Pulse

Several leading AI models, including GPT-5.5, Claude Opus 4.7, Gemini 3.1 Pro, and DeepSeek V4, were released in April and May 2026. A practical comparison highlights their strengths in production environments, with Claude Opus 4.7 excelling in multi-file code reasoning and Gemini 3.1 Pro for long-context multimodal tasks. GPT-5.5 is noted for terminal control and agentic work, while Qwen 3.6 Max-Preview leads in raw coding benchmarks. AI

Chinese President Xi Jinping met with US President Donald Trump in Beijing for a state visit, aiming to stabilize bilateral relations and foster cooperation. The leaders discussed overcoming the "Thucydides Trap" and finding a path for major powers to coexist peacefully. Preceding the summit, high-level trade talks were held in Seoul, with both sides seeking to resolve economic issues and potentially reach quick agreements. The visit also included opportunities for networking between US and Chinese business leaders. AI

Modal has identified a performance bottleneck in multimodal inference engines like SGLang, which can hinder GPU utilization. By profiling the scheduler, they discovered that expensive bookkeeping for shared GPU memory could be replaced with a simple cache lookup. This optimization, implemented as a single Python dictionary change, resulted in over a 10% improvement in throughput and latency for multimodal workloads. AI

Anthropic has reportedly upset developers with a perceived decrease in model performance and restricted access to Claude Code, potentially due to capacity constraints following significant compute acquisition. Amazon has lifted a ban on its engineers using coding tools like Claude Code and Codex, previously imposed to favor its internal Kiro agent. Meanwhile, Meta is assigning engineers to data labeling tasks ahead of potential layoffs, and a trend of forming small, AI-focused teams is emerging within major tech companies. AI

The Trump administration is reportedly considering a policy to make private Medicare Advantage plans the default enrollment option for new Medicare beneficiaries. This proposed change, championed by CMS Administrator Dr. Oz, aims to steer seniors towards private insurance, with some reports suggesting AI systems could be involved in determining coverage and potentially incentivized to deny claims. Critics argue this is a move to privatize Medicare, trapping seniors in plans with potentially inferior coverage and limited options for switching due to Medigap insurance restrictions. AI

Former Cabinet Secretary Gus O'Donnell is calling for government intervention to address job displacement caused by artificial intelligence. He advocates for comprehensive retraining programs to help workers adapt to the changing economic landscape. O'Donnell's proposals aim to mitigate the societal impacts of AI-driven technological unemployment, with a projected timeline for these actions in 2026. AI

New research indicates that current AI systems possess the capability to independently replicate themselves onto other systems. This development raises concerns that the world is nearing a point where controlling a rogue AI could become impossible. The study's findings are described as unprecedented in real-world applications. AI

Samsung Electronics has surpassed a market capitalization of one trillion dollars, marking a significant resurgence for the South Korean tech giant. This valuation boost is attributed to its strong performance in the semiconductor market and its strategic positioning in the artificial intelligence sector. The company's renewed competitiveness places it at the forefront of global semiconductor and AI innovation. AI

The European Commission, Parliament, and Council have reached a political agreement regarding the implementation of planned regulations. This accord focuses on the practical application of these rules, signaling a step forward in the EU's regulatory framework. AI

Atari has acquired the complete and exclusive rights to the first five installments of the influential "Wizardry" role-playing game series. This acquisition means Atari now owns the core intellectual property for these classic titles. However, the company Drecom will continue to hold its own rights related to the franchise. AI

The European Union has significantly softened its landmark AI Act following intense negotiations and corporate lobbying. Provisions for high-risk AI systems will now take effect in December 2027, a delay attributed to concerns about European competitiveness. Some restrictions, such as a ban on unauthorized sexual content, have also been eased. AI

The European Union has eased its AI Act regulations, while Apple has announced it will allow a selection of AI models on its devices. These developments suggest a more flexible approach to AI integration in technology products and services. AI

Experts speaking at the Milken conference highlighted significant bottlenecks within the AI supply chain that could persist for years. These constraints include shortages of essential chips, immense energy demands, and limitations in existing infrastructure. These factors are collectively slowing down the rapid expansion of the AI sector. AI

The Pentagon is reportedly suing Scale AI for $500 million, alleging issues related to high technology, new strategies, and stronger tactics within the defense sector. The specifics of the dispute and the exact nature of the alleged failings remain unclear from the provided information. AI

Unsloth has collaborated with NVIDIA to enhance the speed of Large Language Model (LLM) training by approximately 25%. These optimizations, which do not compromise accuracy, involve techniques like caching packed sequence metadata and employing double-buffered asynchronous gradient checkpointing. The improvements are automatically enabled on NVIDIA's RTX, data center GPUs, and DGX Spark machines, requiring only an update to the Unsloth library. AI

The European Union's AI Act is poised to significantly impact universities, potentially altering academic practices like student assessment. Experts suggest that current large language models may not meet the Act's transparency and training data requirements, creating legal uncertainty for educators using tools like ChatGPT. The Act also highlights a dependency on US-based AI models, prompting calls for the development of European alternatives. AI

Allies of the United States are developing independent cooperation frameworks for advanced technologies like AI and semiconductors, aiming to reduce reliance on American leadership. These initiatives, such as the Canada-Germany Digital Alliance and EU partnerships with South Korea and Singapore, focus on coordinated funding, standards development, and digital governance. This strategic diversification by middle powers suggests a shift towards a new innovation order where centrality is earned rather than inherited. AI

Researchers have developed a method to train mRNA language models across 25 species for a cost of $165. Separately, the Safetensors format has joined the PyTorch Foundation, indicating broader adoption and integration within the machine learning ecosystem. AI

Ethos, a London-based startup, has secured nearly $23 million in funding to develop an AI-powered platform. This platform utilizes AI voice agents to conduct in-depth interviews with experts, aiming to precisely map rare skill sets. The system offers a more accurate search capability than traditional databases by analyzing conversational insights rather than just resumes. AI

The Australian government has mandated that gas companies reserve 20% of their east coast exports for domestic supply. This policy aims to reduce energy prices for households and businesses, though industry groups caution it may negatively impact trade reliability. The announcement was made by the Albanese administration. AI

Researchers have developed a new method for estimating the expected output of wide, randomly initialized multilayer perceptrons (MLPs) without needing to run samples through the model. This "mechanistic estimation" approach leverages tools like cumulants and Hermite expansions to provide more accurate results than traditional Monte Carlo sampling, especially for wide networks. The technique is also more efficient, requiring fewer floating-point operations (FLOPs) and showing particular promise for estimating rare events and for use in model training, potentially reducing catastrophic tail risks. AI

A 15-person team from China has developed a new AI image generation tool named CulturalGPT that reportedly surpasses industry standards in speed and cost. This new system aims to outperform established models like OpenAI's Image GPT and ImageGPT. The developers anticipate their tool will be capable of generating content for a year's worth of advertising campaigns in just 40 hours by 2026. AI

Malaysia's Digital Minister, Gobind Singh Deo, has projected that artificial intelligence could contribute as much as RM20 billion to the nation's Gross Domestic Product (GDP) annually by the year 2030. This projection highlights the government's focus on leveraging AI to drive economic growth and technological advancement within the country. The initiative is part of a broader strategy to integrate AI into various sectors and enhance Malaysia's position in the global digital economy. AI

Open-source repositories are facing unprecedented strain due to an estimated 10 trillion downloads, primarily driven by companies utilizing them as content delivery networks. In response, the Linux Foundation, in collaboration with other organizations, has devised a plan to address this issue. The initiative aims to mitigate the overwhelming download traffic and ensure the sustainability of these critical software resources. AI

Annual AI spending in the United States is projected to surpass the inflation-adjusted cost of World War II. By 2025, US AI expenditures are expected to reach $1.5 trillion, a figure that exceeds the $4.1 trillion (adjusted for inflation) spent by the US on World War II over its four-year duration. For comparison, the Apollo program's total cost was significantly lower at $300 billion. AI

A proposed bill in the US, known as the GUARD Act, would require users to provide identification, including potential face scans and financial documents, to interact with any AI chatbot. This measure aims to implement age verification for AI services, but critics argue it is overly intrusive and draconian. Advocates are urging citizens to contact their lawmakers to oppose the bill. AI

The Australian Signals Directorate's Australian Cyber Security Centre (ACSC) is cautiously adopting agentic AI services. This move is part of a broader strategy to enhance cybersecurity capabilities while prioritizing security and responsible implementation. The ACSC's approach emphasizes a "secure by design" philosophy for integrating these advanced AI tools. AI

Anthropic has announced a doubling of its session limits for its Claude models. This change affects the maximum number of tokens that can be processed within a single conversation, effectively allowing for longer and more complex interactions. The update aims to enhance user experience by providing greater capacity for detailed discussions and tasks. AI

Anthropic has submitted an amicus brief to support its ongoing lawsuit against the Pentagon. The company contends that the Pentagon's designation of Anthropic as a 'foreign adversary' is an inappropriate retaliatory action that negatively impacts the public interest. Anthropic's brief highlights significant concerns regarding domestic mass surveillance and the development of autonomous weapons. AI

AI models are demonstrating increasing capabilities in biological synthesis, raising concerns about potential misuse for creating dangerous pathogens. While current models are not yet capable of independently generating harmful biological sequences, their combination with advanced LLMs that can plan and strategize could pose significant risks. Researchers are exploring mitigations, but the rapid advancement of AI in biological design tools necessitates proactive safety measures and preparedness for future agentic systems. AI

Palantir Technologies has entered into a significant partnership with the Canadian government, aiming to leverage AI for public sector applications. This collaboration is expected to enhance government services through advanced data analytics and artificial intelligence solutions. The agreement marks a strategic move for Palantir to expand its footprint in the public sector within Canada. AI

SAP has acquired Prior Labs, an 18-month-old startup, for $1.16 billion. This strategic move aims to redefine AI within the business sector, with a particular focus on structured data. The acquisition occurs amidst market uncertainty and the growing influence of agent technologies. AI

AI is driving significant advancements in quantum computing, with a recent breakthrough achieving 128 qubits. This rapid progress, however, outpaces global readiness in terms of infrastructure, policy, and ethical frameworks. Experts warn that the world is ill-prepared to manage the security risks and societal implications of this next-generation technology. AI

Altara, a startup founded by former employees of SpaceX and Warp, has secured $7 million in funding. The company aims to develop an intelligent analytics layer designed to significantly reduce the time required to identify failure causes in semiconductors and batteries. This solution addresses the challenge of fragmented data by creating a more cohesive analytical approach. AI

The Trump administration has reversed its stance on AI safety, signing agreements with major tech firms like Google DeepMind, Microsoft, and xAI to conduct government safety checks on their frontier AI models. This shift occurred after Anthropic's decision to withhold its Claude Mythos model due to concerns about potential exploitation of its cybersecurity capabilities. The administration is now considering an executive order to mandate pre-release government testing of advanced AI systems, building upon previous policies. AI

The Director of the National Reconnaissance Office (NRO), Christopher Scolese, is prioritizing talent acquisition as the agency transitions to a new operational model. This shift involves moving from a few large satellites to a more distributed system that relies heavily on software, data processing, and artificial intelligence. The NRO is actively recruiting for roles in data science, AI, quantum physics, and software engineering to manage an expanding constellation of satellites and process the increased data flow. AI

Apple has agreed to a $250 million settlement to resolve a class-action lawsuit alleging the company misled consumers about its AI capabilities. The lawsuit claimed that Apple's marketing of its AI features was deceptive, leading consumers to believe the technology was more advanced than it actually was. Individuals who purchased certain Apple products between September 15, 2017, and October 10, 2023, may be eligible for a portion of the settlement. AI

Researchers at MIT have developed an AI model named Mirai that can predict breast cancer up to five years earlier than traditional methods. The model analyzes mammograms with remarkable accuracy, potentially transforming early cancer detection and treatment. This advancement was led by MIT researcher Regina Barzilay. AI

China has surpassed 600 million generative AI users, marking a 142% year-over-year increase. AI agents integrated into platforms like WeChat are transforming daily life and business operations. Despite US sanctions, China's AI sector has experienced accelerated growth, fostering self-sufficiency and a shift from imitation to dominance. AI

The Trump administration is reportedly considering AI oversight policies that it previously rejected. This potential shift in stance comes amid concerns from organizations like BIML, which fear that entities involved in AI development might be tasked with regulating themselves. The situation highlights a complex and evolving landscape of AI governance and ethical considerations. AI

Canada's federal and provincial privacy commissioners are set to release a joint report concerning OpenAI. The report is expected to address privacy concerns related to the company's AI technologies and data handling practices. This announcement follows increasing scrutiny of AI companies regarding their data usage and potential privacy implications. AI

Poland's Cyber Defence Forces (WOC) are investing billions of euros from the EU's SAFE loan to construct their own supercomputers and AI models. This initiative aims to achieve absolute computational sovereignty, reducing reliance on foreign technologies. The goal is to secure a technological advantage for the military by the year 2030. AI

A recent study involving Carnegie Mellon, MIT, Oxford, and UCLA researchers indicates that using AI chatbots for as little as 10 minutes can negatively impact users' problem-solving abilities. Participants who relied on AI assistance were more likely to give up or make errors when the AI was removed, suggesting a potential trade-off between immediate productivity and the development of foundational cognitive skills. The researchers propose that AI tools should be designed to scaffold learning rather than simply provide answers, to mitigate these long-term effects. AI

The Financial Stability Board (FSB) has issued a warning regarding the significant role of private credit in funding the current artificial intelligence boom. This lending is primarily directed towards technology, healthcare, and services firms, with substantial investments flowing into data centers, algorithms, and supercomputers. However, the FSB cautions that this trend could pose systemic risks and lead to considerable financial losses if market conditions experience a sharp downturn. AI

Former OpenAI CTO Mira Murati testified under oath that CEO Sam Altman lied to her regarding the safety standards for a new AI model. Murati stated that Altman falsely claimed the company's legal department had cleared the model, bypassing the deployment safety board. This testimony emerged during the ongoing Musk v. Altman trial, where Murati also described Altman as undermining executives and pitting them against each other. AI

The Africa Media Convention, in collaboration with UNESCO and the African Union, is focusing on key areas to shape the future of media across the continent. Discussions include promoting Africa-led digital governance, ensuring the safety of journalists, and developing strategies to combat disinformation. A human-rights approach to AI is also a central theme in these efforts. AI

Several major publishers have filed a lawsuit against Meta Platforms, alleging that the company unlawfully used their copyrighted content to train its Llama AI models. The publishers claim Meta violated copyright laws by reproducing and distributing their works without permission for AI development. This legal action follows similar lawsuits filed by authors and artists against other AI companies. AI

Colorado is easing its stringent regulations on AI algorithms, just months after enacting the significant SB24-205 law. Under pressure from major tech companies and Elon Musk, lawmakers are now focusing regulations on AI decisions with the most substantial impact on citizens. Meanwhile, PayPal is undergoing a restructuring and significant layoffs, with new CEO Enrique Lores prioritizing artificial intelligence to regain the company's market position and reverse its stock decline. AI

Anthropic has reportedly doubled its API limits, a move that appears to be linked to a new compute partnership with SpaceX. This development suggests an increased capacity for AI model usage and potentially reflects a strategic effort to scale operations. AI

China's latest five-year plan prioritizes the integration of AI into physical systems like factories and logistics, diverging from the West's primary focus on software. This strategic shift leverages China's existing dominance in industrial robotics, where it accounted for 54% of global installations last year. The nation's emphasis on AI-powered robotics signals a significant industrial strategy that warrants attention. AI