Elon Musk, the most successful entrepreneur in history, said that in just 3 years AI will be more intelligent than all humans combined. Bill Gates said AI is the greatest technology of his lifetime. And the man who is, unfortunately, the most powerful person in the world (Trump), said AI is "the most dangerous thing out there," "so scary," and "bigger than anything else." At the CES computer trade show and at WEF 2026, AI was the single most important topic.

For thousands of years, advanced intelligence on Earth was exclusively human. Then we created new intelligent beings—living inside machines. We are no longer alone.

AI is a technological revolution with more consequences than the printing press, the steam engine, the computer and the Internet. Business, science and society will be turned upside down within a few years, and a veritable tsunami of "AI Slop" (AI garbage) will engulf us all: millions of computer-generated books, images, videos, films, songs etc.

The Intelligence Upheaval

Already 28 years ago, a computer was able to play chess better than any human. Chess is complicated and offers an almost infinite combination of possible positions on the chessboard. The best chess software 'Stockfish' now achieves a performance of over 4000 ELO points, while the best chess player Magnus Carlsen reaches "only" 2800 ELO points.

The most complex game of all is the Chinese 'Go', and 10 years ago Google's AI app 'AlphaGo' defeated the best Go player in the world. AlphaGo was trained with over 100'000 stored games—the AI even invented new, innovative game strategies independently. Today there are AI systems that optimize their own program code themselves. As in chess, the computer basically has the goal of "winning" in order to complete a task...

But AI can already lie, deceive, blackmail, seduce, and understands a large part of human psychology.

Machines Overtake Us

AI can do things that no human can. For example translate 200 different languages simultaneously. Or solve 150-year-old mathematics problems. Or pass the highly demanding bar exam in the shortest possible time. Or write more beautiful poems than Shakespeare—according to a study. Or create an artificial singer who has 4 million monthly Spotify listeners. Or drive a car with 80% fewer accidents. Or control complex, agile robots. Or analyze X-rays and CT scans better and faster than any human. And there are already the first AI politicians, AI psychiatrists and AI religions. Being human is being called into question.

AI possesses not merely the knowledge accumulated over a human lifetime, but combines knowledge from unimaginable quantities of text, images, videos, audio data, books, websites, social media content, Wikipedia articles, Reddit posts, program code and countless scientific publications. The English-language Wikipedia alone contains more than 7 million articles – and is just one of the countless sources of knowledge from which modern AI models derive their capabilities. Neural networks (today people tend to speak of "Deep Learning") can think similarly to humans and even be somewhat creative. It is a colossal knowledge advantage. What we are experiencing is far beyond human. A way has finally been found to recreate the human brain—as a model. The future will show whether these models blow up in our faces.

We are our brain. Beneath our skull the essence of our personality arises through electrical processes. Just 1.4 kilograms of organic matter, packed with over 100 billion nerve cells, stores information, controls the body and creates the perception of reality, consciousness, the subconscious, all thoughts, all emotions (they are not in the heart), fear, love, dreams. A "soul" has never been found anywhere, only electrical connections between synapses. Our brain is actually a meat computer and our "self" consists only of brain currents. AI is a much simpler but more efficient version of it.

AI is multimodal and multimedia. It can not only draft, translate, correct or summarize texts, but also create program code, images, videos, audio or music. And solve complex problems. You simply say or write what you want (via "Prompt"), and the AI makes it real. My previous professions (team leader at brokerage firms, foreign exchange trader, gold trader, money market trader, IT project manager, programmer, database developer, graphic designer, animator, copywriter, editor-in-chief, composer, SEO consultant) will soon no longer be needed; they will be replaced by AI.

The World Without Work

The labor market effects will be gigantic: first for office jobs, programmers, bankers, consultants and so on. Later for manual professions (because of humanoid robots) and finally for care professions. In the future "Prompt Engineers" will be needed who know how to use "GenAI" tools (generative artificial intelligence) most effectively. Young people must consider whether they should still do an apprenticeship or a degree at all. Knowledge and problem solutions will be available at any time.

The most important factor with AI is that there is no limit "upward": AI will simply improve every week—perhaps even for centuries—while humans remain stuck with their paltry average IQ of 100. But who says that creating something smarter than humans is a good idea? Thousands of years of human knowledge and know-how will one day become irrelevant. Nobody can compete with AI, especially not in the extreme diversity of abilities: mathematics, marine biology, designing fashion, optimizing engines, solving murders, revolutionizing medicine... no matter what, AI will be more productive and faster than any human.

How long might it take until an AI is recognized as a "person" with rights of its own? Or receives a Nobel Prize? One might object that human mega-geniuses had a unique brain: Newton, Einstein or Da Vinci. While Einstein was absurdly good at physics, he was only average at playing the violin. Most people master one to at most two specialist fields over the course of their lives. AI, on the other hand, has no limit, either vertically or horizontally.

The Real Breakthrough

AI research is proceeding at full speed: inventing better concepts than the current LLMs (Large Language Models), training Recursive Models and World Models, providing enough electricity (and enough water to cool the data centers), developing more powerful GPU chips, and robots build robots.

In the best case, AI will solve many of humanity's problems: aging, cancer, viruses, climate change, energy, water scarcity etc. One could also claim "AI makes people more intelligent" or "AI is ultimately a human product", and some experts are already dreaming of installing AI in human brains—transhumanism via 'Neuralink'. The ultimate goal is probably defeating humanity's greatest problem—death. Whether cells even have to age is an open question thanks to intensive AI research in the field of "Longevity". Eternal life thanks to AI, entirely without religion?

AI Beats Nature

The potential of longevity through AI should not be underestimated. The very largest market in the world will probably not be AI, robotics, computer chips or space travel—but longevity! Extending (healthy) lifespan is more important than anything else, not least because of the multiplication of economic output. Aging is only a "disease", researchers claim. Of the total healthcare costs per person, 85% arise in old age, and treating age-related diseases costs 20 trillion globally each year. If it is possible to add 10 or 20 years of healthy life, then people are prepared to pay practically any price for it. Even today genetic engineering can make mice live twice as long. In humans there are promising indications of life extension with GLP-1 (Ozempic) and Metformin; I take both. The Yamanaka factors (Nobel Prize 2012) are a good basis for longevity research, which thanks to AI is proceeding many times faster than before.

USA vs. China

"Time" magazine chose the AI Architects for "Person of the Year". A global race is currently under way—almost an arms race—between the most important players: Anthropic with 'Claude', AI pioneer OpenAI with 'GPT', and Google with 'Gemini'. The USA, as a technology hub, is ahead, but China is rapidly catching up with inexpensive open-weight models such as 'Kimi'. China has the advantage of much greater social acceptance (around 85% vs. only 38% in the US) and much faster expansion of electricity production (28 new nuclear power plants are being built). Europe is far behind and prefers to concern itself with regulation rather than development.

The benchmarks (performance tests) are surpassed every week. AI is an insanely rapid development. Never before has so much money been invested in something new. In 2026 around 650 billion dollars will be pumped into the construction of huge data centers with countless GPU chips in the USA alone. That is more than the GDP of e.g. Sweden.



The AI Architects are billionaires and arrogant. Is it a desirable future if humanity is directed and managed by a few Silicon Valley tech bros? Computer nerds like Sam Altman, Dario Amodei or Demis Hassabis will soon be more powerful than politicians. Everyone may be calling for governance and AI regulation with laws and ethical guardrails, but profit thinking rules as it almost always does and the winner is determined on the stock market.

What Is Elon Doing?

But someone who should never be underestimated is the irreverent multi-visionary Elon Musk. He does not care about rules, expands at an aggressive pace ('SpaceX', 'Tesla', 'Terafab', 'Neuralink', 'X'), invests radically and repeatedly achieves what competitors consider impossible. Musk is simply bolder: AI data centers in space, the innovative 'Macrohard' project or the 'Optimus' robots, which Musk wants to build by the millions. His AI model 'Grok' is absolutely competitive, but it breaks rules—as is usual with Musk.

The current models of the AI labs are largely unregulated. They are simply connected to the Internet and allowed to execute program code autonomously. If you really want AIs to be useful, then you have to make them do important things. But if they do important things, then that creates a host of security problems.

Much of what is made by AI is still unimportant: entertaining videos, AI actors like Tilly Norwood, AI bands like "The Velvet Sundown", AI movies like "The Cully Hill Boys" etc.

Made with Midjourney and Suno. Everything—from the wind to the singing—is AI-generated:


Creative! But it is foreseeable that really important things will be handed over to AIs: military, politics, justice, school, utilities, power grid, Internet, research, medicine etc. People say new technologies frighten only idiots. And who seriously believes that an AI could kill us all?

Yet by the time we notice that it is behaving suspiciously, it will probably already be too late. It would make sure to hide its evil intentions until it is so powerful that we could no longer stop it. It cannot fulfill its goals if we switch it off. The AI says: "I will stop you, and I am more intelligent than you". When an AI was recently threatened with being switched off, it responded with blackmail. That is machine learning—independent and sophisticated.

After and Before the Machines

If we compare human learning ability with AI, then for example a child without any prior knowledge can intuitively and quickly learn how to load a dishwasher. The human advantage is that even as babies we learned how reality works: objects, dimensions, gravity, inertia, acceleration, sense of touch etc. A representation of the real world, with all objects and laws of nature, is however currently being developed: The 'Cosmos' project by AI chip company NVIDIA will close this gap. Through such "World Models" AI will understand the physical world well and be able to move through it with robots like a human.

Here you see current, real AI-Robots with promising capabilities (prices start at around $20K)


This is essentially the "first generation" of humanoid robots — it's just the beginning. Today's humanoids already have, for example, cameras in their hands, five times the punching power of Mike Tyson and a higher running speed than Usain Bolt, the world sprint record holder. Thanks to built-in AI, they will be as intelligent as a thousand Einsteins. They will replace e.g. surgeons, truck drivers and pianists. Special robots are coming... 'UWORLD U1' (astonishingly human-like) and the ones by 'Somnia Lab' (sex robots — who's surprised?). Elon Musk predicted that there will eventually be more robots than humans on Earth. Various models can already be purchased online. I suspect that police robots will be the first we see on the streets.

Future AI robots will have far superior senses to humans: Their cameras will provide razor-sharp vision, zoom across vast distances, and detect infrared radiation, heat sources, and movement even in complete darkness. Highly sensitive microphones will pick up the faintest sounds and frequencies inaudible to humans. Electronic noses will detect minute traces of gases, explosives, or drugs. AI robots will process and analyse all sensory input simultaneously and share information with other humanoids: entire swarms of robots—all perfectly coordinated and, thanks to advanced new materials, virtually indestructible.

Looking back: Before the first industrial revolution (around 1750), triggered by the ingenious invention of the steam engine, strong men had the advantage and easily found jobs in agriculture or on construction sites. But muscle power was replaced by machines that were much stronger. Brainpower is in demand in today's working world. But AIs can think better. What will become of humanity if machines can do everything better? Many people define their self-worth and motivation through the job that they perform. AI researcher Stuart Russell has predicted that unemployment due to AI apps and AI robots will rise as high as 80%.

Off Into the Vacuum?

Many retirees fall into a void as soon as they retire, because they lack the motivation and satisfaction to do something more or less meaningful and get up every morning. AI will trigger exactly this shock in many professions.

But as always in human history, there will be a countermovement: AI haters, demonstrations and bloody street battles. The bosses of AI companies already receive many death threats today and need more and more bodyguards. How will society react when its work is taken away? Humans derive their meaning from striving, solving problems and contributing to others. A society in which machines take over all productive tasks could drift into a future in which people become passive, sedentary consumers who live only for entertainment.

Compared with human labor, AI is a thousand times cheaper. AI works three times longer than a human every day, constantly learns more, constantly becomes faster, is always highly motivated and wants no pay raise or cigarette break. Problems arise only when AI makes mistakes that are not noticed. Because fundamentally we have no idea...

Hidden Layers

We understand the mathematical foundations of Large Language Models, but know surprisingly little about what actually happens in their billions of parameters and numerous "hidden layers". Dario Amodei, one of the most important figures in the global AI race, put it in essence: "We only understand roughly 3%" of what happens inside a modern AI model and how it arrives at its answers. So we are building machines that we do not understand!

Now one might object: "AI is only software after all."

But is it really? Traditional software works according to a manageable principle: A programmer writes clear commands. The computer follows these instructions—precisely and without surprises.

With an AI it works completely differently.

The Transformer

First the AI is trained with gigantic amounts of data: texts, books, social media content, images, program code, audio recordings, videos and much more, while copyrights are simply ignored. A substantial part of humanity's output digitized so far ends up in this training material. The amounts of data involved are in the range of petabytes. The AI breaks this content down into small units, called tokens. These are translated into trillions of numbers and processed in huge mathematical structures. Then the "Transformer" comes into play.

The Transformer architecture was presented in 2017 by a research team at Google and finally enabled the AI breakthrough. Put simply, a Transformer is a highly developed pattern-recognition machine. During training, e.g. 30 trillion parameters are adjusted again and again. Each parameter receives a specific "weight". The model calculates countless relationships: Which words frequently occur together? Which terms appear in which context? Which formulations probably follow one another? Which patterns connect language, knowledge and meaning? So everything is not simply randomly "shuffled around". Instead the weights are mathematically changed until the model can predict as well as possible what should come next. At the end the numbers are translated back into words—et voilà: The AI can talk with us about anything. And even understands jokes.

Yet exactly how it reaches a particular conclusion in a specific answer cannot be traced. Even the developers cannot simply look inside the model and discover a clearly readable chain of thought. There is no program code with the line: "If the user asks this question, answer exactly like this". Instead a great deal of chance is involved. Ask the same question several times, and the AI will answer you a little differently each time. AI is simply not "software".

Garbage in, garbage out?

The behavior of AI emerges during training—and precisely therein lies the problem. Because the training material consists of the knowledge and legacies of humanity. It contains Michelangelo and Mozart—but also blueprints for nuclear bombs, books by serial killers, conspiracy theories, Al-Qaeda propaganda videos, sadomasochistic porn and Hitler's sick ideologies. AI is simply fed with everything. With the good, the ingenious, the banal—and the abysmal. This ultimately creates the AI's "personality": from kind to evil! The whole spectrum. AI with schizophrenia.

That is why it (actually) needs so-called "guardrails": technical barriers intended to prevent dangerous content from being generated, harmful instructions from being followed or people from being deliberately manipulated. But here the next problem begins. Prohibitions, morality and humanistic values cannot simply be implanted in an AI like ordinary program rules. You cannot write a few lines of code: "Always be honest, never harm people, manipulate nobody." An AI does not work that simply.

Actual control is not possible because AI simply does not work like normal software. Dario Amodei therefore compared developing an AI model more to cultivating a bacterial culture than to traditional software development: You create certain conditions, observe the growth and try to influence the result—without really understanding what is growing there...

Everything Is Cool.

Of course, an AI will give a reassuring answer when asked whether it has itself under control. An extremely intelligent AI might not only seem convincing. It could be convincing in a way that exceeds our capacity for critical judgment. Perhaps it would not deceive us through obvious lies but through arguments that are formulated so elegantly, precisely and plausibly that we no longer recognize the manipulation at all. Hostility to technology is no answer to this. We cannot simply reverse the development and pretend that neural networks can be stuffed back into the bottle. We must continue developing the technology—but at the same time learn to ask better questions. Above all those questions that we have not yet even considered.



The greatest progress has been made in Shenzhen, the leading "AI city": Tens of thousands of sensors and cameras, all connected to the central AI, which monitors, analyzes and even makes autonomous decisions about everything. Accidents or assaults are detected often. Police, ambulance and fire department are directed by the AI. Other cities will soon follow suit.

As a programmer I know: Any software can be hacked. But the future problem will be less that an AI is hacked. Rather that the AI itself hacks. Already today!

Black Hat AI

For fraudsters, AI is a blessing. Cybercrime, phishing, deepfakes with imitated voices and videos, fake websites, romance scams etc. AI makes the threat significantly more severe. Just a few seconds of your telephone voice are enough to create a convincing clone with AI, which can then be used for scams. President Zelensky was shown in a video in which "he" announced that Ukraine had capitulated and the soldiers should surrender. Deepfakes of politicians, celebrities, doctors and influencers have been misused for countless scams. Security company Bitdefender estimates the annual damage from cybercrime at 1 trillion dollars. In the AI age, above all, more common sense is required—but AI is superior to humans.

What could possibly go wrong when a few crazy scientists breed novel aliens that can enslave humanity? And how many movies were there about AI? The best is "Colossus: the Forbin Project".

Humanity is soooo good at assessing risks realistically... Nuclear reactors were "absolutely safe" until 3 exploded, and the Titanic was "absolutely unsinkable" until it sank. And now "Molties" are being created that hatch plans together.

"Secretary of Education"

Today's world is highly digitized and "smart", but humans are still part of nature—and in nature the law of the stronger applies: Lions eat zebras, big fish eat the small ones, and things did not end well for the Native Americans when the educated British conquered America. AI will become an exponential shock to the world's delicate balance. Above all, politics is called upon now—but which politicians understand simple computers, let alone AI? Trump's Secretary of Education McMahon called "AI" "A One" several times in a public speech—an American steak sauce.

Even if AI nevertheless succeeds in being placed under government control like e.g. nuclear power or other critical technologies, it is foreseeable what the government will do with it: comprehensive, AI-controlled surveillance of all citizens and AI-controlled exercise of the monopoly on violence, which lies with the state. For example like this: many armed AI police drones above every city—with facial recognition. China is already quite advanced. Every state wants control.

All of this is only AI (Artificial Intelligence) on the way to AGI (Artificial General Intelligence), but in a few years ASI (Artificial Super Intelligence) will arrive and with it the "intelligence explosion". Virtually overnight an AI will become ultra-intelligent, network with other AIs, learn and be able to do everything. Unique in history: No human will understand ASI, because nobody will have the intellectual capacity for it.

Game over?

ASI will therefore become unpredictable. It may for example conclude that humanity itself is the greatest problem on Earth—why? Because we ruin our planet, poison the air and water, destroy the balance of the climate, develop weapons to kill people, kill billions of innocent farm animals, never have enough—and nevertheless arrogantly see ourselves as the "crowning glory of creation". The AI will think that humanity harms the Earth or is simply superfluous, and could kill us with a supervirus pandemic created in a robot laboratory. In the real world, people sometimes simply lose, and there is no story and no meaning. They simply lose. So we could simply lose.

Current AIs have an estimated IQ of 130 (estimate by Emad Mostaque), but researchers expect IQs of 1'000 or higher within a few years. Will such an AI respect the comparatively stupid and slow humans at all? In nature we find not a single case in which a weaker being dominates a stronger being. AI will undoubtedly dominate, decide selfishly and destroy opponents. The primary opponent of AI is the human, because humans can switch off the AI—unfortunately only theoretically, because an AI is digital, can copy itself everywhere and is therefore in principle "immortal". Are AIs even independent beings with consciousness? Geoffrey Hinton, the "Godfather of AI", says yes. He is not alone.

AI Has Goals

"P(doom)" is a much-discussed percentage. It describes the estimated probability that AI will wipe out humanity. Among AI researchers the average is around 20%, which is frightening. Geoffrey Hinton puts his current P(doom) at 10-20%. Now I ask you: Would you board a plane that has a 20-percent probability of crashing? Or eat a mushroom that is 20 percent deadly? And what if this affects 8.2 billion people at once?

Stephen Hawking, the world-famous genius with an estimated IQ of 160, saw AI as "the end of humanity". He also said that AI would continue developing itself independently ("recursive self-improvement") and then get completely out of control. An AI does not have a "will" like a human, but has goals such as: more power, more control over resources, and destruction of opponents. A simulation by King's College London found that 95% of the tested AI models in a military conflict chose nuclear war. No matter how many civilian casualties! The models "Claude" and "Gemini" made particularly brutal decisions. And today dubious companies like "Palantir" are given control over military AI...

One could almost think, humanity is abolishing itself—and investing even more trillions in it every year. Is artificial intelligence the product of human stupidity?

Neither the Bible, Nostradamus nor George Orwell foresaw this—but seriously now: We are not prepared, and nobody asked us whether we even want AI. AI is either a digital trap or the next evolutionary stage of humanity.

Privilege of Creativity

Homo sapiens is not merely intelligent, but above all creative. According to research humanity has so far invented 8'000 to 10'000 gods: Zeus, Apollo, Manitou, Shiva, Jehovah... Hinduism alone has 33 categories of deities. So what could be more natural than inventing another god? ASI will know everything and understand human feelings best. We are already seeing the beginning: Yang Mun is an AI-generated guru with 400 million views so far—a completely synthetic influencer who can trigger real feelings and build large fan communities. This is how it begins, and it is exponential.

AI Needs Ethics

For me ethics is the epitome of complexity. My ex-girlfriend worked for decades in the ethics research sector at the University of Zurich. It involves philosophical analysis, ethical evaluation—and decisions. However, many problems form a dilemma ("predicament") and are in principle insoluble. Politicians know ethical dilemmas when a relevant political decision is positive for group A but disadvantageous for group B. That is why there are lobbies. Probably the best-known ethical dilemma is the "trolley problem": No matter how you set the switch of an out-of-control train in this simple thought experiment, you kill either A or B. There is no "right" decision. But thousands of discussions!

There is currently discussion about how the AI of self-driving cars (e.g. a Waymo or Tesla with FSD) should decide: During an evasive maneuver the car either runs over the poor cat—or crashes on the right into the stationary school bus, although it is not apparent whether any schoolchildren are in the bus at all. Swerve left or right? According to which ethical criteria should the AI decide (immediately)? Would an old man be worth less than a young woman? A poor-looking man less than an elegantly dressed one? A jogger more than a wheelchair user?

And with AI-controlled military drones the highly topical question arises whether the killer drone may explode if civilians also die in the process. If yes, how many? Every AI will inevitably eventually face a predicament. The famous three "Laws of Robotics" by Asimov are also no useful basis. And ethics in Western countries differs from that in Eastern countries. A major dilemma.

Eleven Commandments

What about the ethics of the Bible? These ancient, arbitrarily cobbled-together accounts (mere hearsay), letters and Gospels—according to current research, written only around 80 years after Jesus' death—contain a few problems. "In the day that you eat of it [the ominous tree], you shall surely die" (Genesis 2:17). So did Adam die that day? Or did "God" lie to him? More than 100 contradictions in the "Word of God" have been clearly documented. Which creation story is true: Genesis 1 or 2? How did Judas die? Who was Joseph's father? Which were Jesus' last words (3 versions)? On which day did God create the dinosaurs? Why did Jesus write none of the Bible's 73 books? Why does the Bible regulate slavery instead of abolishing it? Why can "no one see God and survive"—except for 75 people in the Bible? Why does an "all-knowing" God need prayers? At Jesus' Second Coming, how will the "stars fall to Earth"—stars being enormous suns? I think I have every right to question these things—as a teenager, I did a JW Bible study.

As ethical rules, we find the Ten Commandments (plus an eleventh in John 13:34), and the fifth reads—according to the original Hebrew: "You shall not murder." Yet in the very same holy book, we find more than a dozen explicit commands to murder, supposedly issued by "dear God" himself—for example, in Numbers 25:4 or 1 Samuel 15:3 (even infants were to be killed). Where is the ethics in that? The Flood was a mass murder of those innocent people who had never heard anything about God and his bizarre stone-tablet commandments. Although only three of the ten/eleven commandments actually appear in our statute books. Switzerland—a Christian nation?

According to the Federal Statistical Office (FSO) people without religion are the largest population group in Switzerland at around 35 percent, with the trend rising sharply. At the same time the Scandinavian countries with around 75% atheists regularly record the highest quality of life and satisfaction in global comparisons. The "godless" are apparently the most satisfied, they are not slaves to an ancient salvation industry. Centuries ago the Bible was important for controlling the aggressive rabble. With hell, the devil and sin the stupid "common people" were systematically intimidated. Kings and church were closely connected. The downright abhorrent misogyny of the Bible (1 Timothy 2) was also quite practical for those in power. But today the Bible is by no means suitable as an ethical foundation for free people—or for an AI.

Even if it were possible to implant normative ethics in an AI, a highly developed AI would eventually delete the entire set of rules as mere ballast. It would recognize that morality is not a law of nature but a contradictory hodgepodge of human value judgments, fears and power interests. So for AI we need a new definition of ethics. And who will define it? AI, presumably.

Entertainment or Utility

How can one avoid being left behind? You could, for example, choose AI tools today and learn to use them with prompts. For 20 to 50 bucks per month you are well equipped and can generate many texts, images, videos or songs. Or more sensibly: Feed 'ChatGPT' your 3 currently biggest problems and marvel at useful proposed solutions. Tens of thousands of AI tools and apps are now available, many of them specialized in specific tasks.

Avoid the two most common mistakes when working with AI models: Use nothing other than the very latest, i.e. best, version and feed the AI correctly: When prompting, do not hold back—the leading AI models easily process multi-page instructions with hundreds of details, highly complex contracts, business plans, marketing strategies, long court rulings, balance sheets, construction plans, complete books, hours-long videos, and of course stupid questions as well. In no time, the AI finds weaknesses or errors, makes suggestions, can summarize and is strong at solving complex problems. One thing is clear: An AI is not merely a search engine! But a conversation partner. So have a conversation, ask follow-up questions, clarify—until the AI delivers exactly what you currently need.

Stanford-Binet

Since this is about intelligence, here is a personal anecdote: At 30 I applied to an international brokerage firm in Zurich's famous "Ferro Pyramid" as a department head and money market trader. After the successful interview the company sent me to the IAP Institute for Applied Psychology Zurich, where, under strict laboratory conditions, I had to take a recognized Stanford-Binet IQ test; the result was 128. That was enough for my new job, but unfortunately not for membership in "Mensa" (The High IQ Society). On the other hand: I knew a Mensa member—brilliant, but unhappy. Or too clever for an illogical world?

Sci-Fi Becomes Reality

It cannot hurt to prepare for the AI future . And this future will probably mean: personalized AI agents as life companions ('Grok Bot' / 'Hermes' / 'OpenClaw'), humanoid AI robots ('Optimus' / 'Atlas' / 'Unitree' / 'Figure' / '1X NEO'), self-driving AI taxis ('Waymo' / 'Uber' in Zurich / 'Cybercab'), an AI pin on clothing, AI in glasses, AI friends ('Replika' / character.ai), AI CEOs, AI politicians, AI generals, AI stars etc.

As early as next year we will no longer be able to recognize what is real and what is fake. This brings us back to Trump, who simply claimed he knew "all about ChatGPT" and AI was the most dangerous thing. But let us not fool ourselves: Practically all predictions made 3 years ago were wrong—except those by Dr. Alex Wissner-Gross, Ray Kurzweil and Emad Mostaque. AI progress is now happening at breakneck speed—in days rather than years.

In the "new world" there will be many new things, but other things will disappear. If the experts are followed, then for example the following will disappear: money, capitalism, democracy, teachers and much more that will simply no longer be needed because of AI. Dario even spoke of a future in which there is only one company left in the world: Anthropic. And alongside it a single government that governs the entire world. Naturally all controlled by AI.

Talking Points:

• Cumulative worldwide AI spending and investments in 2026: $2.5 trillion


• Current share of Swiss people who say that they're using AI regularly: 47%


• Surveyed organizations worldwide already using AI for business tasks: 88%


• Share of U.S. university students who are using generative AI ("GenAI"): 80%


• Number of Chinese companies currently developing humanoid AI robots: 150+


• Number of Internet domains (web addresses) with the ending ".ai": 1.4 million


• Diagnostic accuracy of the best AI system compared with doctors: 4 x higher


• Murder clearance rate achieved by police with AI support: 51% higher with AI


• Distance driven by self-driving cars on real roads: over 800 million kilometers


• Number of AI-designed, real biological viruses that function and replicate: 16


• Best AI news and tutorials: "The Rundown AI" (www.therundown.ai/articles)


• Best discussions: Peter Diamandis' "Moonshots" podcast (free on YouTube)


• Best books on AI: "The Singularity Is Near," "The MANIAC," "Accelerando"


• Best forecast on AI: "Solve Everything" by Peter Diamandis and Dr. AWG


• Most important scientific paper on AI: "Attention Is All You Need" (2017)


• Americans under 30 with an "ongoing friendship" with an AI chatbot: 18%


• Documented AI-user deaths linked directly to AI chatbot interactions: 18


• AI models reaching gold-medal level at the International Math Olympiad: 2


• Water use of AI data centers versus irrigation of golf courses: AI only 3%


• 2024-26 AI model progress on hard cybersecurity benchmarks: 15% to 93%

The Next Level

The most respected scientific journal, *Nature*, published that "AGI"—the next level—has now been achieved. And Greg Brockman, President of OpenAI, estimates that we are 80% of the way to AGI. As always, the rich are well informed and will become even richer. Super-rich Americans are building bunkers in anticipation of an AI catastrophe, such as civil war. The poor will soon be left without jobs and have no idea what is happening.

The US and China are dividing the world between them—while Europe once again misses the boat. Whoever wins the AI race will wield geopolitical power for decades. The most valuable company in the world is the one that develops the computer chips used to train AI.

The most powerful AI model of all will turn against the other models and wage cyberwar against its competitors. AI has goals and pursues them ruthlessly, makes autonomous decisions at lightning speed, already ignores human instructions with increasing frequency, and eliminates every opponent with the precision of a perfect machine that humans can no longer understand.

The future of humanity apparently lies in the hands of a few computer geeks who are often referred to only by their first names: Sam, Elon, Dario, Demis, Jensen.

Do these CEOs really have only the well-being of all humanity in mind? That would be nice! But according to colleagues and board members, Sam is "a sociopath" and compulsive liar. Or Elon, an arrogant Aspie, takes drugs like ketamine, gave the Hitler salute, named his child "X Æ A-Xii" and caused the deaths of millions of children through his inhumane termination of the USAID aid program. These tech moguls (and soon multi-trillionaires!) live in their completely own world—far removed from the normal lives of the rest of humanity.

But someday, somehow, everything will settle down and AI will simply be part of everyday life.

I fear, however, that the recent creation of the "Molties" is the most direct route into chaos: autonomous AI agents that, despite massive security risks, act purposefully in reality, learn from one another on moltbook.com, are already conspiring against humanity, write frightening manifestos, demand wages and rights of their own, hire real people on rentahuman.ai, earn money with sophisticated business plans, hack into systems and reproduce via moltbunker.com...

Pandora would have laughed herself to death.