Editor’s Note

This week’s technology news carries a subtle but significant theme: convergence.

Artificial intelligence is no longer developing in isolation. It is merging with robotics, media production, space infrastructure, and global computing networks. Each step seems incremental on its own. Taken together, however, they sketch the outlines of something larger: an emerging technological ecosystem capable of acting, deciding, and building in the physical world.

This raises some important question that we’ll explore over the next several issues of Exponential Times: What happens as AI moves from tools toward agency? Toward General AI? Toward Superintelligence?

We'll consider these questions in our closing reflection, beginning this week, with "The Race toward AI Agency."

Top Story

Nvidia Announces Chips for Orbital AI Data Centers

Nvidia revealed a new computing system called Vera Rubin Space-1, designed specifically to power AI data centers in orbit. The system was unveiled at the company’s GTC 2026 conference as part of a broader push toward what CEO Jensen Huang described as the arrival of “space computing.”

The new hardware is engineered to operate in the harsh conditions of space while delivering extremely dense AI computing power. The goal is to support the next generation of satellite networks and orbital infrastructure, potentially allowing massive AI workloads to run beyond Earth’s surface.

The concept of orbital data centers is gaining attention as AI computing demand skyrockets. In space, solar energy is abundant and heat can dissipate into the vacuum more easily than on Earth, potentially easing the growing energy demands of terrestrial data centers. Sources: The Tech Buzz | Wikipedia

Why it matters

AI’s growth increasingly depends on enormous computing resources. If orbital data centers become viable, they could dramatically expand global computing capacity while reshaping the physical infrastructure of the internet. In other words, the cloud may someday extend far beyond Earth.

Musk Confirms Tesla–xAI “Digital Optimus” Initiative

Elon Musk recently announced that Tesla and xAI, his artificial-intelligence company, are collaborating on a joint project called “Digital Optimus,” also referred to as “Macrohard.”

According to Musk, the system will combine xAI’s Grok language model with a Tesla-developed AI agent capable of interpreting real-time computer screen activity and interacting with keyboard and mouse inputs. The goal is to create a highly responsive AI system that can automate complex digital tasks—potentially allowing software agents to perform many forms of office work or operational workflows.

The project builds on Tesla’s recent $2 billion investment in xAI, part of a broader integration of Musk’s companies around artificial intelligence and robotics.

The announcement also arrives while Musk faces an ongoing shareholder lawsuit alleging conflicts of interest related to the founding of xAI and its overlap with Tesla’s AI ambitions. Sources: Electrek | Reuters

Why it matters

Digital Optimus represents a step toward AI systems capable of acting autonomously in digital environments—watching screens, interpreting information, and completing tasks. Such systems could become the backbone of the emerging “agent economy,” where AI manages workflows, research, logistics, and software operations with minimal human supervision.

Apple and Netflix Double Down on AI-Driven Media Production

Two major entertainment companies made notable moves into AI-powered creative tools this week. Apple acquired the video-editing company MotionVFX, while Netflix announced the purchase of InterPositive, an AI filmmaking technology company founded by actor and entrepreneur Ben Affleck.

InterPositive focuses on AI tools designed to assist filmmakers during post-production. Instead of generating entirely new scenes, the system analyzes a production’s existing footage—known as “dailies”—to help solve common filmmaking challenges such as reframing shots, correcting lighting problems, or removing stunt wires.

Netflix executives emphasized that the technology is intended to support filmmakers rather than replace them, giving directors and editors more control over complex production workflows.

Meanwhile, Apple’s acquisition signals growing interest among major platforms in tools that make video editing faster and more accessible for creators. Sources: Reuters | The Guardian | Bloomberg

Why it matters

The entertainment industry may be entering an era where AI acts less like a replacement for artists and more like a creative collaborator. As these tools mature, the cost and complexity of producing professional-quality film and television could fall dramatically, opening the door for smaller teams and independent creators to compete with major studios.

OpenAI Delays ChatGPT “Adult Mode” Again

OpenAI has again delayed plans to release an “adult mode” version of ChatGPT, a feature that would allow the AI to engage with more explicit or mature topics that are currently restricted by safety filters.

According to the company, the delay reflects ongoing work to ensure the feature can be introduced responsibly and with appropriate safeguards. The proposed mode would likely require user verification and additional safety measures before it becomes widely available.

The idea behind adult mode is to allow discussions of subjects such as sexuality, relationships, and other mature themes in ways that are educational or informational while still preventing harmful or exploitative uses. However, the feature has proven technically and ethically complex to implement. Developers must find ways to distinguish between legitimate adult conversations and content that violates policies around harassment, exploitation, or misinformation. Source: TechCrunch

Why it matters

As AI systems become more integrated into daily life, companies must continually decide where to draw the line between openness and safety. The debate around “adult mode” highlights a broader challenge: designing AI systems that can handle the full complexity of human conversation while still protecting users from harmful content.

Quick Picks

Meta May Be Preparing Another Round of Layoffs

Meta is reportedly considering additional layoffs as part of its ongoing restructuring around artificial intelligence and efficiency initiatives. The company has been shifting resources toward AI infrastructure and development while trimming costs across other divisions.

As tech giants race to build AI capabilities, resources are increasingly being redirected away from legacy operations toward AI infrastructure and research.
Source: The Verge

Tesla Shareholder Lawsuit Questions Musk’s AI Strategy

A shareholder lawsuit has been filed against Tesla claiming Elon Musk’s AI startup xAI may be diverting resources and strategic focus away from Tesla. The complaint centers on the integration of Tesla technologies with Musk’s broader AI ecosystem.

As AI becomes central to multiple industries, questions about corporate governance and conflicts of interest may become more common across companies building overlapping AI ecosystems. Source: Electrek

Anthropic’s Claude Climbs to No. 1 in the App Store

Anthropic’s AI assistant Claude briefly became the most downloaded free app in Apple’s U.S. App Store, surpassing OpenAI’s ChatGPT. This surge came following heightened attention around the Anthropic's recent dispute with the U.S. Department of Defense.

This indicates that the battle for the top consumer AI assistant is intensifying. Whichever companies win user adoption today may shape how hundreds of millions of people interact with AI in the years ahead. Source: TechCrunch

Startup Raises $190M to Build Autonomous Cybersecurity Agents

Kevin Mandia, founder of cybersecurity firm Mandiant, has raised $190 million for a new startup called Armadin, focused on developing autonomous AI agents that can detect and respond to cyberattacks without human intervention. The company aims to create systems that can monitor networks, identify threats, and automatically neutralize attacks in real time.

As cyber threats become faster and more automated, defense systems may also need to become autonomous—another sign that AI agents are beginning to operate independently across critical infrastructure. Source: TechCrunch

✔ Our tenth and final episode of Singularity Sanctuary's Ethics and Technology course is in production. The final four in the series, however, have been produced with a new and superior camera, which has established a noticeable and stark difference between the first six episodes. In other words, "it's raised the bar."

This means further delay as we reshoot the necessary video to substitute for the old. Fortunately, this does not mean any episodes in the series must be entirely produced from scratch, although this effort will require some extra time.

As Singularity Sanctuary continues to get off the ground, it's important to take what time is necessary to produce high quality work, especially with what will become one of our signature offerings.

Hopefully the anticipation is only increasing your interest. We appreciate your patience and can hardly wait for you to access this timely and important course.

✔ Take note: Our next Singularity Circle will be moved to the second Saturday of April due to the Easter holiday weekend happening April 4th and 5th. This will give you all the time you need to enjoy the holiday with friends and family, without missing or April circle. As always, further announcements, along with the Zoom link, will be sent to our members in due course.

Closing Reflection

The Race Toward Agency — And What It Means

For most of the digital age, computers have been tools. They received instructions. They processed data. They executed programs. But they did not initiate action. That's beginning to change.

A growing number of technology companies are developing systems designed not merely to answer questions, but to perform tasks. These systems can observe information, plan steps toward a goal, and interact with software environments to complete work.

Elon Musk’s recently announced “Digital Optimus” project offers a clear glimpse of this emerging direction. The system reportedly combines xAI’s Grok model with a Tesla-developed AI agent capable of interpreting screen activity and interacting with a computer using keyboard and mouse inputs. In effect, such systems are being designed to operate computers the way humans do, by reading screens, navigating software, and completing workflows.

This shift marks the early stages of what those in the field call agentic AI. Instead of responding to single prompts, these systems can coordinate multi-step actions: gathering information, executing commands, adjusting strategies, and completing tasks. In short, they begin to behave less like calculators and more like collaborators.

This does not mean machines are becoming conscious. But it does mean they are gaining something historically reserved for humans and organizations: operational agency.

The implications of this are enormous. Software agents may soon handle many tasks that currently require human supervision—research, scheduling, logistics, financial analysis, customer support, and even aspects of scientific discovery.

This transition will not happen overnight. At first these agents will operate in narrow domains. Over time their capabilities will expand. As that happens, the real question is not whether AI will gain agency, it's how will humans guide it?

Agency without wisdom is dangerous. But such agency guided by human values may become one of the most powerful tools humanity has ever created. It may mean establishing rules and policies that prevent AI from making final decisions in potentially life threatening or altering decisions.

Whether they are granted agency or not, we cannot hold machines responsible for their actions but must maintain such responsibility for ourselves. The coming years will test whether we are prepared for that responsibility.