Tech Frontiers: Musk vs. OpenAI, Military Smart Glasses, Google I/O, and AI World Models

Welcome to this Q&A deep dive into four major stories shaping technology today: Elon Musk's failed lawsuit against OpenAI, the controversial development of smart glasses for military use by Anduril and Meta, the key things to watch at Google I/O, and the emerging field of AI world models. Each question below provides a detailed look at these topics, with internal links for easy navigation.

Why did Elon Musk lose his lawsuit against OpenAI?

Elon Musk's lawsuit against OpenAI was dismissed because a jury determined he filed it too late – the claims were barred by statutes of limitations. The case centered on whether OpenAI violated its founding contract as a nonprofit by shifting toward a for-profit structure. However, the verdict did not rule on that core issue; it only decided that Musk waited too long to bring the case. The timeline was crucial: OpenAI argued that evidence of a for-profit shift was visible as early as 2017, while Musk claimed he only became aware of the change in 2022. The jury sided with OpenAI's timeline, making Musk's 2023 lawsuit untimely. This means the question of whether OpenAI breached its nonprofit mission remains legally unanswered, and the fight over its structure could resurface in another form.

What is the dispute over OpenAI's nonprofit mission about?

At the heart of the dispute is whether OpenAI abandoned its original charter as a nonprofit dedicated to developing artificial intelligence for the benefit of humanity. Elon Musk, a co-founder who left the board in 2018, argued that the company's transition to a for-profit model – including its partnership with Microsoft – violated that founding promise. OpenAI countered that the shift, beginning around 2017 with moves like creating a capped-profit arm in 2019, was necessary to raise the enormous capital required for advanced AI research. The legal battle did not resolve which narrative is correct; it only determined that Musk sued too late. The underlying tension remains, with critics accusing OpenAI of prioritizing profits over safety and ethics, while the company insists the for-profit structure is essential to compete globally without sacrificing its mission.

What are Anduril and Meta developing for military smart glasses?

Defense-tech company Anduril has revealed new details about an augmented-reality headset it is prototyping with Meta for military use. The system, which combines Meta's wearable expertise with Anduril's defense capabilities, is designed to give soldiers enhanced situational awareness and control. According to Quay Barnett, who leads the effort and previously served in Army Special Operations, the goal is to optimize “the human as a weapons system.” Features include eye-tracking for targeting and voice commands to order drone strikes. The glasses would overlay critical data onto the soldier's field of view, such as enemy positions, navigation routes, and real-time intelligence. This represents a significant step toward integrating consumer AR technology into battlefield scenarios, raising both tactical advantages and ethical questions about autonomous decision-making in combat.

What does the Anduril project mean for the future of warfare?

The Anduril-Meta smart glasses could fundamentally change how soldiers interact with technology in combat. By replacing multiple devices (like handheld tablets, radios, and heads-up displays) with a single pair of glasses, the system aims to reduce cognitive load and speed up decision-making. The use of eye-tracking and voice commands lets operators control drones or request fire support almost instantly, blurring the line between human intention and automated action. However, this also introduces risks: over-reliance on such systems could lead to errors if the technology fails or is hacked, and the potential for autonomous strikes raises ethical concerns about accountability. The project highlights a growing trend of using commercial off-the-shelf AR hardware for military applications, accelerating the pace at which soldiers become cyborg-like extensions of networked warfare systems.

What can we expect from Google I/O this week?

Google enters its annual I/O developer conference as a clear third place in the foundation model race, particularly in coding capabilities. Competitors like Anthropic's Claude Code and OpenAI's Codex have outshone Google's tools for months. However, Google still leads in AI for science – areas like weather prediction, protein folding, and materials discovery. At I/O, the company is expected to showcase improvements to its Gemini model, new coding assistants, and possibly updates to its AI infrastructure. The conference will test whether Google can close the gap in developer tools while maintaining its edge in scientific AI. Observers should watch for demonstrations of real-time code generation, integration with Google Cloud, and any surprise announcements about multimodal AI or hardware acceleration. The stakes are high: proving that Google can compete on both fronts – coding and science – is essential for its AI credibility.

What are 'world models' in AI and why are they important?

World models are a new class of AI systems designed to understand and simulate the physical environment, going beyond the text-based limitations of large language models (LLMs). While LLMs excel at generating language, they often fail to grasp real-world physics, causality, and spatial relationships. World models aim to give AI a richer understanding of how the world works – for example, how objects move, how events unfold over time, and how actions cause outcomes. Recent breakthroughs from Google DeepMind, Fei-Fei Li's World Labs, and Yann LeCun's startup have pushed these systems to the forefront. They learn from video, sensor data, and interactive experiences rather than just text. If successful, world models could enable more capable robots, autonomous vehicles, and AI that can plan and reason in complex environments. However, they also raise new safety challenges because such systems could predict and manipulate physical reality.

What is the upcoming MIT Technology Review event about world models?

On Thursday, May 21, MIT Technology Review will host a virtual event examining the progress and future of world models. Editor in Chief Mat Honan and senior reporters will lead discussions with leading researchers from Google DeepMind, World Labs, and other institutions. The event will cover how world models differ from LLMs, their potential applications in robotics and simulation, and the ethical implications of creating AI that understands the physical world. Attendees can expect demonstrations of cutting-edge systems and a frank look at the remaining challenges, such as scaling these models and ensuring they behave safely. This is a must-attend for anyone interested in the next frontier of AI, as world models represent a shift toward more embodied and grounded intelligence. Registration details can be found on the MIT Technology Review website.