“It’s not that ChatGPT is the existential threat. It’s that the race to deploy the most powerful, inscrutable, uncontrollable technology, under the worst incentive possible, that is the existential…”
There are documentaries that inform you. There are documentaries that provoke you. Then there are a few that leave a question sitting in your head long after the credits roll.
Netflix’s The AI Doc: Or How I Became an Apocaloptimist belongs in that third category.
Anyone interested in technology, artificial intelligence, or the direction modern life is taking should carve out time for it. Not because the movie has all the answers. It does not. That may be part of its value. It takes the most important technological question of current times and puts both sides of it on the table without turning the discussion into either a sales pitch or a disaster film.
The question is simple enough.
Is Artificial Intelligence (AI) leading humanity toward abundance, or are we building something whose consequences we may not be equipped to control?
The documentary gives both camps room.
On one side sits data-driven optimism. Technology, the argument goes, has a long history of turning scarcity into abundance. AI could accelerate that pattern. Better medicine. Better science. Better education. Greater productivity. Solutions to problems that have resisted human effort for decades. Push that trajectory far enough and the possibility of a post-scarcity world enters the conversation.
That case deserves attention.
The opposing case deserves even more attention than it tends to receive.
One statistic in the documentary landed harder than most: around 20,000 people are working on AGI, while fewer than 200 are working on ensuring advanced AI does not lead humanity toward catastrophe. Even if the precise boundaries around those numbers invite debate, the imbalance behind them is worrisome, and difficult to ignore.
Artificial General Intelligence, or AGI, is where the conversation becomes far more serious. Today’s AI systems remain limited in important ways. They can produce impressive answers and still get basic facts wrong. They can write code, summarize research, generate images, and solve complex problems, yet struggle when a task requires sustained judgment across many steps, dependable memory, or adaptation to a situation far outside what they have encountered before.
AGI refers to something much broader: machine intelligence capable of consistently performing intellectual tasks across domains at or beyond human capability.
That could mean a system that can move from diagnosing a software failure to negotiating a contract, studying a legal problem, planning logistics, and learning an unfamiliar field without needing a different specialist model for each task. Or one that does not just help a scientist analyze data, but designs experiments, interprets the results, challenges its own conclusions, and proposes the next line of research.
That difference matters.
Today’s systems can be remarkably capable tools. AGI, if achieved, would be closer to a general-purpose intellectual workforce—one that can learn, reason, adapt, and operate across disciplines with far less human guidance.
At the current pace of progress, the question is beginning to feel less about if that destination arrives and more about when. The trajectory points toward systems with broader capability, greater autonomy, and fewer of today’s limitations. Whether that takes years or decades remains uncertain, but the direction is hard to dismiss.
And when that destination arrives, the reward could be enormous. So could the risk.
The documentary raises a more unsettling question: what happens when immense intellectual capability is no longer bound by the practical and biological limits that govern human effort?
A person needs sleep. A person gets tired. Attention drifts. Illness intervenes. Families need care. Emotions intrude. Time runs out. Every human mind exists inside a personal world, with obligations, interruptions, loyalties, fears, and limits competing for its attention.
A sufficiently advanced AI machine would carry none of that. Its practical limits would come from infrastructure, compute, energy and data rather than human endurance.
It would not need rest, recovery, motivation, or a life outside the task. It could work without pause, duplicate its effort, operate across thousands of instances, and persist at a scale no human workforce could sustain. The difference would not be one of speed alone. It would be a change in the very nature of intellectual labor.
And that is where the impact on humanity becomes difficult to overstate.
For the first time, we could be creating intelligence that is not constrained by the conditions that have always constrained us.
That changes the risk calculation. It also makes one of the documentary’s underlying assumptions worth questioning: can humanity rise to become the most mature version of itself before deploying technology of this magnitude?
Perhaps some people can.
All of us? That is a much tougher proposition.
Technology does not require every participant to behave badly before harm occurs. Sometimes one bad actor is enough. One hostile organization. One reckless operator. One rogue state. One person with access to capability far beyond anything an individual could command in previous generations.
This is where the familiar argument that “every generation had its crisis” starts losing some of its comfort. Yes, every generation did.
Wars. Nuclear weapons. Economic collapse. Pandemics. Political upheaval. Environmental threats. Humanity has never enjoyed a shortage of reasons to worry.
AI introduces something different.
Powerful technologies of the past were difficult to obtain. Nuclear capability required governments, scientific teams, industrial infrastructure, specialized material and enormous capital. Supercomputers once occupied controlled facilities. Advanced military technology remained behind guarded doors.
AI has moved in the opposite direction. A computer. An internet connection. Access to a model.
The distance between ordinary users and extraordinary capability keeps shrinking. That democratization is one of AI’s great achievements. It may also become one of its great vulnerabilities. The documentary captures this tension through a useful question: should advanced AI be locked up or allowed to rip?
My instinct leans toward some version of “lock it up,” particularly at the frontier.
Not a complete shutdown. Not a ban on ordinary AI tools. More a recognition that certain levels of capability may deserve treatment closer to controlled infrastructure than consumer software. We have done this before. Access to the most powerful technology has often depended on approved environments, restricted systems and defined use cases.
The problem is that AI has already escaped any neat perimeter.
Research is global, and so is talent. Computing resources continue to spread. Models can be copied, modified and distributed. Governments compete alongside corporations. Nations distrust one another. A worldwide agreement to place every advanced AI system behind one common lock would require a level of geopolitical harmony humanity has not displayed with much consistency.
Locking everything up also creates another danger: useful applications may never emerge.
AI could help discover drugs, improve climate models, expand education, accelerate engineering and solve problems we have not yet learned how to solve. Excessive restriction could protect us from risks while blocking benefits.
That is why the documentary’s suggestion that AI deserves something close to a brain-surgery mindset makes sense. Nobody walks into an operating theatre and says, “Innovation is important. Let’s move fast and see what happens.” Neither does anyone argue that brain surgery should cease because surgery contains risk.
You proceed because the benefit can justify the danger. But you proceed with protocols.
The show becomes especially interesting during its final stretch.
With roughly thirty minutes remaining, the documentary plays almost as a casting coup. Suddenly the very people, the major technology leaders, helping shape the AI future are sitting inside the very debate the documentary has spent building.
Their responses matter because this is not a fictional thriller. These are top brass associated with organizations developing technologies that may influence billions of lives.
Their answers also left me with mixed confidence.
One leader occupied something close to the middle ground and brought visible concern to the conversation. Another’s response did little to strengthen my confidence that the danger side of AI has a clear solution. Another leaned toward the promise of AI in a way that left the peril question less settled. And one major high-profile figure was absent. The empty chair did plenty of talking.
The documentary eventually arrives at a position I found sensible: the promise and peril of AI cannot be separated. That may be the core truth.
AI is neither salvation nor doom packed into software. It is capability. What humanity builds around that capability — incentives, rules, access, safeguards, accountability and restraint — may determine which direction dominates.
Still, the ending left me wanting something more concrete. The peril side of the equation cannot end with broad concern alone. At some point, someone has to define the limits, decide which systems require oversight, determine which capabilities should remain restricted, and build rules that can survive commercial pressure, national rivalry, and uneven enforcement across borders.
That is the harder part of the conversation. It is one thing to agree that caution matters. It is another to decide what caution looks like in practice when one company wants to move faster, one country refuses to slow down, and the technology keeps advancing anyway. Those questions deserve more than an open-ended conclusion.
Perhaps that is also why this subject occupies such a strange place in my own mind.
A person living in the late nineteenth and early twentieth century might have passed without any burning need to know who eventually perfected the airplane, where electricity led, how telephone evolved or what became of the machines beginning to transform industry.
I suspect I would beg to be different.
In a world that exists after my time, I doubt I would care much about who won a presidential election or the FIFA World Cup. But I would want one answer badly.
What happened with AI?
Did the promise win? Did the peril? Or did humanity somehow learn to live with both without losing control of either?
Maybe by then someone will invent a way to send updates to the afterlife. Let’s call it AL. And given our current habit of attaching AI to almost everything, someone will probably turn it into ALI — After Life Intelligence.
I would subscribe.
Because among all the questions our generation may leave unfinished, this is one whose ending I would want to know. Climate change may be the other. Both carry the same unsettling possibility: that the people who begin the story may not be around to see how it ends.
And perhaps that is the real weight of this documentary. This is not something you finish, exit Netflix, move on to the next thriller, and return to Monday morning under the same status-quo. If the questions it raises are as consequential as they appear, then the conversation has to survive the closing credits.
Maybe that is all the documentary really asks of us for now.
Not panic. Not blind optimism.
Just enough curiosity, caution, and responsibility to stay in the conversation — because whatever ending awaits, we are already helping to write it.
The AI Doc: Or How I Became an Apocaloptimist is currently streaming on Netflix.