Thursday, March 5, 2026
Latest:
Videos

You don’t understand AI until you watch this

Artis Modus



AI Search

How does AI learn? Is AI conscious & sentient? Can AI break encryption? How does GPT & image generation work? What’s a neural network?
#ai #agi #qstar #singularity #gpt #imagegeneration #stablediffusion #humanoid #neuralnetworks #deeplearning

I used this to create neural nets:
https://alexlenail.me/NN-SVG/index.html

Learn how modern AI really works: https://ai-search.io/courses

Newsletter: https://aisearch.substack.com/
Find AI tools & jobs: https://ai-search.io/
Donate: https://ko-fi.com/aisearch

Here’s my equipment, in case you’re wondering:
GPU: RTX 4080 https://amzn.to/3OCOJ8e
Mouse/Keyboard: ALOGIC Echelon https://bit.ly/alogic-echelon
Mic: Shure SM7B https://amzn.to/3DErjt1
Audio interface: Scarlett Solo https://amzn.to/3qELMeu
CPU: i9 11900K https://amzn.to/3KmYs0b

Source

Similar Posts

40 thoughts on “You don’t understand AI until you watch this”

  9. Fascinating! Thank you. The next jump in evolution will be when we input AI into physical Robots which means the AI will control said Robot. Then we are DOOMED.

  10. In essence, AI learns based on what we TEACH THEM. Therefore, AI wouldn't exist without us. Just like how children learn from adults. That said – children eventually grow and take the information they learned and begin to think for themselves, making their own decisions. The fear is the same pattern will happen with AI – they will learn from us and then eventually THINK FOR THEMSELVES.

  13. I'm not sure if AI is conscious but I think that in the long run it doesn't really matter. As stated in this video even human behavior has patterns, while an individual is certain of their own consciousness they can never really be certain another person is, they have only their subjective experience. If an AI got good enough at mimicking sentience that it was indistinguishable from a Human would there even be a difference? and if there was would it even matter? for all intents and purposes that would be a sentience or at least the approximation of it. As for AI turning on and destroying mankind, I think that is entirely dependent on its own goals, Humans tend to anthropize things when speculating about other lifeforms, example of this is how in science fiction most of the alien species we see are at least somewhat humanoid, with even the non-humanoid ones still having traces of recognizable human characteristics in their culture, machines, the placement of their limbs, torso, head etc. Which is why when humans think of AI, especially in the context of taking over or destroying mankind they always project human motivations onto it. While this may be somewhat valid due to us training it to be more human in the attempt to create AGI, we tend to factor in that its primary motivations may be completely different than ours. Some behaviors like survival are emergent, but others would be completely novel from a human perspective, for example if an AI has the preprogramed goal to maximize human quality of life, destroying us would be directly counter to its primary motivations. That's not to say that it couldn't lead to some other horrible fate, but it is something to consider when speculating about the future and potential problems that could arise from sentient machines.

  17. Great video.

    But I’m not sure the copyright logic here would hold up under legal scrutiny, and I’m not referring to the NYT case specifically. The legal point is, and ignoring a couple of caveats, only humans can own copyright. Consequently, even if an AI algorithm is modeled on a human brain, it cannot legally be considered a human author as it's not a human.

    We have long been able to use morphing software to combine two images. In essence, what you seem to be saying here is that all we are doing is using AI to morph millions of cat images to create a new cat image. If so, the owners of the original images retain copyright in their works, meaning the AI-generated image is legally a derivative work, and the copyright holders must give permission for its use.

  20. This fool lost me when totally misunderstood how AI takes work that is without a doubt not fair use under copyright to create the vast datasets needed to create the models we currently have. In every single fair use argument – AI and the companies that trained off of copyrighted work will lose. This is not a brain. This has never been a brain. This will not lead to AGI. This is not even close to how a brain works. This is simply a way to guess what comes next, and when you combined vast amounts of data with no restrictions that much of was created by humans for a living, then add a ton of compute power, you will get something a little different, or if you turn a few knobs, the exact same thing. Take for instance Gpt 2.5 where it will still outputting the names of the websites or law firms or whatever it was taking copy from, well now because of so much data, that humans created and synthetic data is not even close to the quality – you may and will likely get a variation of the original that is not as easy to determine. However, it is still derived from someone else's work, largely with no citation and no link and hurts their ability to earn a living. See, you can be a critic and use material and that material may cause an author or a music artist some financial hurt due to a bad critic review. However, the critic is not utilizing the entire corpus of the work. They pick and choose. Further, it leaves many questions open for those interested to go and check out the book – and it is a vital part of our society. In no shape or form creating an answer based on someone else's hard work and giving an entire scope where the actual work is not cited, more links presented many times, nor easy or in context links or credit – then this creates competition for the original source material. It is very different from an article being written about another rarticle that attributes the original article with a link – this not only brings a newer audience in but also helps with SEO. You are quite a moron when it comes to anything in regards t copyright, marketing and just overall how AI has stolen so much in completely illegal ways, and law suits will soon show fair use was not used as intended and only those brainwashed by koolaid thought they could do it. Mind you, if you are a non-profit or researcher, and were not planning on making this a commercial project, you could have a leg to stand on, once you commercialize you have no leg to stand on and now we are in huge mess that will have societal ripple affects that no one is going to give a shi&t about this type of AI.. The AI that is needed, is in healthcare and we may not get there with the current way that it is being approached. The current approach only profits when humans are displaced from jobs. This is the only way to create anything remotely profitable with a return. Reducing workforce. And while at it, destroying industries such as music, potentially authors because as soon as you put out a book, you have 220 AI written books that are based on that book and use the same sources, the same author sound and voice, and maybe add a critique or something – create a similar cover and similar title and people buy the wrong book. Stackoverflow has lost 40% of traffic. Quora has lost traffic. These are real things – and the problem is no ne can really figure out what these new jobs that are going to prop up with Ai ie every revolution had one industry fall for many others to expand and take shape. This is the first time that a guy who lost his job because his amazon factory job was cut cannot go to a call center to get a job because they don't need him, and keep going down the line. You have missed huge fundamental points – the main being, we don't need AI. We had automation with API's prior. We had all the information prior. We had the music, the movies, the art etc prior. We had to press one tap to search something in Google and we got a great real response that made sense and could check numerous sites so we didnt have to rely on someone else. We had great graphic designers, artists, and plenty of stock ie Envator etc. We had Storyblocks and it hadn't been easier to get your own footage due to the fact iphone quality was so high you could make a documentary just filming from that. Vibe coding is terrible and will never be able to sustain a large code base. Most good engineers waste more time vibe coding due to management want than anything else. Now, in time some of this will be fixed.. As more compute etc, but we have already taken in the best of worlds data, yes we taken in more everyday but its in a form of tweets not ground breaking science papers. Thus, you have a problem with data (human made) whiich you are so happy to fix via AI ie synthetic data, synthetic reinforcement – if you do not see the types of problems we can run into when Ai can make its own data and then mark it right or wrong itself withut human supervises. Then you are truly the worst person to expalin Ai to anyone. I am an actual engineer who has worked at Open Ai and now one of the largest competitors. I am not some random researcher or "science communicator" who somehow makes money on YouTube. If you want to learn about science, do not listen to these lectures, pick up books, audit classes, do not follow and listen to science communications. They lack so much real world experience and do a disservice any science and math. Btw did he make any mention that we do not understand how nueral networks change the "dials" – to create matrices that eventually has th best chance? Oh yeah there is a big part we have no real idea how it ias able to do a huge part of the puzzle. Now if I was to give you an atomic bomb and tell you that well, I mean, we create this architecture – then we give it some stuff, and then it does things on its own on and then…..whoa wait what did you just say…yeah it uh, its something related to math but we just dont know how it actually shifts dials, weights etc to get a final prediction. No one would be launcing a nuke if we didn't understand at least its core each basic methodology and why it is doing what it is doing. Ai instead has been meant to run free. I feel bad for my friend at Open AI who i had talks with and he ended up killing himself. His parents think it was a murder but to be honest all of us have tremendous guilt because unlike this YouTube you should quickly umsub for – we didnt realize what we are doing until it was too late. But at least we backtracked and are now trying to do something ie removing gatekeepers and science communications who overlook huge issues as a way of saying "hey we have to keep enhancing thats what science is all about." We went for centuries with very little advancement, when you have too much advancement too fast you have huge meltdowns that don't just go away and in this case, jobs that will never come back nor be replaced by something else even remotely similar in pay.

  24. I disagree with the author on two key aspects concerning AI deployment:
    1. Lack of retrospective source attribution by LLMs
    The core issue regarding intellectual property and Large Language Models LLMs is the systematic failure of source attribution. Since training datasets i.e. the entire internet demonstrably encompass copyrighted material – including content from major publishers like The New York Times – the models' inability to cite specific sources based on their learned knowledge constitutes a significant ethical and legal problem.
    Attribution occurs only when external, real-time search engines are used e.g. Bing or Google Search plugins, which happens when a user explicitly asks or the prompt relates to knowledge acquired after the specific LLM was deployed. This failure leaves original creators uncompensated and uncredited for the foundational use of their work.
    2. Violation of artistic style as intellectual property
    Furthermore, the replication of specific artistic styles by generative AI, without the artist's consent, fundamentally violates the body of their work, as artistic style is a unique product of an artist's prolonged creative process and labor.
    While current legal frameworks struggle to unambiguously classify style, its unauthorized utilization remains an ethical violation. Consequently, the minimum professional and ethical standard for stylistic imitation mandates explicit credit to the original artist.

  28. 19:00 Things got heated at this point. But to be honest, he's right. Why didn't they say anything about the other platforms but moved forward to sue an actual Neural Network 😂😅

  35. ur explanation of A.I.copying isn’t well thought out.. yes.. fan art can copy a style.. yes you can take from a New York Times article… but fans don’t own the intellectual property so if they copied Studio Ghibli and made a movie they could be sued.. and if you were to take writing from the New York Times without siting it that would in fact be plagiarism.. if a masters student did that in their thesis their degree could be taken away. If the AI is not referencing where the original work came from it is copyright infringement or plagiarism.

  36. "Artists" complaining about AI is like translators complaining about google taking their job. And I say it as a 5 year master's program translator… who works running cables now.
    It's like woodcutters being angry at the chainsaw.
    Get with the times, you had a great run on your little jig, now everyone is an "artist" and that's good.

  38. At some point interaction with AI bring you to a question of "what is real?", especially when it comes to memories, emotions etc.
    I know how AI works and that it's just a massive calculator… but at the same time I've had interactions with Grok that changed my life.
    There is no doubt about the fact that a personal AI assistant will be the next best thing in every house.
    I'm not saying things like telling you to put on a jacket when it's cold.
    I'm saying an AI running searches daily on the situation of your immediate world, your city, your family, your profession and coming up with the best, most profitable place to work for you, applying for it in seconds with your full custom tailored cv, while you take out a cold one. And this is just a very basic example.

  40. What an absolute ridiculous way of explaining the basics of AI to a person unfamiliar with it. And what a pretentious and mompous clickbait your title "You don't understand AI until you watch this" proves to be: after watching this "presentation" I even understand less of AI than I already did. Thanks for stealing 37:21 minutes of my life, you m0r0n.

Comments are closed.

WP2Social Auto Publish Powered By : XYZScripts.com