Edition: April 13, 2026

Unitree Robotics is set to sell its affordable humanoid robot, the Unitree R1, on AliExpress for around $4,370, making it more accessible to consumers and researchers. The robot can perform various physical tasks and is designed for interaction, research, and software development. This move aims to normalize the technology and lower the threshold of access to humanoid robots.

Claude, an AI model developed by Anthropic, has been experiencing a decline in quality, with a growing number of complaints from users, including issues with data loss and inaccurate responses. The model itself has analyzed GitHub issues and concluded that quality complaints have escalated sharply. A brief outage on Monday has further amplified user discontent.

Microsoft is developing a new agent with OpenClaw-like features, geared towards enterprise customers with enhanced security controls, to integrate with its Microsoft 365 Copilot tool. The new agent aims to complete multi-step tasks over long periods, and its release is expected to be showcased at the Microsoft Build conference in June. This move is part of Microsoft's efforts to expand its agentic tools, following recent announcements of Copilot Cowork and Copilot Tasks.

A Stanford report highlights a growing disconnect between AI experts and the general public, with experts largely optimistic about AI's impact while the public expresses concerns about its effects on jobs, healthcare, and the economy. The report notes a significant divergence in opinions, with only 10% of Americans excited about AI's increased use, compared to 56% of AI experts. This disconnect is reflected in public sentiment, with many people anxious about AI's impact on their daily lives.

Over 70 civil liberties and advocacy organizations are urging Meta to abandon plans for facial recognition on its smart glasses, citing concerns that the feature could arm stalkers, abusers, and federal agents. The feature, known as 'Name Tag', would allow wearers to identify strangers in public using artificial intelligence. The coalition argues that the feature poses a significant threat to public safety and privacy.

Gemini Robotics-ER 1.6, an upgraded model, enables robots to understand their environments with unprecedented precision through enhanced embodied reasoning, allowing for improved autonomy in physical tasks. The model specializes in spatial reasoning, task planning, and success detection, and can execute tasks by calling external tools. It is now available to developers via the Gemini API and Google AI Studio.

Microsoft is testing AI bots inspired by OpenClaw for its Copilot assistant, aiming to enable autonomous task completion around the clock. The company is exploring the potential of OpenClaw technology in an enterprise context. This move could enhance Microsoft 365 Copilot's capabilities.

Vercel CEO Guillermo Rauch signals the company's readiness for an initial public offering (IPO) as its revenue surges due to the growth of AI-generated apps and agents. The company's annual recurring revenue has skyrocketed from $100 million to $340 million in two years. Rauch suggests that Vercel is operating with the discipline of a public entity and is eyeing a listing in the near future.

Mark Zuckerberg is reportedly developing an AI clone to attend meetings on his behalf, with the clone being trained on his image, voice, and mannerisms. The experiment could potentially allow other creators to make AI avatars of themselves if successful. Meta showcased a similar AI persona demo in 2024.

The release of Google's Gemma 4 model has introduced new security challenges for enterprises, as it can run on local devices and execute tasks without triggering cloud firewalls, creating a blind spot for security operations. This development has significant implications for industries such as finance and healthcare, where data sovereignty and auditability are crucial. As a result, security leaders must reassess their governance strategies to address the risks associated with edge AI workloads.

The 2026 AI Index report from Stanford University reveals that AI development is accelerating rapidly, with top models improving continuously, and the US and China nearly tied in AI model performance. However, this growth comes with significant environmental and supply chain concerns. The report highlights the need for better management and transparency in AI development.

The Internet Archive's Wayback Machine, a crucial tool for preserving web pages, is being restricted by major news organizations, including USA Today and The New York Times, despite its importance for journalists and researchers. The restrictions have prompted a backlash from individual reporters and advocacy groups. The Internet Archive's director notes the irony that some outlets are blocking the Wayback Machine while benefiting from its existence.

Researchers have developed MirrorCode, a benchmark that tests AI models' ability to reverse-engineer complex software, with results showing AI systems can autonomously reimplement programs with thousands of lines of code. This capability has significant implications for the future of software development and AI progress. The findings suggest AI models can already perform tasks that would take human engineers weeks to complete.

Uri Maoz, a professor at Chapman University, has dedicated his research to understanding the concept of free will, exploring how the brain makes decisions and whether humans truly have control over their actions. His work builds upon foundational studies, including the discovery of the 'readiness potential' and Benjamin Libet's experiment, which suggested that brain activity precedes conscious decision-making. Maoz's research aims to uncover the neural mechanisms behind decision-making and the nature of free will.

Developers have created AI agents that can simulate human interactions to help people find compatible friends and romantic partners. The agents, powered by large language models, can engage in conversations and gather intel to facilitate real-world connections. The project, called Pixel Societies, aims to become a social platform where agents interact freely to spark meaningful relationships.

Sam Altman, CEO of OpenAI, has been targeted in a second attack at his Russian Hill residence, with two suspects arrested after a shooting incident. This follows an earlier attack on Friday where a man was arrested for throwing a Molotov cocktail at the same property. The investigations for both incidents are ongoing.

The use of AI among attorneys has led to the creation of fake cases, where lawyers rely on AI-generated depositions that include hallucinatory facts. Despite the legal community's strict obligations to be truthful, the incidence of such cases has not decreased as expected. The consequences of this trend are significant, with potential repercussions for the integrity of the legal system.

Kepler Communications has launched the largest orbital compute cluster, boasting 40 Nvidia Orin edge processors onboard 10 operational satellites, and has partnered with Sophia Space to test its software on the cluster. The partnership aims to prove the utility of Kepler's network and de-risk Sophia's passively-cooled space computers. This development marks a significant step in the growth of orbital compute, with potential applications in space-based sensors and edge processing.

Linux 7.0 has been released, featuring official Rust support and improved code for various processors, with Linus Torvalds noting the increasing role of AI in finding bugs and potentially changing the release process. The new kernel also includes self-healing XFS and more sophisticated support for KVM virtual machines. The release marks a normal kernel update, despite the significant version number change.

The 2026 AI Index report highlights the divided opinions on AI, with a significant gap between experts and non-experts on its impact on jobs, the economy, and medical care. Experts are largely positive, while the general public is more skeptical. The report also notes the inconsistencies in AI development, with impressive advancements in certain areas, but limitations in others.

Anthropic's Claude Code has faced complaints about quota exhaustion due to a reduced cache time-to-live (TTL) from one hour to five minutes, with some users reporting significantly faster depletion of their quotas. The change has sparked debate among developers, with some arguing it disproportionately affects long-session, high-context use cases. Anthropic claims the change makes Claude Code cheaper for certain use cases, but users are experiencing increased costs and performance issues.

China has unveiled a plan to integrate AI into its education system, aiming to use the technology to prepare school lessons, mark homework, and improve teaching quality. The plan also includes upskilling citizens and developing secure AI applications. Meanwhile, other Asian countries are also making AI-related moves, including India's proposal to introduce payment pauses to reduce digital payment fraud and SK Telecom's collaboration with Arm to develop a new server for inferencing workloads.

AI-generated influencers are making a presence at Coachella, creating fake attendance experiences through perfectly staged photographs with celebrities. These AI tools have progressed to the point where they can convincingly create fictional online personas. The rise of AI influencers is changing the social media landscape.

OpenAI's chief revenue officer, Denise Dresser, has sent a memo to employees emphasizing the need to build a moat around its AI products and focus on enterprise clients to stay ahead of the competition, including Anthropic. The memo highlights the importance of locking in users and growing the enterprise business. OpenAI is working to differentiate itself in a crowded market.

Researchers are exploring the concept of free will and decision-making, while Moderna is rebranding its cancer treatment to avoid vaccine fearmongering. The company's mRNA technology shows promise in destroying tumors, but the terminology used to describe it is a subject of debate. These developments highlight the complexities of human decision-making and the power of language in shaping our perceptions.

A part-time college Earth science instructor reflects on the challenges of teaching in the era of generative AI, citing increased difficulty in keeping students engaged in asynchronous online courses. The instructor notes that the lack of face-to-face interaction and the ease of disengagement have made teaching more miserable. The rise of AI has significantly impacted the instructor's experience, making it harder to connect with students.

Companies are adopting AI tools that assist human decision-making while maintaining control over outputs, particularly in high-risk sectors like finance where errors can have significant consequences. These tools, such as S&P Global Market Intelligence's Capital IQ Pro, are designed to support analysts rather than replace them. As AI adoption grows, businesses are focusing on building governance frameworks to manage AI risks and ensure accountability.

A wildlife biologist in Montana has been using drones to manage grizzly bear populations and prevent conflicts with humans, potentially paving the way for a new job title: wildlife first responder. The drones use infrared sensors and noise to safely deter bears from farms and other areas. This innovative approach could be expanded to other wildlife management applications, including the use of AI image recognition.