Edition: May 6, 2026

Elon Musk attempted to recruit OpenAI CEO Sam Altman to join a Tesla AI lab in 2018, offering him a board seat, as revealed in emails and testimony presented in the Musk v. Altman trial. Musk's lawsuit claims Altman and OpenAI president Greg Brockman stole a nonprofit, using Musk's $38 million investment to create a private company worth over $800 billion. The trial highlights the complex and contentious history between Musk and OpenAI.

Billionaire media mogul Barry Diller expressed trust in OpenAI CEO Sam Altman, but emphasized that trust is irrelevant as artificial general intelligence (AGI) nears, citing the unknown consequences of AI development as the primary concern. Diller believes that most AI leaders, including Altman, are good stewards, but notes that the issue is dealing with the unknown and establishing guardrails for AGI. The development of AGI could have profound and unpredictable effects on humanity, making it essential to consider the potential risks and consequences.

xAI has partnered with Anthropic, selling its excess compute capacity and shifting its focus towards building data centers, potentially positioning itself as a neocloud business. This move could be worth billions of dollars and may indicate a change in Elon Musk's priorities. The partnership also highlights xAI's transition from a consumer to a provider of compute.

The vLLM V1 migration required fixes to match the vLLM V0 reference, including changes to processed rollout logprobs, runtime defaults, and the final projection. The fixes ensured that vLLM V1 returned rollout logprobs in the expected form, allowing for a successful migration. The corrected vLLM V1 run showed close alignment with the vLLM V0 trajectory across key metrics.

Anthropic has partnered with SpaceX's xAI to utilize its computing resources, including the Colossus 1 supercomputer, to support its AI models. This deal comes as SpaceXAI prepares to go public and could bolster its credibility with investors. The partnership may also lead to the development of orbital AI compute capacity, or data centers in space.

Elon Musk's departure from OpenAI in 2017 was sparked by a dispute over control of the company, with Musk demanding full control and the other co-founders refusing to acquiesce. The incident has come to light as part of an ongoing legal battle between Musk and OpenAI's current leaders. Musk eventually stopped donating to the company and left the board, paving the way for the current lawsuit.

A new study from researchers at Carnegie Mellon, MIT, Oxford, and UCLA found that using AI chatbots for as little as 10 minutes can negatively impact people's problem-solving skills and ability to think. The study suggests that relying on AI can lead to a decline in foundational skills, despite potential short-term productivity gains. Researchers recommend rethinking how AI tools are designed to prioritize learning over providing direct answers.

Mira Murati, OpenAI's former CTO, testified that CEO Sam Altman lied to her about safety standards for a new AI model, stating he falsely claimed the model didn't need to go through the company's deployment safety board. This testimony was given in a video deposition during the Musk v. Altman trial. Murati's statement suggests a breakdown in trust and communication within OpenAI's leadership.

SpaceX is considering a $55 billion initial investment to build a semiconductor factory in Texas, with total potential spending of $119 billion on the 'Terafab' project. The factory aims to produce chips for AI servers, satellites, and autonomous vehicles. The project is part of Elon Musk's efforts to ensure sufficient computing power for his AI company, xAI.

DeepSeek, a Chinese AI lab, is in talks to raise its first round of venture capital with a potential valuation of $45 billion, more than doubling its initial valuation of $20 billion. The investment round is reportedly led by China's state investment vehicle, aiming to support homegrown AI technology. This move is seen as part of China's efforts to develop its own AI capabilities and reduce reliance on US technology.

Google DeepMind has partnered with EVE Online developer Fenris Creations to test AI models in the game's complex, dynamic environment. The partnership will involve controlled experiments in a specially designed offline version of the game. The goal is to study intelligence in complex systems and develop AI systems with long-horizon planning and continual learning.

Genesis AI, a robotics startup backed by Khosla, has unveiled its first model, GENE-26.5, which features advanced robotic hands designed in-house, capable of performing tasks such as cooking and lab work. The company aims to bridge the gap between AI and robotics with its full-stack approach.

Samsung's valuation has reached $1 trillion, driven by surging demand for its memory chips used in AI systems, with shares rising over 10% on news of potential chip manufacturing deals with Apple. The company's profits have increased eight times from the same period last year, fueled by high-bandwidth memory chip sales. However, Samsung faces challenges including a potential worker strike and increased costs for its phone and TV divisions.

The US government has added four new AI companies to its list of preferred suppliers, including Microsoft, Amazon, and Nvidia, while reevaluating its relationship with Anthropic AI. The new agreements will allow the companies' products to be used on classified operations, expanding the Pentagon's AI capabilities. The move comes after a public disagreement with Anthropic AI over the use of its technology for surveillance and autonomous weapons.

AlphaEvolve, a Gemini-powered coding agent, has improved the accuracy of DNA sequencing error correction by 30% in collaboration with Google Research and PacBio. This breakthrough enhances genetic data analysis, potentially leading to discoveries of previously hidden disease-causing mutations. The innovation is expected to lower costs and increase accuracy for researchers.

Apple has agreed to pay $250 million to settle a class-action lawsuit over false advertising claims related to its Siri AI features, which plaintiffs say were overhyped and not delivered as promised. The settlement covers US customers who bought iPhone 15 or 16 models between June 2024 and March 2025. Eligible claimants may receive up to $95 per device.

Google is testing Remy, a new AI personal agent for Gemini, designed to take actions for users in work and daily tasks, with a focus on user control and autonomy. The tool is being tested in a staff-only version of the Gemini app and is intended to integrate with Google services and monitor user preferences. Remy's development is part of Google's broader work to expand Gemini beyond chat-based responses.

Marc Lore, founder of Wonder, plans to use AI to enable anyone to open a virtual restaurant in under a minute, leveraging a network of tech-enabled kitchen locations. The initiative, Wonder Create, will allow users to design and launch their own restaurant brand, with AI handling tasks such as naming, branding, and recipe creation. This innovation aims to revolutionize the dining industry by making it easier for entrepreneurs to experiment with new food concepts.

India's Securities and Exchange Board has issued a red alert to the nation's equities industry, warning of potential cyberattacks sparked by Anthropic's Mythos AI bug-finding tool. The regulator has advised market players to revisit their information security systems and develop new strategies to mitigate risks. A taskforce has been established to examine the risks posed by models like Mythos.

Claude has partnered with SpaceX to increase its compute capacity, resulting in higher usage limits for its customers, including doubled rate limits for Pro, Max, Team, and seat-based Enterprise plans. The partnership provides access to over 300 megawatts of new capacity and 220,000 NVIDIA GPUs. This expansion is part of Claude's efforts to improve its services and meet growing demand.

Google is updating its AI search to include quotes from Reddit and other sources, aiming to provide more context and diverse perspectives in search results. This change may help users find answers to niche queries, but also raises concerns about the potential for inaccurate or misleading information. The update adds complexity to the role of Google's AI Overviews, which may now serve as a gateway to various sources rather than a definitive answer.

Ethos, a London-based expert network, has raised $22.75M in a Series A round led by a16z to improve its AI-powered voice onboarding and matching capabilities. The platform uses voice interviews to gather more detailed information about experts' knowledge and experience, allowing for better matches with companies seeking advice. This funding will help Ethos scale its network and further develop its technology.

Inexpensive seafloor-hopping submersibles are being developed to map the seafloor for critical mineral deposits, while conversational AI tools are being used in military war rooms to provide advice on potential targets. These advancements raise concerns about environmental impacts and AI-generated errors. The use of synthetic turf is also being debated due to its potential contribution to microplastics and environmental pollution.

QuTwo, a Finnish AI lab founded by Peter Sarlin, has reached a valuation of $380 million after raising $29 million in an angel round. The company's core product, QuTwo OS, is an orchestration layer that directs tasks to classical, quantum, or hybrid architectures. QuTwo aims to become a leading AI company with a focus on enterprise use cases.

The Open ASR Leaderboard has added new private datasets from Appen Inc. and DataoceanAI to prevent benchmaxxing and improve the quality of speech recognition benchmarks. These datasets cover various accents and speech styles, and can be optionally included in the leaderboard's average WER calculation. The addition of these datasets aims to provide a more holistic view of ASR performance and promote real-world robustness.

Shivon Zilis, a key advisor to Elon Musk and mother of four of his children, has become a liability for the billionaire, with her testimony in the Musk v. Altman trial raising questions about her role and relationship with Musk. Zilis worked for Musk's AI portfolio, including Tesla, Neuralink, and OpenAI, and had a romantic relationship with him. Her testimony has sparked interest and controversy, with many wondering about her involvement with Musk.

Anthropic has raised usage limits for its Claude Code service, doubling the five-hour window limits for Pro and Max plan subscribers, after securing a deal with SpaceX to utilize the company's data center in Memphis, Tennessee. This move aims to increase capacity for Anthropic's users. The deal also includes increased API limits for the Opus model.

Snap's $400 million deal with Perplexity has ended amicably, with the companies parting ways after a planned integration of Perplexity's AI search engine into Snapchat fell through. The deal was announced in November and would have seen Perplexity pay Snap $400 million over one year. Despite the ended deal, Snap reported a 5% year-over-year growth in daily active users to 483 million.

Google has shut down Project Mariner, an experimental feature designed to perform tasks across the web, with its technology being integrated into other Google products. The project was first revealed in 2024 and allowed users to perform up to 10 tasks at a time. The shutdown was announced on the Project Mariner landing page.

David Sacks, the White House's AI and Crypto Czar, has faced a significant setback in his efforts, with the administration considering a government review of AI models before release, a reversal of previous policies. This development comes after Sacks had been a vocal champion of less regulation. The move has significant implications for the tech industry.

Anthropic has introduced a new AI feature called 'dreaming' that enables agents to learn from their experiences and improve performance, sparking concerns about the trend of naming AI features after human cognitive processes. This naming convention can create unrealistic expectations and blur the line between human and machine capabilities. The author argues that AI companies should stop using human-like names for AI features to avoid anthropomorphism and promote a clearer understanding of AI limitations.

Google Search is offering tools to help gardeners plan and maintain their gardens, including visualizing layouts, building planting schedules, and identifying plant issues. The tools can be used to create "chaos gardens" and mini plots, which are currently trending. Users can access these features through AI Mode in Search.

Hackers and cybercriminals are expressing frustration with the increasing use of generative AI in online forums, citing the production of low-quality content and the undermining of their claims to skilled expertise. A recent study analyzed nearly 100,000 AI-related conversations on cybercrime forums, revealing a growing pushback against AI-generated posts. The backlash highlights the complex social dynamics within these communities, where users value human interaction and authenticity.

Match Group, the owner of Tinder, is slowing down its hiring process to allocate more funds for AI tools, aiming to become an AI-native company and increase productivity. The company expects the impact to be cost-neutral, with lower headcount offsetting increased software expenses. This move comes as Tinder shows signs of a potential turnaround, with revenue slightly up after quarters of decline.

HP's AI and Data Science Business Development Manager discusses the challenges of leveraging data for AI ingestion and the importance of governance and integration in automated data processing. The company advises clients to treat model updates like code deployments and implement validation gates to mitigate risks. HP's hardware solutions, such as the Z series, are designed to handle the demands of autonomous AI lifecycles.

Apple has agreed to pay $250 million to settle a class action lawsuit over its marketing of AI features, including a significantly upgraded version of Siri, which were delayed or not fully available at launch. The settlement affects US customers who purchased the iPhone 15 or iPhone 16 between June 2024 and March 2025. Eligible customers may receive up to $95 per device.

TechCrunch Disrupt 2026 will feature a panel on mergers and acquisitions, providing founders with advice on M&A strategies and options. The event will be held in San Francisco from October 13-15, with a limited-time ticket offer available. The panel will include industry leaders such as Aklil Ibssa from Coinbase and Lindsey Mignano from Mignano Law Group.

Google's AI search is being updated to include quotes and perspectives from firsthand sources like Reddit, making it easier for users to find trustworthy information. This change aims to address the growing trend of people seeking advice from others online. The update will link search queries with online conversations around similar topics.

Microsoft's LinkedIn chief, Ryan Roslansky, has taken on additional responsibilities, now leading the company's Work Experiences Group, which includes Microsoft Teams. This change is part of a broader leadership reshuffle following the retirement of Rajesh Jha, executive vice president of Microsoft's experiences and devices group. The move aims to streamline the company's operations and leadership structure.

Google Chrome's AI features may be consuming up to 4GB of computer storage due to a large on-device AI model file. The file, named weights.bin, is connected to Google's Gemini Nano AI model, which powers various Chrome AI tools. This unexpected storage usage may cause unexplained drops in available desktop device storage.

TechCrunch Disrupt 2026 is offering a limited-time discount of 50% off a second ticket for attendees who register before May 8. The event, taking place from October 13 to 15 in San Francisco, will feature industry leaders and founders discussing the latest trends and innovations. Attendees can expect to gain credibility and visibility through interactive sessions and networking opportunities.

Hasan Piker, a prominent left-wing commentator, spends most of his time streaming on Twitch and engaging in online discussions, with a strong stance against AI and its potential misuse. He is known for his outspoken views on politics and social issues, earning him a large following. Piker's daily life involves extensive screen time and online interactions.