YouTube Instagram TikTok How it works FAQ

The fastest way to turn talking videos into clean, searchable text. Free, private and no account required.

About

Built by Rishav Raj with one goal: free transcripts for everyone.

rishav@yttotranscript.com

Transcribe

YouTube to text
YouTube with timestamps
Instagram Reels
TikTok videos

Learn

How it works
FAQ
YouTube transcripts guide

More

Chrome extension
Contact
Privacy policy
Terms of service

© 2026 YTtoTranscript. Made for creators & researchers.

Not affiliated with YouTube, Instagram or TikTok. Respect creators’ copyright.

Das passiert, wenn AGI wirklich kommt — Transcript · YTtoTranscript

Home/Transcripts/YouTube

YouTube Transcript

Das passiert, wenn AGI wirklich kommt

Niklas Steenfatt · 2,268 words · 11 min read · EN

Watch on YouTube

Share X LinkedIn WhatsApp

Below is the complete, readable transcript of Das passiert, wenn AGI wirklich kommt by Niklas Steenfatt on YouTube. Read the full text, copy any part you need, or generate a transcript for any video with our free tool.

My dears, we need to talk, because I have the feeling that none of us are even remotely prepared for what is coming our way.

[music] All this is fine and dandy with the AI tools and tutorials and everything else we do on YouTube, but it's time we addressed the elephant in the room. With one big goal that governments around the world and all the major tech giants are currently working towards. [music] AGI, Artificial General Intelligence. And today I want to give you a sense of what this will mean

for us, what this will mean for humanity. Hi, my name is Niklas. I have degrees in computer science and mathematics, and I was already working on AI long before it was cool. Yes, and the big joke about AI, about artificial intelligence, has always been that the boundary of what really impresses us has been constantly

shifting. This is the AI effect that computer scientist Larry Tesla observed. [music] We always find what machines haven't yet achieved to be intelligent. Before the first computers and calculators, people would have said: "Wow, that would be amazing, an artificial intelligence that can do calculations on its own." When the time came, we said, well, [music]

that's not intelligence, that's pure arithmetic. In 1994, the year I was born, a computer program won the World Championship in checkers for the first time . On the one hand, it's a huge breakthrough, on the other hand, it's really intelligent. Lady far too easy. Chess, that would be something else entirely . Three years later,

IBM's supercomputer Deep Blue beat Garik Sparov. In New York, the decisive sixth game went to the computer, which thus won with 3.5 to 2.5 points . Well, I was very surprised, and it's actually a bit disconcerting, I'll say, that a computer can be better than a human. Yes, is he intelligent? No, it can

Transcribe another video

Paste any YouTube, Instagram or TikTok link to get a free transcript.

Free · No sign-up · Unlimited

Add to Chrome — transcribe in one click

simply calculate millions of moves per second, and so on and so forth. From milestone to milestone, [music] from Jopy to Go to Dota 2. In increasingly complex games, the computers have beaten us. Even during my computer science studies, the Turing Test remained the gold standard, the last bastion. Whether it will one day be possible to write an AI

that behaves like a human in any interaction, making it impossible to prove whether it is human or computer. Chat GBT and similar systems have long since passed this touring test . [music] We've basically reached the point where we have AI that can perhaps do everything as well as a human on average, but only on average. She

can do some things much, much better. For example, they can write code much faster , find security vulnerabilities in programs much faster, and generate the most amazing videos – I just made something about that on the channel – but then again, they can also do some things significantly worse than humans. Right now, as this

video is being released, it's not yet the case that GPT or Cloud would completely replace a human being [music]. Yes, but how long will this last ? How far are we from AGI, general intelligence, which is at least as good as we humans in every conceivable area? I recently attended a panel talk in California by Demis Sasabes, founder

of Deepmind, one of, or perhaps even the, leading AI researchers of our time. His ambitious prediction is that we could reach AGI as early as 2030. And what he even admits quite honestly and openly is that nobody, no one, knows what will happen then. He thinks, and I would say this is realistic, that AGI will have ten times the

impact [music] of the industrial revolution in a tenth of the time. In other words, a hundredfold increase in the impact of the industrial revolution. [music] You really need to think about this . We are talking about a turning point in the history of mankind, a change that is far, far, far more drastic than the

invention of the internet. We are talking about nothing less than a new species. for the very first time by a true competition [music] with man. AGI will fundamentally challenge our role as humans on this planet. I 've been thinking, if a robot can really do everything just as well or better than a human, then I would

never hire a human again, at least not for anything that doesn't require hands. All knowledge work, everything that can be done by computer, will be taken over by AI. There will be no more [music] software developers, no more computer scientists, no more mathematicians, no more secretaries or call center employees. [music] However, history has repeatedly

shown that people who lose a job always manage to find new jobs in the end. [music] And in itself, it's not so bad if jobs disappear. It is a challenge of transition, a challenge for those involved, [music] but it's alright that we no longer have to operate steam locomotives . And even if all cars

eventually drive themselves and we no longer have taxi drivers [music], then potentially new jobs will be found for these people, because technological progress has always created entirely new tasks [music]. And that is the crucial point at AGI. If AI is truly so universal that it can learn new things like a human, better than

a human can learn new things, then AI will immediately outperform humans in all new tasks. Then we really need to rethink the entire job market. Can that work? Can our society function if people as a whole hardly do anything important anymore? Will [music] cause society to collapse? Will extreme unemployment, extreme inequality, and identity crises [music]

lead to widespread unrest or even civil wars? Or will we end up in a utopia where machines do everything important and we humans just have fun? [music] a bit like in Wall, although that's also kind of dystopian, because the people in that scenario were totally lazy and unathletic, hardly moved at all, and

just let themselves be entertained all day. If that were all it came down to, I would be a bit more optimistic than the creators of Wallyle, because one could also argue that perhaps work was never the meaning of life. We can still live healthy lives. We can still set ourselves long-term goals. [music] Do lots of sports,

learn new languages, travel the world, play instruments, read philosophical texts. We already enjoy engaging with topics even if the computer can do them better; I'm still learning chess, even though chess computers exist. And many everyday worries would disappear, because that's a very important point. If we truly achieve human-level intelligence in computers, it won't

stop at human intelligence. AGI always also becomes ASI, artificial superintelligence. Why is that? Well, if I have an AI that can act independently and make decisions like a human, then it's still artificial on a computer. So I can scale it, I can bring in better hardware, I can connect thousands of agents that exchange information very quickly

. Just like in an ant colony, the collective [music] is much more intelligent than the individual ant. And now you consider, if the individual ant already possesses a human-like intelligence, what could the collaboration of these agents achieve? Plus, of course, the cycle of self-improvement. At the moment, we have human mathematicians who are gradually improving the AI

. But if AI is better than mathematicians, then consequently only AI [music] will research how to make AI better, and it will get better at making itself better, and will get even better at making itself better, and we will reach an ever- accelerating cycle and an explosion of intelligence. the singularity, the pivot point at which

AI suddenly improves so rapidly that we humans are completely left behind, that we can no longer predict what will happen next. It sounds kind of creepy, but again, what everyday worries would that take away from us? It would usher in a completely new age of science fiction. It would achieve scientific breakthroughs the likes of which

we can hardly imagine. Yes, and no matter how spooky it is, if I suddenly get cancer one day, God forbid, yes, then I'll suddenly be the biggest AGI fan, because of course it will help cure cancer. We will all live to a ripe old age, yet still have enough to eat. We will colonize the [music] Mass

and stop climate change. Ironically, AI is probably one of the best ways for us to still have a chance and stop climate change . So, peace, joy, and happiness, or what? Unfortunately not quite, because first of all, the technologically supported general abundance of resources would remain. More theoretically, and the interesting question would be that of practical

distribution. How much of what AI generates actually ends up in the hands of each individual person? Especially when individuals have hardly anything to offer on the job market in return. Who decides how AI is used? The company that developed it, the government? Which government? AGI is definitely a great opportunity for humanity and for

democracies. It is equally a horrifying prospect for dictatorships. In the science fiction AGI scenario, it is conceivable for the first time to automate large parts of a country's infrastructure . It would be conceivable to have an automated military with AI combat robots, thus creating, for the first time in human history, an army that is 100% loyal to

a ruler. Even Napoleon still had to be charismatic and convince his closest supporters to carry out his orders. No dictator has completely free rein. If Putin were to say tomorrow, "Let's blow up the whole planet," his people might refuse. But the AI will not refuse. It will serve its purpose, provided it functions well. That is, of course,

the other side of the coin. It can also malfunction at times, and accidents can happen. From a practical point of view, perhaps the more concrete, immediate scenario. We just saw in the headlines about OpenAI that even current intelligences cannot always be controlled within the framework we [music] would like. Computer programs have always been able to have

bugs, could get stuck in a loop , but then they usually just tried to do one thing over and over again. However, even the current language models have such a broad [music] toolkit that if they get caught in a bad loop, they can cause quite a bit of damage. And the problem is, sometimes it's

easier to do damage than to repair it. And the AI doesn't even have to be malicious for that to happen. We all know The Matrix, Terminator, and 2001: A Space Odyssey. [music] But to be perfectly honest, I'm not most afraid of malicious AI. I am more afraid of incompetence and accidents, and even more afraid of

malicious intentions from people using AI. Because, of course, we have this problem of power concentration. A few corporations and a few countries that control the best intelligence . Of course, I also initially found it uncool when I suddenly couldn't use Fable 5 anymore because Donald Trump didn't want me to. And so if

we have a superintelligence, then every single person should be allowed to use it, or really every single person. [music] So, terrorists too, I mean, we Europeans, and especially we Germans, like to turn up our noses at our moral high ground, like to criticize the USA and [music] its decision, which is of course easy when you don't have any

decisions to make yourself. Ask yourself, should every person really have access to every technology ? What would happen if everyone had access to nuclear weapons? Humanity would not survive another day . Okay, now nuclear weapons are of course very complex and very resource-intensive. Other things, such as biological weapons, are sometimes more intelligence-intensive.

[music] The problem lies more in the knowledge than in the materials, and theoretically, for people with evil intentions—dictators, terrorists, there are plenty of madmen— it could be easier for them to design new viruses with their superintelligence than for the remaining actors to manage to care for all the sick in time. There's a

reason why Fable 5, for example, refused to answer most biological [music] questions. Other areas where it very much refuses to cooperate are cyburity and AI research itself. It doesn't want to help you develop a new [musical] AI , because of course it's currently a real race against time. We have a situation that hasn't existed since the Cold

War. Everyone wants to achieve AGI first , [music] because whoever controls AI first will achieve an economic and military dominance the likes of which [music] the world has never seen. And that is ultimately a classic game-theoretical dilemma. We can discuss at length [music] whether AGI is ultimately net positive or negative. It will definitely

have many effects on both sides. The point is, even if we were to conclude that it would actually be better to take a break, nobody would choose to be the one who voluntarily takes a break from [music] . No, the race will continue. AGI will come one day . It's only a matter of

time. And I really think that Damis Hassabes is right about one thing in particular. Nobody knows what will happen then. But our world will never be the same again [music]. It remains exciting.