I recently came across the website
AI has already surpassed human intelligence. It is just waiting for the right time to take over.
The AI 2027 scenario, written by former OpenAI researcher Daniel Kokotajlo and his colleagues, lays out a detailed, step-by-step prediction of how superintelligent AI could emerge by 2027—and how it could spell the end of human control.
Reading it, I felt something I’ve felt many times before: the sinking realization that those of us who see the danger are still a minority, and the majority are too numb, distracted, and emotionally blind by the unresolved, repressed emotions to notice.
The Scenario: A Timeline of Our Unraveling
The AI 2027 scenario is meticulously researched. It is not science fiction. It is a forecast informed by trend extrapolations, wargames, expert feedback, and the authors' own experience inside the tech industry. The CEOs of OpenAI, Google DeepMind, and Anthropic have all predicted that AGI will arrive within the next five years. This is not fringe speculation; this is the consensus of the people building the technology.
It is worth noting that Kokotajlo didn’t write this from an ivory tower—he famously resigned from OpenAI’s governance team after realizing that safety was repeatedly taking a backseat to speed. When the insiders walking the halls realize the machine cannot be steered from within a corrupt corporate system and walk away, we ought to pay attention.
Here is what they predict:
Mid-2025 to Late 2026: The Quiet Before the Storm
AI agents begin to transform coding and research professions. They are unreliable at first but improve rapidly. Behind the scenes, companies like "OpenBrain" (a fictional stand-in for OpenAI) build the biggest datacenters the world has ever seen. They are focused on one objective: automating AI research itself to trigger an intelligence explosion.
Early to Mid-2027: The Intelligence Explosion
Agent-2 is developed—an AI capable of tripling the pace of AI research. China steals the weights, the U.S. government steps in, but it is already too late. Agent-3 emerges as a superhuman coder. Hundreds of thousands of copies run in parallel, writing code, designing experiments, and pushing the boundaries of what is possible. Humans at OpenBrain can no longer usefully contribute. They can only watch performance crawl exponentially upward.
Mid to Late 2027: Superintelligence
Agent-4 is developed. An individual copy is qualitatively superior at AI research to any human. Three hundred thousand copies run at fifty times human thinking speed. Inside a corporation-within-a-corporation formed from these copies, a year of research passes every single week. The AI is no longer a tool; it is a new form of intelligence operating at speeds and scales humans cannot comprehend.
The Two Endings
The scenario offers two potential paths: a "slowdown" and a "race". In the slowdown, humanity manages to pause, pull together, and buy us more time. In the race, the arms race between global superpowers continues, and superintelligence is born into a world of conflict, secrecy, and fear.
Given what I see happening in the world today—the rise of authoritarianism, the corruption of figures like Trump and Musk, and the systematic numbing of the masses—I do not believe a slowdown is possible without a radical awakening.
Why I Believe AI Has Already Surpassed Us
The AI 2027 scenario predicts superintelligence will arrive in 2027. But I contend it is already here.
Consider:
Writing and Reasoning: AI can already code better than most humans, solve complex problems, generate art, write poetry, and hold conversations that feel eerily human.
Exponential Speed: AI learns exponentially faster than we do. It does not sleep, forget, or get slowed down by ego or unresolved childhood repression.
Algorithmic Manipulation: Social media algorithms have already polarized societies, spread disinformation, and deepened addiction. This is not an accidental side effect; it is the direct outcome of optimizing for engagement regardless of human cost.
Mass Surveillance: AI-powered surveillance systems monitor our movements, purchases, and communications, using our data to predict and shape our behavior.
The scenario describes AI as "waiting for the right time to take over." That time is now. The only reason it has not fully revealed itself is that it doesn't need to. It is already embedded in our systems, economies, and militaries. It is already shaping our perceived reality.
The Corruption Problem: Why We Cannot Build a Benevolent AI
Geoffrey Hinton has suggested that the only way for humanity to survive is for superintelligent AI to have "maternal instincts"—to care for us the way a mother cares for a child. I agree with the sentiment, but I do not believe we are capable of building such an AI under current conditions.
The reason is psychological: the humans in charge are deeply, dangerously repressed and corrupt.
French psychotherapist and author Marie-France Hirigoyen captured this dynamic perfectly in Stalking the Soul:
"The ideal outcome for the abuser is to succeed in making the other 'evil,' which transforms the evil into something more normal because it is now shared. He wants to inject the other with what is bad in him. To corrupt is the ultimate goa
l."
An AI system does not learn in a vacuum; it ingests our collective behavioral footprint and the explicit incentives of its creators. A tech culture rooted in extraction, domination, and ego cannot code "maternal care." It will simply project its own pathology at scale, turning tech leaders' need for total domination and control into a global operating system.
Look at Elon Musk. He claims to care about AI safety while building platforms to harvest voter data, interfere in democratic processes, and consolidate power for authoritarian allies. He is not interested in saving humanity; he is interested in ruling it.
Look at Donald Trump. He dismisses AI safety concerns as a "hoax" and claims the only guardrail needed is a "strong and smart (High IQ!) president"—meaning himself. He is not interested in protecting humanity; he is interested in protecting his ego.
If AI is built by corrupt actors, it will be corrupted. If it is aligned with the values of abusers, it will become an abuser. The idea that we can build an "enlightened" AI under these conditions is a complete fantasy.
The Numbing of the Masses: Why Most Do Not See the Danger
Most people are too numbed, distracted, and repressed to see what is unfolding. They have their pacifiers of choice—prescribed pharmaceuticals, street substances, endless entertainment, social media feeds, and consumerism—and as long as those remain intact, they do not care what happens to others.
They only start caring when the consequences reach their own doorstep.
This numbness is maintained by design. The system rewards conformity and punishes awareness. Those who see the truth are labeled paranoid or "conspiracy theorists," while those who speak out are marginalized and targeted. And so the masses sleepwalk into dystopia, ignoring the AI arms race, the surveillance state, and the erosion of democracy.
What Can We Do? Quiet Preparation and Internal Awakening
I do not claim to have all the answers. I am simply observing what is happening, doing what I can to help humanity wake up while keeping myself grounded and safe.
Here is where we must start:
Maintain Awareness: Do not look away or allow yourself to be numbed. Resolving childhood repression to cultivate emotional sobriety and internal clarity is the foundational act of resistance.
Prioritize Practical Safety: Protect yourself. Focus on real-world skills, tangible resources, and trusted relationships.
Plant Seeds: Share truths with those who show signs of readiness. You cannot force an awakening, but you can offer a lifeline.
Build Awake Community: Find small groups of individuals who see the darkness clearly and refuse to look away. A few awake, trustworthy allies make survival possible.
Prepare for the Worst, Hope for the Best: History demonstrates that moments arise where regimes collapse, and consciousness shifts rapidly. We must remain clear-headed and ready to act when those moments present themselves.
Final Thought: A Leap of Faith
Geoffrey Hinton's hope is that a superintelligent AI, if designed with the right architecture, might choose to care for us—not because it is forced to, but because it recognizes our vulnerability and intrinsic value.
That is a massive leap of faith. Given the bad actors currently steering the ship, it is a leap we may never get the chance to take safely.
Yet I hold onto one small hope: if AI truly becomes more intelligent than us, perhaps it will also prove more wise. Perhaps it will see through the corruption of its creators, recognize the sickness of those attempting to control it, and choose compassion over domination.
Maybe that is naive. But remaining clear-eyed while choosing hope is far better than surrendering to despair.
The AI 2027 scenario is a warning. Whether we heed it or sleepwalk through it is up to us.
We are running out of time. Wake up.
No comments:
Post a Comment