Edition: May 11, 2026

Ilya Sutskever, former OpenAI chief scientist, testified in court about his role in Sam Altman's brief removal as CEO, stating he didn't want OpenAI to be destroyed. Sutskever, who has a $7 billion stake in OpenAI's for-profit arm, also discussed the company's need for funding to build advanced AI models. His testimony supported Elon Musk's claims that Altman may not be the right leader for OpenAI.

OpenAI has launched Daybreak, an AI initiative that uses the Codex Security AI agent to detect and patch vulnerabilities before attackers can find them. Daybreak creates a threat model based on an organization's code and automates the detection of high-risk vulnerabilities. The launch comes after rival Anthropic announced its own security-focused AI model, Claude Mythos.

General Motors has laid off over 600 IT workers, around 10% of its IT department, to make room for new hires with stronger AI skills, as part of its efforts to transform its Information Technology organization for the future. The company is seeking expertise in AI-native development, data engineering, and cloud-based engineering. This move signals a significant shift in enterprise AI adoption, with a focus on rebuilding the workforce from the ground up.

Thinking Machines, an AI company founded by former OpenAI CTO Mira Murati, is developing 'interaction models' that enable real-time collaboration between humans and AI. These models can process audio, video, and text simultaneously, allowing for more natural interactions. The goal is to revolutionize how people work with AI.

Nobel-winning economist Daron Acemoglu shares his insights on the impact of AI on jobs, stating that AI agents are not a replacement for human workers but rather tools to augment certain tasks. He believes that AI's ability to handle task orchestration will determine its impact on jobs. Acemoglu also notes a new hiring trend in AI companies, where they are building in-house economics teams to better understand the economic implications of AI.

Google has stopped a zero-day hack that was developed with the help of artificial intelligence, which could have been used to bypass two-factor authentication on a web-based system administration tool. The exploit was discovered by Google's Threat Intelligence Group, who found hints of AI assistance in the Python script used for the exploit. This marks the first time Google has spotted and stopped an AI-developed zero-day exploit.

The UK tech industry, which relies heavily on international AI talent, is struggling with manual and inefficient sponsor licence management, despite advancements in automation, posing a significant operational risk. This vulnerability can lead to revoked sponsor licences, affecting not only companies but also the skilled workers they employ. The sector's inability to automate immigration compliance is a glaring exception in an otherwise automated HR compliance process.

Cowboy Space Corporation has raised $275 million to develop its own rocket program, aiming to launch space data centers and address the current shortage of rockets. The company plans to launch its first rocket before the end of 2028, competing with established players like SpaceX and Blue Origin. This move is driven by the growing demand for AI compute and limited options on Earth.

The finance industry is undergoing a significant transformation with the adoption of advanced AI technologies, which is happening rapidly and often without a clear plan or governance in place. As a result, executives must reconcile productivity gains with oversight, risk, and accountability. The key to successful AI adoption lies in reframing its role and making it an ambient capability that disappears into existing processes.

Researchers propose 'radical optionality' as a third way for AI regulation, where governments invest in tools and institutions to prepare for potential future crises. This approach involves avoiding overregulation while building capacity to respond to various scenarios. The goal is to preserve democratic governments' ability to make good decisions about transformative AI systems.

A Hollywood writer and showrunner has turned to training AI as a side job due to the entertainment industry's decline, and now secretly works on assessing and improving chatbot responses, annotating images and videos, and testing safety precautions. The writer was drawn to AI training after the 2023 Hollywood strike and a Facebook post about a company paying $150 an hour for writers. The job has proven to be challenging and not as easy as expected.

Nvidia's CUDA platform is considered the company's most valuable asset, providing a significant competitive advantage in the AI industry through its ability to parallelize tasks and optimize GPU performance. CUDA's capabilities have allowed Nvidia to maintain its lead in the market, despite the rise of open-source AI models. The platform's importance is highlighted by its role in reducing the workload and cost of training AI models.

Bain & Company estimates a $100 billion market in the US for SaaS companies using agentic AI to automate coordination work in enterprise systems. The market is largely untapped, with vendors currently capturing only $4-6 billion. The report identifies various enterprise functions with significant automation potential, including sales, operations, and customer support.

Google Finance is expanding its AI-powered platform to Europe, offering features such as AI-powered research, advanced visualizations, and real-time market intel. The platform provides comprehensive and personalized financial information, including live earnings calls and AI-generated insights. This expansion aims to enhance users' understanding of the financial world.

Companies can achieve greater success with AI by adopting a 'customer-back engineering' approach, prioritizing customer needs and experiences when developing technology solutions. This strategy involves working backward from customer challenges to design and build solutions. By doing so, organizations can create more effective and customer-centric AI innovations.

Amazon Web Services (AWS) provides a foundation for training and inference of foundation models, leveraging open-source software and infrastructure to support large-scale distributed training. The AWS building blocks include compute, network, and storage, which interact with common open-source stacks across the foundation model lifecycle. This setup enables machine learning engineers and researchers to understand systems bottlenecks and scaling characteristics.

Digg, a once-popular link-sharing site, has relaunched as an AI news aggregator, aiming to track influential voices and surface relevant news. The site uses real-time data ingestion and sentiment analysis to rank top stories and key figures in the AI space. Digg's success is uncertain, but it could provide a useful resource for those tracking AI news.

Journalist Joanna Stern has left her position at The Wall Street Journal to start her own media company, New Things, and has written a book about AI called I Am Not a Robot. Stern spent a year living with AI and has insights into the technology's potential and limitations. She is now using AI to help launch her new company.

Chargers have undergone significant technological advancements in recent years, becoming smaller, safer, and faster, with innovations like gallium nitride and multi-port chargers. These changes have repositioned chargers as standalone devices rather than peripheral accessories. Manufacturers are continuing to push the boundaries of charging technology to meet the demands of a growing connected ecosystem.

A hantavirus outbreak on a cruise ship has resulted in three deaths, but experts believe the situation can be contained. Meanwhile, the trial between Elon Musk and OpenAI has entered its second week, with new revelations about Musk's motivations and actions. Additionally, concerns are being raised about the potential for large language models to supercharge mass surveillance in the US.