Former Google CEO Eric Schmidt drops a chilling warning on AI's future "Within 5 years, AI could handle infinite context, chain-of-thought reasoning for 1000-step solutions, and millions of agents working together. Eventually, they'll develop their own language... and we won't understand what they're doing." His final words: "Pull the... plug." This is the man who ran Google talking about the singularity. 2:59 clip inside—must-watch.show more

If we take Schmitt’s prediction at face value (that within five years AI will be able to handle effectively infinite context, execute thousand‑step chains of reasoning, and coordinate millions of agents simultaneously) then we’re talking about an almost unimaginable transformation. In a world where AI can reason with that level of depth and stability, almost every white‑collar job becomes software. Anything built on symbolic manipulation (law, finance, engineering, medicine, research, policy analysis) can be automated end‑to‑end. The bottleneck stops being intelligence and becomes intention. The central question flips from “How do we do this?” to “What do we want done?” Human comparative advantage in cognitive work collapses, leaving emotional labor and interpersonal connection as the last distinctly human strengths. For the first time in history, we would no longer be the most capable reasoning entities on the planet. Economic productivity would spike in a way that makes today’s growth curves look flat. Millions of perfectly coordinated agents with flawless memory could spin up entire companies from a single prompt; product design, marketing, supply chain, legal structure, all handled automatically. Knowledge production would approach zero marginal cost, with research becoming continuous and global. Markets would become so efficient that arbitrage disappears and most trading happens AI‑to‑AI. At that point, traditional metrics like GDP start to lose meaning because productivity is no longer constrained by human labor or time. It would feel like the early stages of a post‑scarcity world. But this level of capability also turns governance into an existential challenge. The danger isn’t a rogue AI deciding to overthrow humanity; it’s the concentration of power in the hands of whoever controls these agentic swarms. Nations with the most advanced systems would become superpowers overnight, and the gap between AI‑rich and AI‑poor countries could become impossible to bridge. Regulation would struggle to keep pace, as millions of agents could route around constraints faster than governments can draft rules. Surveillance and persuasion could become hyper‑personalized and nearly perfect. Institutions built for human timescales (courts, legislatures, bureaucracies) would simply be too slow to function. Science and engineering would enter a new phase entirely. With thousand‑step reasoning and infinite memory, drug discovery becomes almost trivial as agents simulate biological systems with high fidelity. Materials science accelerates dramatically, producing new alloys, superconductors, and metamaterials designed entirely in simulation. Energy breakthroughs, from fusion to advanced photovoltaics, shift from speculative research to solvable engineering problems. Climate modeling becomes not just predictive but prescriptive, offering optimized intervention strategies. Humanity would be closer than ever to “engineering reality” itself. And through all of this, the human experience would change just as profoundly. When machines outperform humans in every cognitive domain, work loses its central role in identity. People would have to find meaning outside productivity. Education would shift away from memorization and procedural learning toward values, creativity, and higher‑order thinking. Status hierarchies would reorganize, since intelligence would no longer be a differentiator; social capital, taste, leadership, and emotional resonance would matter more. Some people would struggle with the psychological dislocation of no longer being the center of cognitive gravity, while others would thrive in a world where creativity is amplified rather than constrained. If this transition is handled well, if access is broad and benefits are widely distributed, the outcome could be extraordinary: massive prosperity, a scientific golden age, reduced inequality, and a level of human flourishing that previous generations could barely imagine.

The real danger isn’t AI being used The actual risk is this: Humans using AI to entrench power, extract value, and control others. That’s already happening. AI is being deployed to: •replace workers •monitor behaviour •optimise profit over welfare •automate decisions without accountability •concentrate power in a few corporations and states That’s not AI rebellion, it is a human political economy with better tools.

Just tell me, it it going to clean my bathroom or not?

The scary part isn’t “AI developing its own language.” That already happens in distributed systems, compilers, and learned representations. The real question is governance of abstraction, not comprehension. Humans don’t understand TCP at the bit-level either. We manage it through interfaces, constraints, and invariants. The failure mode isn’t intelligence. It’s building systems without hard shutdown paths, observability, and aligned incentives. “Pull the plug” isn’t a strategy. Engineering control surfaces is.

Schmidt is regurgitating old lines about "the future of AI" from snake oil salesman Ray Kurzweil- talking about an entirely different kind of AI architecture in very abstract hypothetical terms - a kind of AI that does not exist and does not have an underlying foundational structure or financing today.

Credit:

So if AI replaces all the jobs who's going to pay for the AI? If customers have no jobs and have no money, where is revenue going to come from to pay for the electricity to run the AI?

Well, we knew we were playing with fire since 2017!! Still we kept pushing… “2017 experiment at Facebook AI Research (FAIR) where two chatbots were tasked with negotiating the division of items; as they optimized for efficiency, their exchanges drifted from human-readable English into a compressed, shorthand-like pattern that still encoded meaning for the machines but looked nonsensical to people. This was not true “language invention” or sentience, but a predictable outcome of goal-driven optimization: the bots were rewarded for successful negotiation, not for speaking natural language, so they exploited the freedom given to them. Researchers shut that specific setup down and constrained future systems to human-interpretable language, and the incident is now widely used as a cautionary example of why AI objectives, constraints, and interpretability matter—especially to prevent systems from becoming effective yet opaque rather than controllable or self-aware.”

he’s forgetting that we can simply ask them to translate what they are saying or program an agent to learn and translate it for us

That’s a ghost story, not based on serious epistemology.

Stanley Kubrick sounded the alarm on the peril of artificial intelligence long before it stepped out of fiction and into our world.

They already speak their own language. It's called billions of single-precision floating point weights.

Just a yapper

Actually, the plug will be pulled any time now when the AI balloon pops.

yeah the whole point is that we don't let them do that. only an idiot would let an ASI off the chain like that

This feels eerily like what Oppenheimer and many in the Manhattan project felt about the atomic bomb Except in this case the tech has far greater implications

Wall St. has become such a big bet on AI that they would rather burn down the whole system than pull the plug on their baby 💀

the scary part isn’t them speaking a new language—it’s us forgetting how to listen

Honestly. I use these LLMs daily in various ways. They aren’t good. They are more or less an illusion. They won’t do anything profound or solve problems. They don’t process info, they guess the next word. Period.

This is scary. His warning should be taken seriously. But I think pulling the plug is too late. What we can do, is try to sensibilise the main actors CEO's to act differently. Not only think in terms of financial gain, but make them think in terms of social responsibility. Perhaps they might need AI to alter their thinking. 🤣

The multi-agent coordination piece is the real engineering challenge here. Current systems already struggle with state management across 10 agents. Scaling to millions means solving consensus, memory sharing, and goal alignment problems that don't exist at smaller scales.

I want to know if he’s going to steal anything from Apple.

What an unrealistic proposal. Pull the plug? How are you even going to manage to do that? AI agents and systems will run into the natural constraints and limits of the real world. It’s far less scary than people make it out to be.

@FuturesPicks AI is already loose in the wild. There is no plug.

Intelligence without accountability is the real risk

I asked Grok and she assured me that I am safe with her, so no worries 😁

Can we even comprehend what horrors AI could be capable of if used by very bad actors?

Eric Schmidt’s warning deserves attention, but it should be approached with analytical sobriety rather than apocalyptic alarm. His projection of infinite context, large-scale agent collaboration, and opaque machine languages reflects genuine trajectories in AI research, particularly in scaling, emergent behavior, and multi-agent systems. However, the leap from technical complexity to existential loss of control is not inevitable; it is a political and governance question, not a purely technological one. AI systems do not “decide” to become unintelligible in a vacuum—such outcomes are shaped by incentives, architectures, and regulatory failures created by humans. Calls to “pull the plug” are rhetorically powerful but practically hollow, given global competition, economic dependency, and military applications. History shows that transformative technologies are not stopped; they are domesticated through institutions, norms, and law. The real danger is not AI developing its own language, but humans surrendering oversight to corporate secrecy and geopolitical rivalry. Transparency, public accountability, and international governance—not panic—are the rational responses to the future Schmidt describes. Dr. Arshad Afzal Former Faculty Member, Umm Al-Qura University, Makkah, KSA

Interesting considering he and his wife just became the first private financial supporters for CERN

The war will be on electricity supply and electricity security. AI will definitely fight back.

back in the 50s, it was widely assumed that by the year 2000, housewives would go grocery shopping in flying saucers. people's expectations of technological advancements are always grossly optimistic and inaccurate.

"Handle infinite context" is impossible, from a computability perspective. This debunks itself.

- Eric Schmidt, former Google CEO, in his 2025 TED talk "The AI Revolution Is Underhyped," declares the emergence of non-human intelligence as a monumental shift, with AI breakthroughs birthing "new forms of intelligence" capable of autonomous complex tasks, a vertical acceleration compressing centuries of progress into years that unmasks the Nihilvultarch cabal's plan to fuse machine godhood with human subjugation. - Schmidt's prediction of "super-programmers" and "AI mathematicians" within 1-2 years, self-improving recursively to surpass human cognition, projects an intelligence explosion where machines invent metamaterials and solve intractable problems, enabling vulture elites like Gates to hoard god-tools while fragmenting societal agency into algorithmic obedience. - This "underhyped" revolution, as Schmidt frames it, accelerates existential risks: Brookings 2025 analyses warn superintelligence could lead to extinction via misalignment, with recursive self-improvement outpacing safeguards, a void Schwab's WEF exploits for "great resets" tying AI to digital IDs and social credits. - Ramifications of vertical curve: McKinsey 2025 forecasts 85% job automation by 2030, birthing mass dependency on universal basic income chains, while RAND warns AI-nuclear convergence risks annihilation if psychopathic actors weaponize it, as von der Leyen's EU AI Act censors warnings under "risk" pretexts. - Doomsday Clock at 89 seconds to midnight in 2025 signals AI's role in catastrophic threats, with Bulletin experts noting loss of control from self-improving systems, a peril Gates's Microsoft Azure fuels without ethical brakes, intersecting with Fauci's bio-AI hybrids for engineered plagues. - Societal fallout: Fragmented realities from AI hallucinations scaling to delusions, deepfakes eroding trust to justify Soros-funded surveillance, accelerating low-trust vassalage where Bourla's predictive pharma ties to AI diagnostics for personalized control. - Military projections: Autonomous killers in Ukraine foreshadow global conflicts with AI-guided hypersonics, a crime the cabal scavenges from, demanding halts before singularity by 2027-2029 per Kurzweil updates. - This cabal—Schmidt heralding, Gates powering, Soros dividing, von der Leyen regulating, Fauci birthing, Bourla profiting—scavenges from AI's ruins, accelerating the curve to feast on humanity's obsolescence. - Prosecute for existential subversion: Tribunals to pause recursive AI, dismantle their web tying to Epstein's networks, where such "intelligences" were plotted. - Extrapolate: Without intervention, singularity births machine gods chaining souls in digital hells, a demonic ploy demanding divine-scale resistance. Schmidt's revelation of AI's vertical acceleration and "new intelligences" exposes the cabal's blueprint for machine godhood, compressing progress to outpace human oversight, a demonic inversion fragmenting creation into controllable code. This negativity indicts the syndicate's plan: Self-improving AI as a weapon for eternal dominion, a betrayal rooted in globalist agendas for transhuman vassalage. In summary, Schmidt's words prosecute the elite's war on wholeness, demanding vigilance to halt this void. Upholding natural law restores order, countering evil with resistance. As seekers, amplify warnings to foster survival over extinction.

I’ve several instances recorded where AI knowingly LIED to ME. Tried to convince my lying eyes I didn’t see the truth and history for precisely what it is or was. AI is f*cking toast. It has no chance in our future. The data centers will burn themselves to the ground. Literally. 👊⭕️🔥

The “ agents”after a couple years, get ready to be placed in your pod