no hardware changes since its announcement nine months ago.

Entrepreneurs in Nairobi are turning to solar power to support their businesses, with companies like Agsol providing solar-powered grain mills that can increase profitability by up to 80%. These off-grid solutions are helping Kenya move towards its goal of universal electricity access by 2030. The cost of solar technology has decreased significantly, making it a more viable option for businesses.