Edition: May 4, 2026

AI systems may soon be able to build themselves, with a 60% chance of no-human-involved AI R&D happening by 2028, marking a significant shift in AI research and development. This could lead to an automated AI R&D process, with profound implications for society. The author cites public information and trends in AI capabilities as evidence for this prediction.

OpenAI cofounder Greg Brockman testified in court, defending his $30 billion stake in the company and stating that his financial interests are secondary to OpenAI's nonprofit mission. Brockman's testimony came as part of the Musk v. Altman trial, in which Elon Musk is attempting to remove Brockman and CEO Sam Altman from power. Brockman emphasized that he and others had put 'blood, sweat, and tears' into building OpenAI

Cerebras Systems, an AI chipmaker and partner of OpenAI, is preparing for an initial public offering that could raise $3.5 billion and give it a $26.6 billion market cap. The company's AI-specific chip challenges GPU-based AI chips and has garnered significant investor interest. A successful IPO could pave the way for larger tech offerings, including potentially OpenAI and SpaceX.

Image AI model releases are driving significant growth for AI mobile apps, generating 6.5 times more downloads than traditional model updates, with examples including ChatGPT and Gemini seeing tens of millions of new downloads after releasing their respective image models. However, increased downloads do not always translate to higher revenue. The trend marks a shift from earlier days when chatbot upgrades drove more demand.

Google announced several major AI updates in April 2026, including the Gemini Enterprise Agent Platform, Gemma 4 open model, and new tools like Google Vids and Deep Research Max. These updates aim to make AI more accessible and helpful for various tasks. The company also highlighted its progress in using AI to improve healthcare and education.

Elon Musk's expert witness, Stuart Russell, testified in the OpenAI trial, warning of the dangers of an artificial general intelligence (AGI) arms race and the need for regulation. Russell, a University of California, Berkeley professor, expressed concerns about the pursuit of AGI and safety. The trial highlights the tension between corporate interests and AI safety concerns.

Colin Angle, the creator of Roomba, has unveiled a new robot companion called 'Familiar' through his startup Familiar Machines & Magic. The robot is designed to interact autonomously with family members and has an expressive face with movable features. It is set to be showcased at the WSJ Future of Everything conference.

Sierra, an AI startup founded by Bret Taylor, has raised $950 million in funding, valuing the company above $15 billion. The company plans to use the capital to become the global standard for AI-powered customer experiences. Sierra's platform is already handling billions of interactions for over 40% of the Fortune 50 as customers.

Elon Musk sent ominous texts to OpenAI's president Greg Brockman, suggesting they settle a lawsuit two days before the trial began, with Musk implying they would be hated if they didn't comply. The exchange was deemed inadmissible in court, but has raised questions about Musk's true motives. The trial, which could significantly impact OpenAI's structure and technology, is ongoing.

Anthropic and OpenAI are launching joint ventures to provide enterprise AI services, with Anthropic's venture valued at $1.5 billion and OpenAI's at $10 billion. The ventures will offer preferred sales access to investors' portfolio companies and allow for more engineering resources to be devoted to individual clients. This move comes as both AI labs are fundraising at a rapid pace and considering possible IPOs.

Elon Musk is suing OpenAI, alleging the company has reneged on its original nonprofit mission, with the trial potentially setting the company back as it plans to go public. The case centers on whether Musk was deceived by OpenAI's leaders about the company's for-profit operations. The trial has drawn significant attention due to the high stakes and the feud between Musk and OpenAI's Sam Altman.

The Gemini API has introduced Webhooks, a push-based notification system, to reduce friction and latency for long-running jobs, allowing developers to build complex applications more efficiently. This feature eliminates the need for continuous polling, providing real-time updates when tasks are completed. The implementation prioritizes reliability and security, adhering to the Standard Webhooks specification.

The rise of 'shadow AI' has led to the development of AI-BOMs, which provide visibility into AI components and their interactions within an organization. This is crucial for securing enterprise supply chains infused with AI applications. Companies like Cisco are open-sourcing tools like AI-BOMs and Model Provenance Kits to help track and secure AI assets.

Nvidia CEO Jensen Huang predicts an era of 'AI factories' where intelligence is the primary output, but a crisis is brewing in cloud storage architectures due to increasing demands of agentic AI. Current cloud storage systems were not designed to handle these demands. This could impact the development of AI technologies.

The AI market is targeting various healthcare functions, from curing cancer to streamlining administrative tasks, but execution can be difficult due to the complexity of healthcare. AI applications for healthcare are rapidly increasing, with over 1,300 AI-enabled medical devices approved by the FDA. However, poorly designed or inadequately trained AI tools can put patients at risk, and regulators are keeping a close eye on development and adoption.

The increasing use of Physical AI in autonomous systems, such as robots and industrial equipment, raises governance questions around testing, monitoring, and stopping their actions. As the global Physical AI market is projected to reach $960.38 billion by 2033, companies like Google DeepMind are developing models like Gemini Robotics to control robots directly. Ensuring safety limits and escalation paths are part of system design is crucial for these physical systems.

Google has introduced agentic AI governance as a native product feature, addressing a significant gap in enterprise AI deployments. The Gemini Enterprise Agent Platform provides a comprehensive platform for building, scaling, and governing agents, with governance and security at its core. This move is seen as a response to the lack of centralized governance in enterprise AI, with most organizations still struggling to deploy AI agents at scale.

TeamViewer ONE is transforming IT operations by utilizing agentic AI to proactively identify and resolve tech issues before they become problems, reducing digital friction and increasing productivity. This shift from reactive to proactive IT support can help organizations improve efficiency and competitiveness. With the help of AI, IT teams can now monitor thousands of endpoints and intervene early to prevent issues.

The Five Eyes security alliance has warned that agentic AI technology is too unstable for rapid rollout, citing potential misbehavior and amplification of existing organizational frailties. The alliance recommends slow and careful adoption, prioritizing resilience over productivity. A guide outlining 23 risks and over 100 best practices has been released to address these concerns.

A Chinese court has ruled that it's illegal to fire human workers and replace them with AI, establishing a new legal principle that using AI does not automatically justify terminating a contract. This decision was announced just before Labour Day, a holiday that celebrates workers' rights. The ruling may have significant implications for employers in China who are considering replacing human workers with AI systems.

A new enterprise AI services company has been formed by Anthropic, Blackstone, Hellman & Friedman, and Goldman Sachs to bring AI solutions to mid-sized companies across sectors. The company will work with clients to identify areas where AI can have the most impact and build custom solutions. This new firm is backed by a consortium of leading alternative asset managers and will become a member of Anthropic's Claude Partner Network.

OpenAI's president, Greg Brockman, testified in a trial against Elon Musk, showcasing a defensive demeanor and correcting minor details during cross-examination. Brockman's journal has been a key witness for Musk's case, and Brockman himself has become a significant figure in the trial. The trial's outcome may have significant implications for the AI industry.

Microsoft has reversed a change to VS Code that added a default AI attribution notice, giving credit to Copilot for human-authored code, after developers complained about the feature. The change was intended to add transparency but ended up causing frustration among users. The fix will make the attribution trailer opt-in instead of opt-out.

DoorDash has introduced new AI-powered tools to streamline merchant onboarding, enhance food images, and create websites based on app listings. The tools include AI Retouch and AI Replate for photo editing, as well as a feature to automate website creation. These additions aim to reduce friction and help merchants succeed on the platform.

TechCrunch Disrupt 2026 is offering a limited-time buy one, get one 50% off deal on passes for the event, which will take place from October 13-15 in San Francisco. The offer ends on May 8 and allows attendees to bring a colleague or partner to the event, which features 300+ showcasing startups and 10,000+ founders, investors, and tech leaders. Attendees can save up to $410 on passes by registering now.