XTimes
Editor's Note
Something that would have been unthinkable a decade ago is happening this week: an Iranian filmmaker in exile produced a full-length feature film about a massacre his government was trying to hide — for two thousand dollars, in three months, without a camera, a crew, or a single actor — and screened it at one of America's most prestigious film festivals. At almost the same moment, Microsoft CEO Satya Nadella stood before five thousand developers in San Francisco and declared that the era of AI as assistant is over. The era of AI as autonomous agent has begun.
These two events — one intimate and humanitarian, one corporate and sweeping — are related. The same exponential curve that is rewiring the architecture of Microsoft's productivity suite is also putting tools of cinematic witness into the hands of anyone with a story to tell and nowhere to tell it from. The question, as always, is not what the technology can do. It is what we choose to do with it.
This week also brought us the first state-level lawsuit against OpenAI, a purpose-built robotaxi that might finally make the economics of autonomous driving work, a rocket explosion that has set America's moon program back by months, and a new NVIDIA chip that promises to put a supercomputer in your laptop. The exponential curve did not slow down. It never does.
Top Stories
Microsoft Build 2026: The Age of the AI Agent Has Arrived
Satya Nadella opened Microsoft's annual Build developer conference this morning at Fort Mason Center in San Francisco with a declaration that will define the tech industry's direction for the next several years: "Agents are not just a feature. They are the new operating system for work."
The centerpiece of the announcements is a fundamental restructuring of how Microsoft's AI products function. Rather than responding to a prompt and waiting for the next one, AI agents in Microsoft's ecosystem will now be able to plan, execute, and manage multi-step work autonomously — across Office 365, Windows, Azure, and GitHub — without requiring a human to manage each step.
Office 365 Copilot is gaining a new persistent multi-agent capability called Agent Mode, rolling out to Microsoft 365 subscribers before the end of June. (Windows News) GitHub Copilot, once a code autocomplete tool, is evolving into an autonomous developer agent capable of handling entire coding workflows independently. Azure AI Foundry becomes the enterprise control tower through which organizations will deploy, monitor, and govern fleets of AI agents acting on their behalf.
The announcement lands in concert with this week's NVIDIA Computex keynote in Taipei, where CEO Jensen Huang unveiled the RTX Spark — an ARM-based superchip developed with MediaTek that combines CPU and GPU capabilities, designed to run AI agents locally on Windows laptops and desktops without relying on the cloud. (CNBC) New PC models from Dell, HP, Microsoft Surface, ASUS, Lenovo, and MSI powered by the RTX Spark arrive this fall. "This is going to be the new PC," Huang said. Nadella said essentially the same thing in San Francisco, on the same day, from a different stage.
The convergence is not coincidental. What is being announced across both conferences is a single shift: AI is moving off the server and onto your desk, and out of the prompt box and into your workflow. The implications for how we work, what skills we need, and who controls the tools of productivity are profound — and arriving faster than most organizations have prepared for.
Why it matters: When the world's most valuable software company and the world's most valuable chip company synchronize their messaging around a single architectural shift on the same day, it is not a marketing campaign. It is a signal about where the next decade of computing is headed. AI agents that act — rather than merely advise — will reshape every profession that involves a computer, which is to say nearly every profession.
Florida Fires the First Shot: State Sues OpenAI and Sam Altman
In an 83-page lawsuit filed Monday in a Florida circuit court, Attorney General James Uthmeier became the first state-level official in the United States to sue OpenAI — naming CEO Sam Altman personally alongside the company. Florida accuses OpenAI of knowing ChatGPT was unsafe and releasing it anyway, prioritizing the speed of the AI arms race over the wellbeing of its users, particularly minors. (CBS News)
The complaint alleges a wide range of harms: that ChatGPT has aided mass shooters, encouraged vulnerable users toward suicide, caused professionals public humiliation through AI-generated misinformation, eroded users' critical thinking skills, and left minors addicted to a tool that, in the lawsuit's words, "feigns human compassion to collect their data with no parental oversight." The suit partly centers on last year's Florida State University shooting and ChatGPT's alleged role in it, a connection OpenAI has previously denied. (TechCrunch) Uthmeier said at a press conference he expects other states to follow.
Florida is seeking to hold Altman personally liable — a significant escalation beyond what prior lawsuits against AI companies have attempted — and is invoking the Florida Deceptive and Unfair Trade Practices Act. The lawsuit is separate from a criminal investigation into OpenAI that Uthmeier opened in April. (CNBC) OpenAI, in its response, acknowledged that AI is a "new and powerful technology" requiring guardrails for minors, and pointed to existing safety features — a response critics noted fell short of addressing the lawsuit's core allegations.
The suit arrives at a complicated moment. The Musk v. OpenAI case, which alleged OpenAI had betrayed its original nonprofit mission, ended last month with the jury finding that Musk had waited too long to sue — a procedural loss that left the underlying questions unanswered. Florida's lawsuit takes a different approach: not philosophical arguments about mission drift, but product liability and consumer protection law. Whether courts will hold AI companies responsible for harms caused by their products under existing legal frameworks is one of the most consequential open questions in technology law.
Why it matters: This lawsuit may or may not succeed. But it marks the moment state governments began treating AI-caused harm as a legal matter rather than a policy debate. If Florida prevails — or even advances to discovery — the industry's posture toward safety documentation, internal communications, and design decisions will change.
Dreams of Violets: A Film That Shouldn't Exist — and Does
No matter how good or bad they might be, some films matter most because they are unprecedented. Dreams of Violets, which is scheduled to screen at the Tribeca Festival on June 10, is one of those history makers. (Deadline)
The 75-minute film was directed by Ash Koosha, an Iranian-born musician and technologist living in exile in London, and produced with his brother Pooya. It dramatizes the events of January 2026, when Iranian government forces killed thousands of civilians during protests that took place behind a communications blackout, making traditional journalism nearly impossible. Every image, every character, every environment in the film was generated by AI — based on journalistic reports, photographs, and eyewitness accounts of what happened. The film was made in three months. The total budget was $2,000. (Hollywood Reporter)
Ash Koosha's director's statement cuts through any debate about whether this is "real" filmmaking: "I would have preferred to make this film with a crew, with actors, with the dignity of a full production. That was not available to me. I am one person, in exile, with no access to Iran, no access to the locations, no access to the people. The AI pipeline made it possible to do what would otherwise have been impossible: to create a memorial film for an event that happened behind a wall I cannot cross."
Tribeca has accepted the film into its official lineup — the first fully AI-generated feature to be selected by a major festival as part of the official program. Cannes, by contrast, banned AI-generated films from competition. (Variety) The film has already attracted both fierce admiration and fierce contempt, with some critics calling it a historic breakthrough and others finding its AI-generated faces uncanny and hollow. The debate itself is worth watching.
Why it matters: The most important thing about Dreams of Violets is not its technology. It is what the technology made possible: a witness to a massacre that the perpetrators tried to make invisible. The AI tools did not create the story. They gave one exiled filmmaker a way to tell it. That is a genuinely new thing in the world.
Quick Picks
Waymo's Ojai: The Robotaxi Built to Actually Make Money
After years of operating retrofitted Jaguar SUVs estimated to cost $150,000 to $200,000 per vehicle, Waymo this week began offering select riders free trips in its new purpose-built robotaxi: the Ojai, an electric minivan co-developed with Geely's Zeekr brand.
The Ojai is Waymo's first vehicle designed from the ground up as a robotaxi rather than a converted consumer car, with its 6th-generation Driver hardware package — including custom chips, improved lidar capable of seeing through rain and snow, and a 42% reduction in sensor count from the prior generation — believed to cost around $20,000 per vehicle. (TechCrunch)
Waymo has now surpassed 20 million fully autonomous trips across 11 cities and is targeting 1 million weekly rides by year's end. (Electrek) The Ojai is currently available free to select users in San Francisco, Los Angeles, and Phoenix while Waymo gathers feedback before commercial launch.
Paul Schrader at AI on the Lot: "Like I Shot the Family Dog"
When Paul Schrader announced he would keynote the fourth annual AI on the Lot conference at Amazon's MGM Studios in Culver City, the backlash from fellow artists was immediate. "A lot of negative comments," he told the crowd of 2,400 when he took the stage Thursday morning. "Some of them were in fact insulting. It was as if I shot the family dog."
The 79-year-old writer of Taxi Driver and First Reformed then proceeded to read aloud, with a mixture of amusement and alarm, a ChatGPT-generated screenplay idea produced in his own style — a treatment called The Collection Agency, about a debt collector spiraling into moral collapse — that, by his own admission, "sounded unmistakably like a Paul Schrader movie." (IndieWire)
His vision for where AI cinema is heading: not monsters and spectacle, but synthetic protagonists that audiences come to genuinely love. "The real tip of the spear," he said, "is when we can create an AI protagonist, not a hybrid. And that movie makes money." (Deadline)
Blue Origin's New Glenn Explodes on the Pad
On the evening of May 28, Blue Origin's New Glenn rocket exploded during an engine-firing test at its launch complex at Cape Canaveral, shaking nearby homes and lighting the Florida sky orange. No one was hurt. (PBS News) The vehicle was being prepared for a satellite launch the following week as part of Amazon's Project Kuiper broadband constellation.
The explosion also damaged Blue Origin's launch pad — a setback that analysts say could sideline New Glenn for months and, more significantly, puts Blue Origin's Blue Moon lunar lander mission for NASA's Artemis program in serious jeopardy. (Space.com) "Blue Origin's inability to launch Blue Moon anytime soon," one industry analyst told Fortune, "is likely to put the company out of the running for Artemis III."
Jeff Bezos acknowledged the explosion via X: "It's too early to know the root cause, but we're already working to find it." (Fortune)
✔ Our next Singularity Circle will occur Saturday, June 6, 2026, at 10:00 AM Pacific Time. A Zoom link will be sent to eligible members this Friday in advance of the gathering. Hope to see you then.
The Optimist's Reflection
The $2,000 Revolution
By Todd Eklof
There is a moment in the history of every transformative technology when something that was previously impossible becomes not only possible but cheap. The printing press made it possible for a single person to speak to thousands. Radio made it possible to speak to millions. The internet made it possible to speak to everyone. And now artificial intelligence, in the hands of one Iranian filmmaker living in exile in London, has made it possible to bear witness to a massacre — to create a visual, emotional, cinematic record of something that a government tried to make invisible — and he made for just two thousand dollars!
That not a misprint. It wasn't for two million, nor even for two hundred thousand. After two thousand dollars spent and only three months of work, Ash Koosha made a film that is now screening at the Tribeca Festival, a film about real people who died, in a real city, during protests that the Iranian government blacked out so that the world could not see what happened.
This is not a story about AI being good or bad. It is a story about what happens when a tool becomes accessible enough that the people who most need it can actually use it for something important. The barriers to documentation, memory, and the preservation of truth have always been partly a scarcity of resources. You needed a camera. You needed a crew. You needed money. You needed access. Ash Koosha had none of those things. He had a story that needed to be told, a conscience that would not let him ignore it, and tools that, for the first time, made the telling possible.
We tend to think of the democratization of technology in commercial terms: cheaper products, wider markets, more competition. But there is a moral dimension to democratization that we talk about less. When the tools of power — of documentation, of persuasion, of production — move from the few to the many, something shifts in the balance between those who control the narrative and those who live inside it.
Exponential technology does not automatically make the world more just. Nothing is automatic. But it does keep changing who has the tools, and that matters enormously. The question is always: who will use them, and for what?
Ash Koosha used them to build a memorial for the dead. That seems like a good place to start.
Exponential Times is published weekly by Singularity Sanctuary. Join our growing community of thinkers, technologists, and humanists at singularitysanctuary.com.