AI is a Lie – Cutting Through the Hype

43 thoughts on “AI is a Lie – Cutting Through the Hype”

1. Correction for the sponsor spot at 1:02 – We meant to say "14700 KF", as is shown on screen 🙂

2. Great job trying to keep the explanations simple. The best job I have seen so far. Thank you.

3. The worst thing about this cycle of AI hype is not the tecbros pushing it to make green, it is the average citizen idiot that believes the hype and facilitates its harmful integration in society.

   There’s a reason why the French resistance despised collaborators more than the actual nazis.

4. Your title is a lie. Content creators, CEOs, and other business suits who don't understand how R&D works are the ones pushing this narrative. I can tell by the way you present this that you think you've got a deep understanding, which you don't. You are just another content creator looking for clicks.

5. rare linux social W

6. On a scale of 1 to Clown, where would you position this video in anno 2H 2025?

7. Funny thing. Grok 3 will openly admit to everything you just said about chatbots. You don't even have to ask it outright. Just pretend like you don't understand what it is and where it gets its information.

8. 4:56 making things up as they go. Bullshitting is about as human as it gets.

9. why is this video so wide

10. AI is dumb. And now makes big fortune from phones and other electronic prices. Usually, I have to remind AI that this is 2025, not 2024. It keeps giving me suggestions from previous years, but never from this year. AI is not AI. AI is like to the moon far from actually being an AI.

11. "ai" is a pretty old technology edit: achiving curent power by the way of stealing data.

12. Also: why agi? What needed for?

13. Triple Huray for late capitalism!!! Yay!!!

14. We should really create laws for this.

15. understanding is not computing

16. Here's a marketing opportunity: fast foods can jump on the bandwagon and say their products are "Made with A.I."… Artificial Ingredients! 😄

17. It's strange how machines can give you correct answers and yet they have no idea what the answers mean.

18. AI will never go further than a narrow domain, might touch into context based.

19. The question is not merely what it can do, but whether it can simplify and enhance everyday life for everyone. That was the promise of the internet—and it will be the defining impact of artificial intelligence. The scale or justification matters less than the outcome: making life easier.

20. Excellent presentation and summary. Learnt a lot, thanks. You confirm a lot of my impression of AI being over hyped

21. For now

22. Real AI will never happen I keep telling people this all the time and nobody listens unless it can imitate and think like us with an artificial brain then it's not real AI not even close it's all just cleverly disguised botting software.

23. It's infuriatingly tiring trying to explain other people that what's being sold as an overblown tech advancement is nothing but a new way to market decades old technology. My parents, more specifically my father, won't believe no matter how long a discussion we have about this. And I really wish they'd hear, because these companies don't deserve success in trying to make a fool of other people.

24. I feel like the people who NEED to see this will never see this. We live in a whole new era of snake oil sales businesses

25. Sometimes I compare large language models with students that can learn everthing but they need a million repetitions to get there but as soon as they have gotten there once they won't forget. It's like having to explain a child how to use knife and fork litteraly a million times and rewarding them with a gummy bear each time they do it right.

26. The percentage of people that misuse AI tools is noteworthy.. and the whole marketing of AI in the tech space doesn't help at all.

27. I once just wrote “sugma” to a llm and it went into a reasoning circle that made no sense, never spit out an answer, and the “thinking part” was it going in circles

28. C’mon…..”a lie” is not really a very nuanced descriptor. I agree that it creates the illusion of intelligent agency which is very convincing on certain levels. Being aware of that helps keep the whole thing in perspective. And yes, I get that the occurrence of very convincing seeming intelligent agency might cause some real perhaps super-snarly problems….particularly in the hands of multi-nat corps which have a self-serving agenda. Again, being aware of this kind of thing is key

29. I was a bit shocked at how uninformed the writers are here, but then I've checked the date. I'd say this video is not aging well.

30. we need a part 2 for this

31. A.I. is going to be just like computers in general (even though I love them.) THEY ARE NOT MORE EFFICIENT. They created three times as many problems as they solved. They never work like they do in the commercials. They're like the food you see in advertisements that are really made of painted glue or covered with polyurethane. It doesn't look like that in real life.

32. Mostly correct, but still you made a lot of errors. Maybe you are really just ANI?

33. Hearing an American use the term "Ai Wank" warmed the British cockles of my heart ❤

34. "AI is a Lie – Cutting Through the Hype" : so what? Let them free writing and drawing from AI "training" to stop this over-exploitation of human labor and then I will recognize the benefit of such admissions that AI is a fraud. Because as long as the exploitation continues, revealing the truth that AI is a lie is of little use. At the same time, I CONGRATULATIONS YOU! The truth must be told. But without consequences for the guilty, the truth has little meaning. ❤

35. Glad to see a mainstream creator not pushing AI hype 😇. Tired of folks thinking it’ll replace everyone at work—employers just use that fear to squeeze more labor. Reality check: people are cheaper, smarter, and can fix themselves. Robots are pricey, and no ChatGPT is making complex parts in a factory anytime soon.

36. high grade panderslop

37. AI is a search engine.

38. Thank you Linus for this video. As CompSci grad, I knew most of it but couldn't word it in a way to explain it to normie people. I loved the monkey and octopus analogies. Great work!

39. Ai is A-Lie. But A-why?

40. my biggest issue is still not addressed in this video properly. Machine learning and statistical algorithms like chatGPT are fundamentally incapable of what we humans call "understanding a concept". It's not just "a narrow version of what our brain does" because a human brain, even if you're only talking about turning experiences into new output does this by understanding the experiences, and the concepts and using that understanding as a basis for new creation. chatGPt is fundamentally INCAPABLE of this. A great example is "full glass of wine". I've neverr seen a full glass of wine but i can imagine it because i understand the concept of "wine" (separate from it's container) and "full glass", chatGPT simply has no understanding of what "wine" is in the way we do, which is why it simply cannot generate a full glass of wine, because it doesn't know what "wine" is, all it knows is the word "wine" (or some set of numbers that refer to that word) and what, statistically, it usually refers to when that word is used to describe something. It's not an understanding, it's just "all of these images are described using the word "wine" and this is what all these images have in common so the word "wine" = "that thing which it has in common". It's not even abstracting the concept of wine, it's literally just A=B with some statistics thrown in.

41. "What we have now are decent summarization machines and lukewarm guessing machines"
    On 🎯……exactly….
    ….. stock purchase goals…..

    And that is where the problem is. Some are treating the tech as an infallible guide incapable of error.

    California lawyer fined $10,000 for an AI generated BRIEF. Claimed precedents that did not exist. A "drunken librarian" could have done better.

42. AI is just a few years old, and really only a year from public release. It leans as it goes and shares with other systems. Add the dozens and dozens of data centers being built all over. look at YouTube and see it can be trained to make games play games. The growth will be expediently huge in the next few years. AI may not be able to think, but task such as programming, research, and such it will be able to train itself in days or weeks to be better at a large percent on humans.

43. if it's not conscious, it's not intelligent. it's really that simple

Comments are closed.