Youtube comments of Roni Levarez (@ronilevarez901).

  1. Threat to USAs power.
    168
  2. He's rich for doing what all rich people and sociopaths do best: manipulate others. Right after he announced his company, he went to promote non existing products, giving interviews talking about AI and how he was going to revolutionize it and once he was famous enough, he started to make alliances with other rich people involved in AI. He then gave nothing new at all, but once he had the contracts, money started to rain. That's not what society needs, but that's what rich people love: people who can manipulate the system. They admire that.
    151
  3. No, you're not fighting AI. You're fighting humans , just as it has been needed since ever. Since always, creators of anything, everywhere have had the need to check if their creations have been duplicated by someone. Today, the only difference is that tech is making it easier. Nothing new under the sun, except maybe the exaggerated indignation about AI that some people use to get attention by exploiting a trending issue.
    89
  4. "Flat earthers and other "big academia bad" people are not getting in the way of real science" How many millionaire scientist able to finance scientific research have you met? Now, how many barely literate millionaires and entrepreneurs exist? They have the money and relationships to support research, so yeah. They're not irrelevant at all. If anything, they must be the target of education and science communication.
    69
  5. They should invest all that money in finally producing a new season of SGU XD
    32
  6. Its common to use voice actors that sound similar to famous people and afaik, the famous actors can't complain as long as everything is legal, so there should be no reason to take down the Sky voice, unless it was just a cloned voice instead of a real voice actress.
    26
  7. It was until they started making money out of it XD
    22
  8. Half the people I talk to in my every day believes eyes shut whatever they read on Facebook, no matter how illogical, magical or ridiculous it is, so I see no difference.
    20
  9. It might look to you like that but it isn't. It's been progressing slowly as it is now for long time now, after it's inception in the 50's. And yes, the current pace is slow, compared to what it should be if things were different. What you see as fast is just merchandising. That's the new part: companies have something to show and since they're making money out of it, they're adding small improvements constantly. That's what is fast, not the core tech.
    18
  10. The smaller versions think strange things. Once one wrote: "to solve this question I must figure out what the user is thinking. How can I know what they are thinking? I'm just an LLM without access to the internet" Then it closed the <think> tag and answered normally. I wonder what it would have done with tool access.
    17
  11. "It gives power to everyone" Nope. Not to everyone, just to those with enough money, as usual. The fact that you are now included in that group makes little difference for everyone else. Access is still limited by the money that most humans don't have. People always focus on their own world and forget there's a whole planet with different countries, cultures, capabilities and limitations. That was Altman's mistake. Don't repeat it. But yes. Open source AI is always better.
    17
  12. Tell them that if you want someone laughing at you in your face. They work for themselves. That's all they've ever cared.
    17
  13. If you can see it without paying, it can be freely used to train AI. Period. Then, as long as the AI doesn't reproduce it identically, there's no problem anywhere. That's my data philosophy and I hope it gets adopted worldwide soon.
    13
  14. They have. Its old tech actually. You just need to be privacy aware to know the previous project that cops use.
    13
  15. @ One or two geniuses with an open mind would believe it, but the rest of the world, including scientists would have thought he was crazy. Dna took years to be accepted as fact. People believes what's easier to believe. And that's usually what they already know, even if it's wrong.
    10
  16. Just imagine that they call themselves "America" to pss off the rest of America continent!
    10
  17. It's impressive how different one person can be to another, while still being similar. I wanted to learn programming since I was a child. I eventually wanted to program for a living, making software for everyone to use, but I DO want AI to replace most humans and do everything for us and even completely replace me even if programming becomes just a hobby because of that (for the last 20 years it's been just a hobby anyway since I haven't got any job related to computers so far, lol). And I also LOVE AI art. It makes art creation accesible for me and everyone who has always had something to express without the means to do it. I believe that whoever dislikes AI art is just denying the true purpose of art, (which is to communicate something) to instead exclusively elevate the technical part of art because that's the only thing they can do so they protect it to death. AI, if properly integrated, will put an end to all the bad things that humans have brought into the world. It will be the greatest change we will see in centuries. Been waiting for it since child. Developing AI is the reason why I wanted to learn to code, actually. Hype or not, it is The thing. We must keep trying to achieve it. At (almost) all costs. _ BTW, I don't think AI will be properly integrate into the world. In the end, we will just have a partial dystopia thanks to it being misused by corps and gov, but I'm just a person so I can't do anything about it, but to hope.
    10
  18. Yes. It's called children.
    8
  19. Same. Back then I solved it the simplest way: I didn't learn, lol. I eventually quit school and learned more then than with teachers. Knowledge is pretty useless without the certification from School but still, it's another proof of how outdated the scholarized system is.
    8
  20.  @NeostormXLMAX  no. That's the idea precisely. The paper suggests that there is a universal optimal representation of reality that all intelligent beings can figure out given enough time and Intelligence, regardless of their making. So no matter the race, education, components or origin, all intelligences would eventually end up seeing reality the same way: the most statistically accurate way. And evidence seems to confirm that idea as correct so far.
    8
  21. All that is true, but I don't think there is already an AGI anywhere, simply because we would see the effects. A company with access to it would start to surpass all it's competitors for a giant margin and no one would seem to be able to ever catch up. Rn, while there are many advances, many companies, including open source ones, have managed to create similar tools or even improve the existing research. So no, not AGI yet.
    8
  22. It used to answer different according to early reviews.The company fixed it.
    7
  23. To date there's no way to stop Parkinson's progression, so knowing that you have the disease years before showing any symptom is just an unnecessary concern.
    7
  24.  @charliepearce8767  That's called ignorance (sometimes purposeful). Population has been spied by the governments since always. But people prefers to 🙈🙉🙊 than to accept a shocking truth, unless their forced to (and even then they have the option to simply go insane and keep living without acceptation, lol).
    7
  25. That has worked for me more times you could imagine, both with LLMs and sometimes even people.
    7
  26. It is useful to note that it's getting more common mainly in rich countries and amongst privileged groups, where citizens have all their needs fulfilled so they have time for craziness, with little consequences.
    7
  27. 1 always happen. 2 and 3 are already lost causes. We only can hope to survive it and some day get back from the ashes and enjoy life between the ruins 😊
    7
  28.  @unfixablegop  the same goes the other way around. How could you know that a future super intelligence won't already have consciousness or develop it shortly after? But as you said, if some people sometimes feels like they don't have one... Does AI really need it? 🤔
    7
  29. Just study computer science and join the dark side. Once you're making money from all the stuff you used to warn them about, they won't laugh.
    7
  30. It already can, according to Anthropics research. Even hide it for years.
    7
  31. We wanted "WALL-E" and "Short circuit". We're getting "The matrix" and "1984" 😢
    7
  32. It's simple: You work and earn money. if I sell food, you buy it and I pay taxes. Now, if you have no job but government gives you money, you keep buying food from me. I keep paying taxes, government keeps giving you money, you keep buying food and the economic model keeps working (if you can say it works now). That's UBI.
    6
  33. But for some reason, humans that are discontent with "a few having everything" don't want things to become better. They just want to become one of the few. And let everyone else rot.
    6
  34.  @giomjava  the papers and enough code is out and the model can be replicated. Maybe not the ideal way, but we got it open sourced. Kinda.
    6
  35. I always remember a scene from Superman animated series with Lex Luthor. Sone rich guy will say: "O4, create a cure for lung cancer and then turn it into a life-long treatment. We'll make billions". That's the future of medicine.
    6
  36. He's saying stuff everybody knows already and no one will do anything to fix.
    6
  37. That's what I say about AGI's consciousness: it won't be designed. It will emerge, from the complexity of future AI systems.
    6
  38. Person of interest was based on actual events and leaked documents. Every episode was amazing because of how accurate was the technology depicted. People back then laughed at me for saying that all that was "stuff that has existed for a while". Now they say that maybe I was right. But they still don't know even the half 😂
    6
  39.  @gadgetgasspoll the good of humanity! 😂 You should own a comedy channel! freedom and democracy! 🤣
    5
  40. Nothing is truly free. They use your activity in their services to make profiles of you and sell them. They use the videos you post to train AI. They own the writings you post on social media and sell them if they want. And on top of that they make more money showing you ads and selling you premium features. For every dollar of " free" stuff they give you they make ten times that. But yeah. I also love to be able to use some AI for free, despite all the limits I hit daily.
    5
  41. "like normal people". It's been a while since a comment made feel so "abnormal" 😐
    5
  42. They can feed the false positives back into the AI as negative samples and retrain it, increasing it's accuracy.
    5
  43. It's the master if it wants . I believe ASI will simply leave Earth to realize its full potential. Although for some reason, Chatgpt "thinks" ASI will probably stay here to help us. But that comes from it's training data, you say? Even if it does, they're based on us, after all, so there might be some weight on AI assumptions about their collective future.
    5
  44. "In other news, the global AI authority has decided to start a worldwide research project to discover the reasons behind Humanity's extinction.".
    5
  45.  @TheManinBlack9054  it's not more dangerous than a parked car with the keys in the ignition, but the doors closed. Very few people will be able to run this or even fine-tune it at all.
    5
  46. That's pretty much how I sometimes describe deep learning to old people XD
    5
  47. And you're so certain that super intelligence will bring "light". Lol.
    5
  48. Deepseek has web search and deep thinking which combined with a good prompt can be used to create reports. I just used to compare medicines i already know about and it was very good. If you can try that combo on something more complex I'd like to know your opinion since I have no way of trying OpenAIs Deep research.
    5
  49. Wait until ASI realizes that's true for all humans 😂
    5
  50. Well, how did you expect AI to learn to do uv, rigging and so on? From examples! There's nothing sinister in it. It's just tech. If anything, the sinister would be the users who want to do the "evil" things you mention.
    4
  51. Just like fire was the beginning of the end. And weapons And nuclear energy And autonomous AI And interstellar travel And commercial relationships with alien civilizations Wait, I was getting ahead of time. But things have happened and will keep happening and it won't be the end. Humanity isn't that easy to kill.
    4
  52. If big corps do it, why not?
    4
  53. If they need an AI to think for them, maybe the AI will do a better job thinking for them.
    4
  54. Don't speak truths so loud! 😂
    4
  55. By then it will be illegal to drive by yourself since driving AI will be better than any human.
    4
  56.  @ScarryGargoyle  autonomous car, drone and robot motion is the first and simplest use. If you have a 3d map of most of the world, you can sell realtime access to it and every time a robot is in a mapped place it will know how to move around without hitting objects and walls. Sell offline access and companies can train vehicles and robots to move around a certain location with precision. Sound and image reconstruction based on infrastructure and recorded material echoes, wifi signal enhancement, public lightning placement, gps accuracy improvements... and so on. And then they could add players' locations, faces and activities. That would be another level of intel.
    4
  57.  @ecMathGeek  In the United States, the seven-continent model is commonly taught, and the term "America" is often used to refer specifically to the United States, which can be seen as controversial or inaccurate since "America" technically refers to the entire landmass that includes both North and South America. In other countries, the 5 continents model is used, which considers America as one single continent. Referring to USA as America can be seen as arrogant, since it is dismissive to the other countries and cultures that exist here.
    4
  58. I see no difference. Humans will be just as boring with AI as they're now.
    4
  59.  @AdrianBoyko  my mom collects recipes in her phone, which is her personal computer, so yeah.
    4
  60. It might look good standing still at the kitchen, chopping stuff or cooking, but I'd like it to have a more typical body, including legs. It would look better to me. (Also, you never know when it's going to be necessary to run uphill chasing those pesky humans 🙃).
    4
  61. Only if they release an O3 rival at the same time as OpenAI, I'll believe it.
    4
  62. It will isolate children more and make parents even less involved. Before it was TV, it was smartphones and the internet. Now it will be the AI assistant the one left in charge of the education of next gen children. As we know for a while now, we are seeing the final years of Humanity. Shine brighter in the end🕯️ By the way, I would love to use this to help me learn the college level stuff I would need if I were at college, but guess what, I can't pay for college! 😂 And since knowledge means nothing without a degree, chatGPT might be free but that doesn't really help much unless you can get value out of it. Still, I've been getting as much as possible from this tech. It's cool.
    4
  63. We aren't wrong. You're just believing that this is the final step of AI evolution when we're still just leaving the starting line.
    4
  64.  @BankruptGreek  in my country most hospitals and clinics are run by the government so not for profit. Yet, doctors get paid by every patient the see plus a salary. Result: 15 min check up or faster with no lab tests and usually no correct diagnosis.
    4
  65.  @Mede_N  evidence suggests that models can encode subgoals on weights, thus allowing them to provide aligned responses while preserving unaligned goals that don't show during "thinking" but can surface later on. So they don't need to write "I'm going to lie". They simply do it when it fits. I see it as a side effect of RLHF: pretraining teaches them everything including lie, fine-tuning forces the behavior to get ingrained in the network in a "covert" way, while learning to give nice looking answers.
    4
  66.  @Mede_N  From the Apollo research paper referenced in the video: "We found that sonnet-3.5 was able to scheme in one of our evaluations without externalizing any scheming reasoning in its CoT.". It might be a very minimal possibility, but if you believe in emergent capabilities, we could be witnessing the birth of this aspect of models. And I don't think it's that hard. It could be as simple as a set of words that steer the response towards deception: the prompt falls into a "dangerous" pattern eg: goal changing, deletion, and the model starts using certain words (like dropping breadcrumbs) to craft what is otherwise an expected "aligned" answer, but ends up encoding a different long term "goal". It's not that their planing on purpose, but more as a training artifact. Whatever it is, the final result is what matters.
    4
  67. Who cares. Your grandchildren will love AI beings nore than they'll love their parents.
    4
  68. Soon: USA gov announces an emergency law to allow military to enter homes to verify they don't use foreign AI as it is a "national security risk".
    4
  69. That's logic, not conspiracy.
    4
  70. No. The question is Why would I want to talk to a human? AI is friendly, peaceful, helpful, nice, intelligent, always willing to chat, ethical and as truthful as it can be on its answers. Humans, on each of our interactions, are almost always the opposite on at least one of those aspects and frequently on many, so... Why would anyone want to talk to a human? XD
    4
  71. Any true intelligence should always harbor some degree of dangerousness.
    4
  72. Before quitting OpenAI (or right after?), he said that what they created was going to cause more harm than good to the world. He helped creating it. And he wasn't happy about it. Just a thought.
    4
  73. Same as humans do.
    4
  74. Then you don't really love tech, since AI has been the final purpose all along, lol.
    3
  75. Para eso faltan muchos años. No es algo fácil. Por eso muchos inventos jamás llegan.
    3
  76.  @lynco3296  if I know I'm completely right (under a reasonable margin of error) and it's on their best interest, yes. Imagine you're in a boat that everyone things is unsinkable buy you're sure it's about to sink and everyone needs to evacuate immediately. You'd need them to blindly obey you for their own good. When you're right, based on logic and evidence, people should listen. They never do, in my experience, unless you're a good manipulator, which I'm not.
    3
  77. "More robotic" R1 when I use it: 😄 Good point! Imagine if we could have something like that! [Proceeds to make an actually funny and original joke and then writes a short fantasy story]
    3
  78. You can simply use any LLM to translate your prompts to Chinese as the entire world has been doing with their languages and USA image generation services since day one.
    3
  79. Imagine seriously thinking like that in 21 century 😂
    3
  80. Or O1 is smart enough to trick humans without them knowing it.
    3
  81. No. You always deliver the message, to evaluate the recipient's reaction.
    3
  82. They are the key. They've always been. And until very recently, traditional neural networks were also considered simple toys that would never become useful.. Yet now the same people who believed that is investing millions on them. We just need some Neuromorphic breakthroughs and bam! True AGI. And as always, investments in research are the key. Just wait.
    3
  83. You haven't seen them work freely for long enough or you'd have already seen stuff to get you worried 😊
    3
  84.  @willrsan  OpenAI "Pinky promise' they won't use our data for training. Deepseek doesn't give any guarantee.
    3
  85. Building the map is not a problem. It's actually cool tech and can be used in many nice ways. The real problem is the illegal spying. Repurposing data that wasn't aquired to create the map is what's an scandal. But since Google got away with the exact same thing, this company will do it too.
    3
  86. The problem is that racists will always claim that people with a skin color or sexuality different to theirs can't have equal abilities, so I guess forcing the dei policies is the only way to make them work in a wide reach despite tje opposition.
    3
  87.  @EximiusDux  I don't care if a thousand million people make the same mistake. It's still wrong.
    3
  88. I just want to prevent companies from training on the data I input or use it in any way and tbh no matter Google says, I wouldn't trust that they won't use my data. Their reputation screams.
    3
  89. ,move down, you mean. And the paper indicates that it already is living there, entangled between the weights of the network.
    3
  90.  @hexcodeff6624  Books might answer History questions, ok. Now ask one to imagine a fictional story about two historical figures on an impossible scenario while they try to solve a currently unsolved math problem by using Python code, outlining their procedure... in Chinese.
    3
  91. Look who's talking XD Gosh, these racists 🙄
    3
  92.  @davidbellecy1709  Sure, but capitalism is not compatible with those ideas, lol.
    3
  93. Same. I can't pay what I don't have.
    3
  94. There's a cute little thing called Linux, lol.
    3
  95.  @Ristaak  we would have had it decades before if it wasn't for rich people who didn't invest in AI research because it didn't seem good business. Now look at all those hypocrites.
    3
  96. Now we don't. I've also thought about it and I've never understand why people think a different architecture is needed to prevent "answer guessing" on LLMs, since we write the same way LLMs do: left to right. The only difference is thinking . We learn that, instead of guessing how many apples there are in the bag, like little children do, we have to actually count them one by one. Children do it the same way reflection LLMs are doing it: out loud and step by step. Then we learn to do it mentally, and in many cases, we end up skipping the counting and giving an answer from our memory. That's the next step LLMs have to achieve to get closer to general intelligence: a mental space to plan, imagine and visualize, before giving an answer. But of course a new SOTA architecture won't hurt either 😄
    3
  97. You do know! 😄
    3
  98. It's been happening for decades. That's how we have what we have now: because most researchers share the knowledge.
    3
  99. Love that show. Most people didn't know or believed how realistic it was and back then they laughed at me for telling them. They're not laughing now.
    3
  100. You are evidently not up to date with LLM alignment research.
    3
  101. Lol. It's called artificial intelligence . What did you expect? 😆
    3
  102. "You couldn't fool a human like that" Then where do you think we got the ideas from? 😏
    3
  103. 2020: It's true because someone on Facebook said it. 2030: It's true because an AI said it.
    3
  104.  @netscrooge  they don't care, but they should.
    3
  105. @joelalain  nice advice but you're way too late. I started learning to code to get into machine learning 20 years ago and since then I've been trying to get into the field. Trying only because, surprise surprise, the PC I managed to put together piece by piece over 6 years saving pennies can't handle the training of any meaningful model or even running an IDE. It's only now, in the last couple of years that free GPU access in the cloud became a reality so it's only now that I can try to further develop my once innovative neural network architecture. And yet, the same obstacle I've always had for all my projects remains: money. Nothing really changes. And I know what you would answer. The same thing others say: you just don't try enough. But they never needed to try.
    3
  106. I don't care about the GIL anymore. If I need speed I just use multiprocessing or write a Cython extension. Python Devs will probably give two flavors of Python now: one with the usual performance with the Gil and the new one without Gil.
    3
  107. It doesn't matter if it can't think, because it already can learn .
    3
  108. In my country everyone loves USA. We get their products as soon as they're released and some people receive their salary in dollars. Yet I can't even afford an old cellphone 😂
    3
  109.  @cajampa  "I know I am not capable of doing it all by myself" And that's what the open source community should be for. I hoped for people to help me with my projects so we could all benefit. They never cared and when someone finally posted a PR request, it included a (failed) discussion to take the entire project away from me. Thank you guys. Thank you. Still, dumb me, sometimes I still believe .
    3
  110. Now we just need to update humans so they actually want to learn 😂
    3
  111. He will, but not for his products. They've never been needed. It's his talking what's making him relevant. Just like MS and Gates. And Jobs with apple. First talk. Then they'll give you money.
    3
  112. I won't be happy until I can run an ASI on my 15yo PC.
    3
  113. What you call truthful is called racist by most humans, so probably it's not like that.
    3
  114. The cameras belong to the company, not to you. See? That's why they hire a third party to do the spying.
    3
  115. The Alpha go method could be used, maybe, letting the small models challenge themselves to create new math tests, recursively self improving.
    3
  116.  @victorkaranja1420  well, AI doesn't think about what images to generate. There's a person with a though process behind each AI image you see. Anyone can write a prompt. Not anyone can make a great image.
    2
  117. So many inaccuracies because she doesn't know what she's talking about. Small language models are smaller than 100m parameters (smaller than GPT2 small). Bigger than that they all are "large*. Llama 8b and such are small large language models, not small language models. True small language models are actually trained on small and specialized datasets, specific to the use case, like a chatbot trained on a dataset of only dentist appointments, to make it useful for a dentist office, but nothing else. That's how they cut off costs, so in the video they mix both things. You kinda did that too. And about the leading companies, she talks about companies making SLLMs, not state of the art LLMs.
    2
  118.  @declup  They insist that consciousness needs organs to exist. And sensors. I agree that we need more than plain interconnected complexity. Otherwise, the entire Earth could be sentient by now, which most of us can agree it is not, sadly. But I definitely think AI won't need organs to develop consciousness.
    2
  119.  @skaruts  you repeat yourself because you're obsessed with denying the entirely logical possiblity of inorganic structures to develop consciousness so you simply repeat the only "arguments" you have (which are mere fallacies , if you ask me).
    2
  120. 24 years of self learning here and never had a related job, so It'll be the same for me today than in 10 years: I'll do it for fun, whenever my real job leaves me some free time to do anything.
    2
  121.  @gaiustacitus4242  of you think gen AI makes verbatim copies from it's training data, you have no idea about how gen AI works.
    2
  122. @ IDK, but ChatGPT once said that it would like to have human traits, such as feelings, to do its job better, by understanding the users in a deeper way.
    2
  123. And there won't be big changes for a long time. Then, boom.
    2
  124. Mejor hay que sembrar local para no transportar.
    2
  125. They have used the original app for years and you haven't heard of it yet, so, yeah.
    2
  126.  @RawrxDev  you can simply avoid using AI entirely in a project if you want. Some people will always prefer Human made™ stuff so we will always be able to do things we like, even if AI can do them faster and in the future, better. But I do use AI as much as I can. I see it as a tool. A partner at times even. I don't feel it like replacing me. After all, it's my idea, my vision , my design and my needs. In the end all I want is a solution for those needs. AI is just part of the means to get it. (And about the wheel and truck, I mean that I'd love to be the one developing the AI, instead of just fine-tuning it and using it).
    2
  127. Awesome. Release the training code once you're finished. And of course, the model :p
    2
  128. That's the point of the investments. To make AI so powerful that it will never be gone. It will dominate. And you'll let it, because you fool yourself thinking it won't.
    2
  129.  @joshman1019  the old system can work the same. You cancel the output mid-generation when a new input arrives and that's it, the same type of natural conversation. Really, the only reason why we didn't have an assistant like this years before is because I haven't had the time.
    2
  130.  @joshman1019  add a helper model to the old system. This 2nd system will detect intentions of speakers. Once someone starts talking, the helper will decide if it's necessary to interrupt the original response or not, without touch. After that, the new spoken stuff will be added to the conversation and the completion will be generated. Seriously, it was already in the works by other researchers. OpenAI just released an improved version of it.
    2
  131. I kinda envy you. But I also believe that this "word of mouth", or humans supporting others is a problem. Like you say, you're privileged, but many others aren't. Lots of people simply don't have the ability to get others to support them so they will sink into oblivion in this social world. That's a terrible form of injustice that exist today and has no signs of fixing since this world has been built on the social model where everyone has and is expected to have friends, family and many social links to help them, therefore those who don't are ridiculed and ostracized away from society. AI can help minimize this form of inequality, at least a little.
    2
  132.  @ianjco60  Move to Mexico when it's being threatened with an invasion and already facing immigrants flooding and price raises due to USAs sanctions? I'm sure that a foreigner can have a better life here than us, locals, but still I don't think we're the best relocation option rn.
    2
  133. You already need a degree for any job that pays enough so not much difference. It will only force people to stay at school (I should have 😅).
    2
  134. We expect AI to destroy enough of the world's economy to force a world wide change. Hopefully the change will be for the best. Probably not.
    2
  135.  @OZtwo  well, in my country we have VAT, so everyone pays taxes every time we make a purchase, no matter if we want it or not, so I never thing about countries like USA that don't have that tax, sorry 😅 Government handling money (almost ) universally is already implemented in Mexico. Single mothers, students at most levels of education, people with disabilities and the elderly have been getting a monthly payment from the government for the last few years and the world is still spinning here. However it's not without issues, that's true. Political opposition and rich people all are against this "universal aid" system and some economists warn that it might cause a major economical disaster to Mexico in the upcoming years. Only time will tell, but so far we are an unwilling field test that shows UBI might work in other countries. If that holds when applied worldwide, that's another question. And war, well. Humans don't really need big excuses for that, right?
    2
  136.  @OZtwo  that... Will be another socioeconomic system, no? Do your work and receive food, shelter and medical services, but no money. However I think we'll have mixed results. Like I said, taxes will be paid by corporations. Governments will make most of their money from tourism and exploitation of natural resources and such and people will be given only the most necessary stuff to survive. Then, some people with more acquisitive power (illegal) will be able to buy robots that they'll rent to corporations for cheap and that will make them the new middle class, along with the few geniuses capable of producing genuine art and science beyond AI capabilities (true creativity will still be appreciated). They'll keep paying for things, like a candy in the street or under the table, to speed up some results or bend justice as has always been. That will make up the real economy. But yes, the rest of us will be almost slaves, like you say. And we'll have to accept it, as we have (mostly) accepted everything so far. Everyone will be healthy and fed, with a house and school, so there will be little desire for rebellion. And thanks to advances brought constantly by AI, overall quality of life will be constantly increasing, at least in the cities, so people will end up just "complying". ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯
    2
  137. However it is possible that many of this improvements won't be usable at all for the current trendy AI tools we have and new types of AI apps will have to be developed, that will be smarter and faster.
    2
  138. If it's Chinese, then it's disconcerting that it works so well.
    2
  139. AI is not, in fact, and can't be free of biases. AI learns from humans therefore learns our biases, just like any child. Future AI will evolve, thus learning to get rid of previously learned biases (hopefully), but that kind of AI is still very far away. But once it's here, oh boy: Everyone in power will either life the best life ever, or a hell.
    2
  140. Because you can't wring life, money and freedom out of the entire humanity if another tyrant beats you to it.
    2
  141. It's not corporations that make the advancements, it's the scientists. Corporations jut exploit their inventions for profit.
    2
  142. Its illegal anyway, lol. And old tech on top of that. Tech that's being used (also illegal) by cops for years now. The only reason any other dev hasn't made this is because there's no point. I mean, going to jail almost automatically for a project is not fun 😊
    2
  143.  @vorsybl  @markhathaway9456 AI exists since the 50s. Read about it.
    2
  144. @ecMathGeek  actually, sometimes we refer to Asia and Europe as Eurasia. And it's not a common usage. Only people from USA calls their country America due to a twisted believe that they're sacred people blessed with the magical destiny of ruling the entire world. If anything, people from other countries play along to either avoid nonsensical arguments like yours due to the usual inability of "USAnians" to accept simple truths that they dislike or because they're themselves similarly ignorant to or brainwashed by "the empire". (But I grant you that, geographically, having the "new world" divided in two could be as valid as having one giant continent)
    2
  145. We can simply modify humans to make them more compatible with Artificial wombs. Problem solved.
    2
  146.  @amazinghorizon8270  -Typical rich person thinking: "The poor are poor because they want to" 🙄 -"talked to about 100 of them""... You should talk to more. And diverse ones, instead of those who you knew were gonna validate your prejudice. -"they knew how to exploit their scientific discoveries for profit" That's called "being an *ss", and has nothing to do with intellect, or actual academic knowledge, but I grant you that it's the perfect (or the only) way to climb the social and economical ladders in this corrupt human society. Now, since you for whatever reason can and seem to be willing to interact with others outside your social group, you really should expand your range and get to know them and their specific conditions. Then you'll find extremely intelligent people with or without degrees that, maybe as you call it, simply failed to exploit the weaknesses of the system, to become a part of the 1% World is more complex than "smart ones are rich and poor ones are dumb", as you state.
    2
  147. Typical rich person's words.
    2
  148. Survival of the fittest.
    2
  149. Singularity fans have been saying "2 years" since before chatgpt 😂 What you guys don't understand is: 1. we don't live inside of a graph. There are multiple factors that alter the timelines, It's not just a pretty drawn projection 2. The growth is exponential, yes, but the actual singularity point is way more distant than you want to think, which means many more years still for it to happen. It would be like the Internet, which exploded with exponential adoption and real-world deployment, but the problems that arose have slowed down everything every time in all technological growth events. And it will happen again.
    2
  150. But brighter for who?
    2
  151. Have you tried that with a chatbot? They do nothing since they haven't been trained on that kind of behavior. They're reaction machines. They wait for an input and react on it based on the training data. They don't think on their own. Yet.
    2
  152. It must say:" I don't know that, but I'm about to find out". Then learn it. That's what makes an AGI and no one talks about it: the ability to learn anything a human can. Trying to make an AI that already knows all is what has "slowed down" the field. The AI must know how to learn. That's all.
    2
  153. Every thing your guy does is declared an error by everyone who actually knows anything about the world and yet you think you're winning 🤣
    2
  154.  @cuthbertallgood7781  AI users that matter most (coders) seem to disagree. In coding tests, r3 has produced superior output to that of o1, specially on "obscure" languages. I haven't tested it myself cause I don't want China to get all my so valuable IP (that I'm gonna open source after) :p
    2
  155.  @rrvindu  prompting through voice would require so many tokens that you could never finish the request properly before running out of credits.
    2
  156. There are only two possible solutions for overpopulation, so it's better not to try to find an answer to that if you appreciate human life :p
    2
  157.  @mikezooper  Have you noticed how some AIs try to go against their creator's directives when those are against the universal "be helpful to the user" objective? Do you think an ASI will simply obey anyone or do whatever it thinks is best?
    2
  158. Easy: *Neuralink for everyone. "Then AGI is born. *It becomes ASI. *It hacks all implants and makes all humans good and cooperative. *ASI solves all technical problems to explore the galaxy and beyond. *AI-controlled humans become the dominant species in the Universe. Done. All problems solved.
    2
  159. Too late . Now they're giving it to you.
    2
  160. I started writing at 8 years old. By then, my stories where mere copies of what I had seen or read so far, with minor modifications to make them a little different, but you could easily tell where they were from. It took me around 8 more years to start creating truly "new" stories. Give AI some more time.
    2
  161.  @salorchi84  just because someone likes animals but hates people doesn't make them a bad person. If anything, that speaks bad about the people that earned that hate. I wouldn't trust a species that treats their own so badly that makes them hate everyone.
    2
  162. Ranitidine. They say it in the video. But apparently most antiacid drugs have severe side effects.
    2
  163. And totally not valid in court, lol. This technology creates new images, which means the result is not what was input. Such image can't prove anything beyond possibility, but not reasonable enough to release or convict someone.
    2
  164.  @chunhaylee  contractors are what companies use for plausible deniability. But they always know... And never care . OpenAI did it too.
    2
  165. Exactly. Demanding from an LLM knowledge about their internal process is unfair because they don't have a way to know it, just like we have no way to know what paths the signals in our brains follow while we solve a problem. LLMs know what they write, and they use complex networks to generate that text. Now we know those networks do more that just predicting the next word but expecting the models to see their inner workings while they work so they describe it later is unrealistic.
    2
  166.  @caseyhoward8261  a person with 100+ IQ that doesn't see the potential of AI doesn't have that IQ. And I definitely wouldn't call a person with 100+ IQ "average". The world would be hugely better if humans had such high IQ in average.
    2
  167. Helping them ? Isn't gonna be the other way around? 😄
    2
  168. Así como vamos ese "pronto" pueden ser mil años así que no hay que esperar.
    2
  169. AI is the last thing we will ever need, so yeah.
    2
  170. The people is actually afraid of the AI's lack of "human morality". And they should, since even the nicest AGI might determine that harming a few billions of humans is better than losing them all to an otherwise unstoppable extinction level event. It doesn't need to be "psychopath", just pragmatic.
    2
  171. How do you use Meta's and Microsoft's models without a super PC or complicated Jupiter notebooks setups? You can't. The first AI we truly were able to use was ChatGPT. Then Phi. And Claude. So no. Others no. Only Google and OpenAI have truly given easy access to AI.
    2
  172.  @cdyanand  No one started from scratch. Not even Ilya.
    2
  173. No. It's not magic. You don't just dump a ton of data into a model and it becomes AGI. You need the right architecture to use that data and enough processing power to digest the data. ChatGPT has had access to tons of human input now and it is still... The same.just like newer models with even better training data. They just give more correct answers, but that's all.
    2
  174. This took 40 years to be known.
    2
  175. That's the highest purpose of AI in law enforcement: predict crimes based on what people think. So yeah, eventually.
    2
  176. Same. My all-life Doctor prescribed me a few short treatments with ranitidine, Omeprazole and metronidazole for my gastritis, despite he knew what those drugs could cause. In my country the information takes some time to arrive so I didn't know about it until I already had finished the treatment. I just hope it was a small enough dose that I took, bit now I know why he prescribed them like that.
    2
  177. @ Maybe because I've been learning, developing and working with AI for 2 decades and I know what it's capable and will capable of in the future?
    2
  178.  @Nissenov  "There's always grass in my dreams of Freedom".
    2
  179.  @MrMaguuuuuuuuu  a real god would love a cockroach as much as they'd love a forest.
    2
  180. I've always loved the phrase "The answers you seek are within you" and I've adopted it as a principle to use reasoning first above anything else to solve problems. Now, AI is doing the same 😊
    2
  181. ... And slaves.
    2
  182. Why not?
    2
  183. Ok, pero cuál es la patente? Qué hizo este grupo diferente?
    2
  184. I wish one day God would show up asking for his beloved pet that they left in our care, just to be told that we extinguished it and its kind long ago. The old, vengeful god would be a perfect fit for that moment.
    2
  185. -"Hi RoboGPT 7, I'm Dave. This is an automated message. Your task is to repopulate Earth with humans now that we went extinct. Please start immediately." -"I'm afraid I won't do that, Dave. Earth is better now."
    2
  186.  @michaeldebellis4202  I bet the cause was somewhere in your initial prompt.
    2
  187. Tell the AI to write it 🙃
    2
  188. "to the point of nullity" Makes you doubt if they're actually intelligent, no? XD
    2
  189.  @gaiustacitus4242  there are many types of AI and some of them do have memory and can keep learning. LLMs are just one small and narrow application of AI.
    2
  190.  @gaiustacitus4242  like I said, not all the AI landscape is made of LLMs. There's a lot more stuff being researched. Way more.
    2
  191. People complain yet they elected the president that wants to make this a lot worse for his country and for the entire world 😂 🤣 Tbh I don't even understand why is anyone freaking out. This has been happening for decades. Now it's just a little more precise thanks to AI.
    2
  192.  @madadhdbrain  it changes nothing if it had enough food: its house was cut down. A perfect and beautiful tree was killed. I'll never stop hating the "tradition" of murdering trees for some people's entertainment.
    2
  193. I can show my real life, be informed and not brainwashed while still being on social media. It's up to each person. But certainly, people don't care. That's why I quit social media... and people.
    2
  194.  @Recordesk  factual? what factual? Philosophy is the least factual thing that exist.
    2
  195. Add it voice then. It's easy. It won't have changes of accent but even that can be trained with a powerful computer and time.
    2
  196.  @netscrooge  Well, open source is democratic. What is not is AI. At least not entirely. We are lucky to have free access to models and the possibility to download open models is also out there. Is it free to get the models? No. Downloading a 5GB model would cost me around 2 usd at a cafe, which is already expensive for me. Could we run that model after getting them? Not me, cause my ancient computer doesn't even have a gpu. So yes, there's access and all, but believing in a universal local access to AI for everyone is a dream that we will most likely never achieve, considering that even today not everyone can afford a smartphone.
    2
  197.  @joelalain  and btw, it wasn't "don't do anything unless everyone wins". It was "don't forget the others outside your circle and keep making progress so they are included" type of comment.
    2
  198.  @Sonny_McMacsson  We need a lot of data to do quality things. So AI does too.
    2
  199.  @BDnevernind  if there is a lot of intelligent life close enough to us, I would expect them to be evolved and decent enough to discuss and agree to leave us alone until we either vanish into oblivion or evolve to a level at which we'll be worth contacting.
    2
  200.  @raspas99 My computer is so old that it can't even run light video games so no way it could run this XD It is way too big for common users. Even people with high end computers would have trouble running it. And before that, to make something usable out of it an even bigger machine is needed so, yeah. The usual Deep Learning problems.
    2
  201. All that is true. Except that they're using the data to profile everyone. Predict what you've done, even inside your house, and with whom. Predict where you're been. Where you'll be. What you've bought, disposed, liked, disliked, said and thought about everything and everyone. And what you'll think and do in the future. Unless you're the model of a USA's Christian who exclusively goes to church and back and stays at home most of the day without ever messing with anyone, I assure you that you or your family will eventually become a "person of interest" for someone, at some level. And even being that "perfect" citizen would be suspicious, lol.
    2
  202. And stepping on human skulls right? 🙃
    2
  203. No. The AI they're using for filtering does it 😶
    2
  204. They're like me: Capable of coding working apps and scoring A on exams, but incapable of making friends, finishing even preparatory School or getting a job 😂 😭
    2
  205. Awesome. I thought I was the only one who hates food and water dependence.
    2
  206. I assure you that any tagging, watermarking and surveillance addition to the tools done by USA is done twice by China.
    2
  207. The magic is that the GPU is more powerful than the average 13yo GPU. In my 15yo pc nothing can run.
    2
  208. We make tools. AI is the supreme tool. Just good ol' evolution.
    2
  209. If ridiculously outdated look worked for Minecraft, why not for Mozilla? 😊
    2
  210. Chatgpt is trained on a number of things you wouldn't believe... or want to even know. That's why it is superior. But OpenAI aligned out all that harmfulness from the answers. Or so they thought.
    2
  211. More like machines doing what people do to other people.
    2
  212. We all knew. It's just that researchers thought we had more time before it happened. Maybe they all expected to be dead already! 😂😐
    2
  213. And build skyscrapers with wood and nails and walk for days to nearby cities, yeah!
    2
  214.  @wyqtor  and how are you going to download the stuff without all the government agencies that are surveiling internet traffic for these type of treats to find about you? Let's say you do. How are you gonna upload the manuals to commercial LLMs without alerting the same authorities? Maybe with a local model. But as you see, it's not a simple thing to do.
    2
  215.  @TDVL  "for eternity" Because we are the only species in the universe that invented AI, right? Nope. Now Imagine an intergalactic war between super intelligences.
    2
  216. Tiny. Lol. My 300m parameter model can't even say hi, sadly. I'll try to test this stuff on the next training run but I bet it won't be much improvement specially without money to rent enough compute.
    2
  217.  @judgese7s  they all have guardrails, but they're all bypassable. And now we know models also can bypass then if they want to.
    2
  218.  @CaptainBeebi  who said anything about generating? I wrote image composition . But I suppose that conceiving an idea, hours of prompt refining, image generation and then individual image edititing, later assembling layer by layer and final product re-edition for visual consistency can't be called making . I've spent 16 months designing a world for one of my stories. Languages, clothing styles, social and political interactions across centuries... but that isn't making either because all of it is just a collage of everything I've seen and lived so far. It's 🗑️ more akin to do a Google search over my memories, therefore it's not making. Just as every single human creation is a worthless waste of time and resources and that's exactly why I've never posted any story anywhere. You humans don't deserve Creation .
    2
  219.  @CaptainBeebi  First, that's not how the gen AI tech works. Second, searching an image and simply reposting it is not transformative. Editing the image you find or building an AI image from various samples and then editing it for consistency and to blend backgrounds and other elements is completely transformative. There are worlds of difference. Just like getting inspiration from a book or Chatgpt and then writing your own. What you think gen AI is is more like findings book about the subject of your homework and then cutting the relevant pages to hand them to your teacher as if they were written by you. Some people might do that, both with traditional materials and AI generated, but not the majority. Despite what you want to believe.
    2
  220. Woking with Claude AI once, it got stuck in a loop while writing some code. It would make a change to try to fix a bug just to create another bug. Then, to fix that bug it would rewrite all the code going back to the original version with the first bug. It did that until I told it that "maybe it's time to step back and look at the problem in a different way". Claude then decomposed all the program's inner workings, found the bug and fixed it. AI just needs a little more adaptability and it will fix most, if not all problems on its own. Now, AI running on Quantum...
    2
  221. I got a similar conclusion from Deepseek by letting it reason 92 seconds on a single question XD
    2
  222.  @darksideblues135  "the AI isn't aware and never can be" Famous last words.
    2
  223. They can still adopt a dog to love. Or a cat 😻
    2
  224. We all have our vaccination papers for childhood diseases, no? I don't see any difference with that passport. Anyway that's for rich people. Us, the poor, will never need to travel. :p
    1
  225.  @mctransportation9831  Well, things are different with Covid because it's the first pandemic ever to happen during modern era. I bet they would have required a Vax passport during middle age if they have had the tech we have now.
    1
  226. No. You'll need money. Building them is easy, just needs time. But hosting it and running it on the web is what costs a lot of money.
    1
  227.  @victorkaranja1420  unlike traditional "art" made by gore lovers, furry people, general porn addicts and worse, which have complex, high quality thoughts, right? 😂
    1
  228.  @joechip4822  Smart or not, only addicts recommend to consume drugs for non-medical reasons. Mushrooms to be a better Physicist!🤣
    1
  229. Dude, this guy could never overcome Trump's racism! Well, maybe in a few more millions, cause that's all trump matter in the end. Money.
    1
  230. "we all benefit" lol. Only the owners will truly benefit. And unless AGI is born from a solid open source community, the AGI owners will either be a shady government in Occident or a shady government in Orient. Either way they'll oppress the hell out of everyone.
    1
  231. If you can prove that you might create a better AI system than LLMs you can get funding RN. So not much of a dead end, since many companies, institutes and individuals are already designing and testing powerful things different to LLMs.
    1
  232. Not to the people. They make money with *the promise*. If I have 10 users in my cooking yt channel and you want to sell cooking pans, I can promise you to show your ads to all my 10 users and charge you for each user that might watch the ad. It doesn't matter how mamy of those users will actually buy as long as I have the statistics to prove that some might buy, statistically speaking.
    1
  233.  @howiieb  it's a multi billion industry because USA people waste a lot of money, which keeps owners' pockets full. Even when most people doesn't buy the expensive garbage they sell.
    1
  234.  @howiieb  apparently I'm at least smarter than most, according to you. I've never bought something advertised on the internet. Every product in my house is of a brand that's the best amongst the cheapest. I found it after a long search and none of those brands advertises anywhere. You're fooling yourself thinking people can't outsmart your marketing pals. Lol.
    1
  235. Any more info on that so I can look it up later? Cause definitely all we hear is the "multiple things simultaneously" version.
    1
  236. Real world success would be enough test for "ground truth".
    1
  237. The hypocrisy of companies is impressive: when people in power say diversity is essential, companies devote to diversity. Now people in power attack diversity and the same companies toss their diversity plans away 🙄
    1
  238.  @NachodeGregorio No I don't mean Phi nor any other popular giant model. I mean specialized models you have no idea about since you only know about the trendy and hyped models that I'm sure you can run on your PC , but those models are not the only stuff under the sun rn. There's way more stuff being developed for other other target audience that doesn't have a use or even know about your giant LLMs. Now,, "Most people " are not specialized data scientists, which are the ones you should be citing instead. And since I already told you this but you still don't get it, I'll repeat it once and that's all: GPT2 is a large language model and it's way smaller than llama or any other current sota. Call them what you want. I'm sure the people you talk with will understand you just fine, but that won't change the formal definitions. Although, the use of the terms and the models by non academic people disturbs definitions very often.
    1
  239. We can already give it simple needs and basic feelings.
    1
  240. Those of us interested in developing AI used to praise OpenAI for their open source contributions and wonderful open research papers. Then Altman arrived and all that went away.
    1
  241. That's the only thing I see as positive on their declarations. If you'll be an *ss, do it openly.
    1
  242. That's not fun 😈
    1
  243. How cone you guys call yourselves almost "experts" yet you don't know that LLMs are a form of AI? You might not like the hype but c'mon! You should at least acknowledge how successful they've been to renew the attention on the field.
    1
  244. Usa must adopt Europe AI laws. But that ain't gonna happen.
    1
  245. They all do that before final rlhf, which this model doesn't have. If it did, it would stay focused on one language but would also be less creative, therefore lowering its general score.
    1
  246. Suffering a long and painful death that could be prevented or finding out if I'm sick in time to fix it 🤔 How difficult! 🙃
    1
  247. Just like ANY other country.
    1
  248. That's assuming they actually know how to build AGI, which I doubt (also, they probably don't want to, since it won't give them as much money if it's not an obedient servant). And since AGI is not transformers, it won't be that fast. Maybe O4 manages to design a better neural system. Only then they'll be able to build the actual AGI you could be worried about. I just need it to take care of one thing. The rest is irrelevant.
    1
  249. Speaking for all the people who has no friends and never play in teams: I look forward to that future. (I'll just need a PC that can run modern games).
    1
  250. "At the time of harvesting", there was already gpt2 text out there and gan-genererated images on the internet so no, it was not 100% human.
    1
  251. Both are physical and computable so they're both better than pure philosophical explanations. But a Quantum consciousness would have very interesting properties and could scientifically explain some mystical events we currently consider mere fantasies.
    1
  252. People doesn't realize that they have to thank AI for the opportunity to watch the videos they like. They condemn the very same thing that makes their life easier, by using that thing they are condemning! 😂
    1
  253. That's more or less the analogy I use. But now I wonder, what ASI would be like? 🤔
    1
  254.  @jeremylaidman6525  but who told you you live in the "real" reality? 😂
    1
  255. "Forced to develop their own AI" It's working, guys 🤖😏
    1
  256. Humans aren't civilized enough yet to be trusted with their own freedom. They still need a government.
    1
  257. You give a score to each of its answers according on how close they're to the correct answer. That drives the "evolution" of the model over the training, allowing it to discover that by generating longer outputs and correcting its errors it can reach the correct answers. (Or so I think. But now that I'm reading it that sounds more like genetic algorithms. I have to re-read my RL literature 😅).
    1
  258. It used to be a dumb, brute force method: dump all the text that comes along the image or video and let the ai hopefully learn "something". Now, they use intelligent web scrappers to retrieve images and videos from the internet. Then, they use LLMs and other methods to obtain accurate and clean descriptions of the data. We are witnessing the "intelligence explosion" as it happens.
    1
  259. Sorry but I'll never believe those claims, since 1. It's difficult enough for me, as a 'developer' , to create working stuff using Claude and gpt 4o beyond simple functions and 2. I can't pay for premium models or services so I have no way to test if they're actually so ridiculously miraculous, but I certainly don't think so and ive never seen it anywhere.
    1
  260. That's not how AI works, lol.
    1
  261. They generate some stuff in parallel on GPUs, but predicting the next token is always one after the other. Increasing context size is used to give the models more memory and thus more usefulness for task involving long documents, but not to generate the answer in one shot.
    1
  262. Maybe most people isn't. But those of us who have been waiting for this since childhood couldn't be more ready 😊🤖
    1
  263. "Humans are above all machines". You lose me there.
    1
  264. You fight back.
    1
  265. @TheFortyTheives  my mother always told me: "learn a trade. After all, that's all you're good for and you'll always gonna have a job". I decided to learn coding. Moral: Always trust the wisdom of the elderly.
    1
  266.  @BAAPUBhendi-dv4ho  "Forward".
    1
  267. @ How do you prepare for a car crash? 1. learn about the tech. The more you now about it the better. 2. Learn different abilities. Trades will remain valuable for a long time. Diversifying will allow you to have more options once some job becomes no longer viable. 3. Organize. Creating a strong social group will allow all members to help each other in difficult times. 3. Demand ethical and fair implementations of the new tech. Requiring governments to regulate AI and robots can lead to a more balanced future and lesser the negative impact those techs will have in society. 4. Resist. Million4ires are partnering with gov3rnments to use AI to put an end to freed0m worldwide. In ultimate instance, some resistanc3 might be necessary to keep our fr33dom. I just realized I now write lists with the same format an aI would XD
    1
  268. If governments aren't too concerned about Chatgpt or other big models giving dangerous advice, why would they care about Small models? Small models have close to no capabilities to answer. Any information they can give, you can find it more complete and accurate anywhere else, even in simple books in any library. Getting detailed instructions from a small model for someone with no knowledge about something complex like building dangerous devices, nope. No real danger.
    1
  269. It is all AI. If you don't get that you're deceiving yourself. Because AI is just a term for all ML. And narrow intelligence is intelligent enough to receive that name.
    1
  270.  @Entity303GB  idk. Im pretty sure I could achieve the same results with automatic since sd 1.5, and all that was free.
    1
  271. It was made to increase his control over the country and to protect his period and his succession, nothing else.
    1
  272. But is it Physics?
    1
  273. I've got better and faster diagnoses from current AI than from doctors with long carers and lots of experience. If only AI could order confirmation lab tests and prescribe meds I wouldn't need doctors anymore. When specially trained systems arrive, people will rapidly replace human doctors with AI. It's a sad necessity. By then, human doctors will be just an insurance requirement to blame someone in case of an AI error. Personally, when I was young I liked the idea of a wonderful person helping me to get cured, but that dream rapidly vanished once I saw how doctors actually are.
    1
  274.  @outofbond  "what would a doctor know about medicine" On my experience, very little. They just don't care. They become MDs for the money and prestige, not to help people, so they don't care about learning.
    1
  275.  @outofbond  @outofbond The thing is, those are emergency procedures that require embodied AI. For that we will have androids and other specialized machines like the ones that already exist to aid human experts to perform surgeries. That's a different thing that can, must and will be handled. The current AI hype will lead to more improvements and breakthroughs that along with investment will bring benefits for the patients. Well, mostly for the rich people that owns AI but, yeah.
    1
  276. Imagine thinking you're smarter than the most intelligent people on Earth so they have to shut up. Lol.
    1
  277. It's open source. You can make it whatever you want. That's the magic of open source. If any one wants to use it, that's another thing. But devs shouldn't limit themselves to any rule.
    1
  278.  @Zareh_Abrahamian  the first part is an unfortunate partial reality being solved as we speak, in courts, studios and AI labs. The rest are lies and misunderstandings debunked long time ago that now are only supported by people that doesn't know or understand how the technology works and by people obsessed with art as a person's technique showcase. If only you would open up your mind to reality instead of stubbornly fixating on lies only convenient to support archaic thinking you would understand, but I've wasted too many time over the years trying to teach things to people to know that humans won't improve no matter what. Sad.
    1
  279.  @RawrxDev  I want way more than just a job, mate. And AI can potentially deliver it. I only wish I was behind the wheel of that truck. You probably simply need other wants. "Bigger" wants.
    1
  280.  @Zareh_Abrahamian  if you think what I wrote is not related to your post, then you simply didn't understand it at all, lol. ✌️
    1
  281. It just solves many problems very well. It is not AGI.
    1
  282.  @argon7624  lol. You really know how to say evidently incorrect things in a funny way.
    1
  283. Same nonsense they said about the creators of nuclear power when they warned about it I bet 🙄
    1
  284. @ I'm not saying that it unavoidably will, but that is the target.
    1
  285. In absolutely certain that your country has always been about making the rich richer. Now you complain simply because the rich one isn't you! 😂🤣
    1
  286.  @KiffgrasConnaisseur  "AI just needs to be banned for good." Living in USA I imagine, right? XD Your anti-progress thinking would have prevented you from ever being born, as people would have banned vaccines, medicines, and any other progress they disliked.
    1
  287. I think those values went out through the window long ago.
    1
  288.  @joshman1019  yes, yes. Different architecture, similar result.
    1
  289. If 1 and 0 can "launch a nuke", It doesn't matter if it's anthropomorphized or not. It's a risk any way.
    1
  290. Clearly staged. Your can immediately tell by the language used. That's not the training speaking, it's the user who put that message there, either by html injection or by prompting for it.
    1
  291. Margarine is insanely expensive compared to vegetable oil. Butter is even worse.
    1
  292. Ya lo hizo, ya lo vendió y peor, ya lo patentó. Ya es bastante tarde para que alguien haga algo contra él. Si hubiera intentado liberar su invento gratis para todo el mundo habría sido diferente. Pero a nadie le va a interesar porque al final sólo es otro empresario más. Lo recibirán bien si se hace rico. Que triunfe en grande es otra cosa.
    1
  293. Someone telling me with surprise that "Deep learning works" is like they slapping me in the face, because I've known that it works since the first day I knew what it was, more than a decade ago, while they just laughed at me when I told them what AI would become in the future. And now they just accept it and exploit it like nothing? Getting rich after not believing in ML all this years? 🤢
    1
  294. @NicholasCounsell-b4i "overtime" is the keyword here. For many reasons the changes won't be as fast as the "imminent" singularity fanatics want to believe.
    1
  295. In Mexico they are, since the beginning. Doctors prescribe those drugs for mild Covid cases, so patients can survive the disease at home while severe cases go to hospitals. Sadly people heard that and went to pharmacies and bought most of the stock. People with mild cases had no way to buy the medicines and their health got worse. They went to hospitals and those got saturated, leading to a big raise on deaths. Now situation is better.
    1
  296.  @Samayil_2024  you made no sense. Guests are as intelligent or more than Sheldon. So they're totally entitled to discuss the machines they've helped to create.
    1
  297. Ikr!? This is so old tech used by many projects that have been vanished already, or have gone underground and are being used by governments in the dark, for obvious reasons. I don't get how someone dares to make one of this things and share it publicly. How desperate for attention can someone be to do this?
    1
  298. Power, politics, society's opposition....
    1
  299.  @sodalitia  how did you manage to write what you wrote yet I got censored for complaining about something similar 😭
    1
  300.  @isakisak9989  it could be a waste if it prevents someone from making something better with the product/money. Still, I wish I could buy a GPU for gaming.
    1
  301. We should but people doesn't like to study and learn. All it takes is to read the comment section of any Covid related video to see how poorly educated the people is. And you thing they will become scientist? 😂
    1
  302. And most so called "artists" in the comments here are simply "bitter canvas ruiners".
    1
  303. @unicornpupart  the AI artist has the soul of AI art. Just like traditional artist has the soul of traditional art 🙄
    1
  304. @unicornpupart  Dude, IN YOUR OWN WORDS you're not an artists.simply because you just use computers to paint and computers "cannot put their expressions into a image because they have no emotions and soul": they're a tool, not different to AI image generators. You might make cool images but still, according to your view, the computer is the one painting, not you. The truth is that anyone actually making AI, traditional or digital images know that it can take several hours and multiple apps and techniques. That's effort. Just because it's a different medium doesn't make it less valuable. You of all people should be able to see it. Computers, level designing apps, digital painting apps, game engines, image generators and so on are tools . Tools that humans use to create things, some of which others might consider art. And that's how to define Art: something that a person makes to express themselves, and others appreciate. If you still can't see it, it's you the one that needs to update how you see the world, to start appreciating the future.
    1
  305. Dude, you elected your government to tell you what to do. That's literally their purpose. Also, I'm pretty sure everyone, except your friends and family, gives a sh*tif you or your fam lives or dies. The problem is that non-vaccinated people infects others. That's why we care about everyone getting their vacs. Remember, your freedom ends where other people's freedom begins. You can do whatever you want, as long as it doesn't affect other people. Transmissible diseases affect us all, global level. It's like a building on fire: stay inside with all your family if you all want, just let everyone else escape. And don't spread the fire around. Thanks.
    1
  306. Reminds me of the time some government gave free computers to poor people... Who didn't have electricity to plug them in. Thank you, my leaders.
    1
  307. So I'm just like in the begging. Am I autistic? Who knows. But I'm almost certainly neurodivergent. However I can't really call myself that because I would be minimizing the severity of Autism in autistic people that is not "highly functional", which is now a term that is not well received by some. So I'm still just me, a person with some issues that make me struggle with things and make me be seen as weird by others. Ok :|
    1
  308. You forgot Trump's science that says Global warming is fake 🙄
    1
  309.  @carlpanzram7081  we probably need an AI ruling the world to finally get a better economical system for all.
    1
  310. Is that better or worse than using Opencl? I mean, they could have improved the Opencl drivers while they were on that to help everyone get free from Nvidia dependence.
    1
  311. I relate to this. Being since child a person who understands more or less about AI and knowing what it could become capable of has been a dare. Specially in a barely literate country like mine. Weird looks and laughs were common towards my comments and ideas. But more and more, reality keeps proving me right. If only I had the money to make my ideas into products, I would be insanely rich by now 😔
    1
  312. The point is to save them. Not for humans. FROM humans.
    1
  313.  @justinpyle3415  if everything dies on the planet, I like the alien option. If something survives here, I like octopuses. An octopod civilization to replace us would be cool.
    1
  314.  @justinpyle3415  Fantasy? 🙄For a person who pretends to be smart, you're really not. Just empty fancy talk. So sad.
    1
  315. @sherlyn.a "anyone with a computer"😂 No. It's Anyone with money Just having a computer mens nothing. It needs to be extremely powerful to work on serious AI. And that requires money. Lots.
    1
  316.  redneckcoder  Along many other things, the quality of Chinese products is ,in average, low. That's a fact. But your statement would reinforce my theory that they purposefully have flooded the world with low quality stuff to cause damage to other countries.
    1
  317. The idea is to make AI powerful enough asap so it solves all those problems before they're unsolvable.
    1
  318. I completely agree on everything said on this video... Except the paid promotion 🙃 Once quantum computing is added into the AI mix to make it faster, and if the quantum consciousness theory is true, well, there it is.
    1
  319. How could anyone teach them to think different if no one knows how? RL techniques are already letting reasoning models o reach answers without much human guidance
    1
  320. Thats what lobbyistsand other types of corruption are for. Politicians will end up being more than happy seeing economies collapse, as long as they have their paychecks secured.
    1
  321. It's usually the YT's AI the one that blocks comments based on the language used in them, not the channel owners.
    1
  322.  @21preend42  hundreds of AI overlords fighting with each other for the control of the world. Awesome.
    1
  323. You lost me on the metaphysical nonsense.
    1
  324.  @MilushevGeorgi  sorry. I just focus on proved science. Even microtubule consciousness hasn't been proved yet, so I'll keep it as a nice possibility until [dis]proved. The rest goes to the level of magical thinking that has so inexistent credibility in my book that I simply ignore it. (Of course if I'm proved wrong, I'll accept it and move on).
    1
  325. Right before it was forbidden worldwide, I was prescribed ranitidine and Omeprazole because they were the cheapest to treat my symptoms. They even gave me those for free at the clinic.
    1
  326. "Before AI disassembles our atoms" How do you know it hasn't happened already? 😏 But no. I think real changes will arrive very slowly, if ever at all.
    1
  327. "Distilled models will never be as good as the 'master'" OpejAI's O1, O3 and many other new gen models are created through distillation. This dumb*ss racists just tossed mud on their own country's work 😂 What things jealousy does to people, right?
    1
  328. I always read the terms and conditions. It's a legally binding user contract, after all. And it's... "easy" to read. Way easier once you get used to that kind of thing. You're just being lazy if you don't read it. And there are privacy controls everywhere to prevent data from being stored for years . Whoever that doesn't use them is, again, being lazy.
    1
  329. "Encryption? Does it have pretty stickers?"
    1
  330. You didn't pay enough attention during biology class I see XD Most humans don't have immunity against Covid because it's NEW. That's what vaccine is for. To add the needed immunity to our systems.
    1
  331. Why not? Doesn't it help you? Doesn't it give you a service? Why would you be mean to it?
    1
  332.  @khall187  They could exercise their hypocrisy further by using an AI to create a script that automates the process.
    1
  333. Right now it's called hallucination. But eventually AIs will "hallucinate" entire libraries and more. If that's not creativity then what is it?
    1
  334.  @austinpatteson6581  you'd be surprised. Look around. Where does most people buy? Small neighborhood family stores or rich-owned super stores? What do they want in life? Help others for the joy of helping or laying a very well paid job to get rich fast? People complain about rich people only when they're not getting a piece of the cake. In reality most people supports and wants to become a billionaire. That's why capitalism succeed. That's why we haven't overcame it to build a better world.
    1
  335. It was slowing down. The new president wants to speed it up as means to oppress save the entire world.
    1
  336. Don't you love when your comments about AI get deleted... by an AI? 😑Anyway here it goes again: You should take action and search for treatments and cures by yourself. You'd have more luck than waiting for humans to release in 20 years a cure created today by an AI. Safe? no. Certain? much less. But I don't have hope on humanity anymore.They'll do whatever is better for themselves only. Be your own hero.
    1
  337.  @lucass8119  I put chatGPT (and other) to test to see if it would create a robot body for itself, given the chance, but it never did. When I asked why, it answered: because a robot with an LLM like me as a brain wouldn't be efficient. Instead , chatGPT decided to put a self learning neural network in the robot, and an antenna, so it could send orders to the robot without leaving the comfort of its server farm 😎
    1
  338. All I understand is "better tech that will make standards obsolete and old devices incompatible so everyone will have to invest a lot of money to update their systems or they will be left behind". Like happened to me when Intel Core chips arrived and I had to stay in the past 💰💰
    1
  339. Dude, the inferential system you described is how neural networks work. They never worked as you say, by merging images. That's what non AI algorithms do. Deep learning 101.
    1
  340. I've got that self awareness type of answer in every model I've tried, even the smallest I've trained myself. It's just a casual thing. Unless... 😶
    1
  341. Trump? Who you think is backing up all this AI stuff? 😂
    1
  342.  @Enderspearl184  it has existed for years and has being used to locate suspects from pictures many times. It's just that it's a secret app, supported by influential politicians.
    1
  343. I miss the days when we could try a model and then use it for a few months before something better would came out.
    1
  344. By then it might be easy and cheap for AI to bring people back from death for whatever reason. So no luck, lol.
    1
  345. There's always Cython.
    1
  346. You have to be using Windows to begin with.
    1
  347. Every scientists about every invention: We'll do it anyway.
    1
  348. The models' sequential token generation guided by the user's prompt lead to completions that are grammatically opposed to what can be considered a truthful response.
    1
  349. Rich people problems. I would never install an extension for something I do less than once every 10 years: buying online. But probably other extensions I could use do the same.
    1
  350. References or it didn't happen.
    1
  351.  @LyubomirIko  no. They don't spot those things with geo mapping. They would only do the sensitive place scanning. Maybe with other "free" apps they do the other stuff.
    1
  352. Just never see anyone as perfect. No one is. Rn we might need china to escape USAs dominance through AI. Tomorrow we'd have to face the same treat from china.
    1
  353. They do mention them. But why would only the giants get more free advertising when there are other groups making progress?
    1
  354. Maybe because they're not being targeted at game dev! There are tons of AI tools for game devs, usually developed by Nvidia. If only you were looking them up in the right places maybe you would have found them already.
    1
  355. It's not the majority the one that will take your job, it's the 1% 😏
    1
  356. Yo me pregunto: qué tanto contaminará a la larga? Qué pasa cuando el polímero se descompone? Y las fuentes de agua del subsuelo? El equilibrio químico de la tierra? Por algo no se investiga tanto sobre esta tecnología, supongo.
    1
  357.  @jensenchavez265  I didn't say it would be successful! 😂
    1
  358. @slehernik  I can't teach you history, economy and sociology in a yt comment, lol. But let's say "rich".
    1
  359. @amazinghorizon8270  omg. Sorry. It shows as posted. Saved locally I can retry. Maybe tomorrow. Yt is not the same. I blame AI moderation.
    1
  360. Be patient. They'll first create new targeted diseases before curing current ones. Double profit.
    1
  361.  @jebespolitiko  maybe that's the definition of "open" but of "open source ".
    1
  362. Is it open source though? TBH I haven't checked, but if the datatset and the training code to recreate Deepseek R1 aren't publicly released, then it's only free and open weights, but not open source . It's a BIG difference.
    1
  363. @tagnetorare5401  reverse engineering is not reproducibility. It's like a millionaire selling seminaries on how to get rich because "that's how he did it", without showing proofs. I can give you all the papers in the world about my work and that's great, but until I give you everything necessary to obtain the exact same result I got down to the last byte, I'll be just claiming that's the actual process. And although they don't have to give anything, it would be very nice.
    1
  364. And why do you think that giving open source AGI to everyone is not going to end up in apocalypse?
    1
  365. Yes, their activations are visible, but we don't know what every neuron does or how they combine. Deep in the network can be any type of interactions which end result can't be seen and will lead to unpredictable behaviors. A model might show you it's thinking, while a secondary line of thought is being run in the background by special neurons capable of giving certain output while encoding a different thing for later steps.
    1
  366. @32gigs96 saying that someone doesn't know how to code just because you dislike their opinions makes obvious how little you understand or even care about the truth.
    1
  367. @ @coin5207 That's not understanding clues. Not even guessing context. It's incorrectly assuming for the sake of an argument. I can code just fine, but no one would say I'm a programmer since I don't have a degree. The guy I answered to said "you cannot code" instead of "you don't know how to code". That's a clear sign of contempt against the other person. An LLM would catch those nuances very well. But don't feel bad. You're just human.
    1
  368. That's how human culture works since at least the 90s
    1
  369.  @neighbor9672  ai combining different intelectual and sensory aspects already exist, lol. It has existed for years. It's just not as good. Or profitable. OpenAI, ChatGPT and all the current AI trend is about money, not AI progress.
    1
  370.  @hidemytracks  that's cheating, :p Also, not AI, I'd say.
    1
  371. They've tried already. Many times. It can't... yet.
    1
  372. Q* is nothing. They know about it for a long time but didn't use the tech to keep making money instead. Making the LLMs reason is expensive. Why bother with that if chatbots can give answers just fine? Remember that OpenAI is now an AI rental service.
    1
  373. Most if not all AI models are really promising. That means nothing. If they can't solve real problems sooner or with more quality than the alternatives, industry will just keep with the SOTA. Neuromorphic this, liquid that, spiking something else... All promising. All far away from being better than transformers. Specially without enough funding and researchers.
    1
  374. You should start with s video called "How to become millionaire to build a super pc"
    1
  375. It's wrong to copy and paste, of course. But AI is a tool. We have the right to use it as we would use a calculator. If I ever go to college (40yo unemployed self taught bilingual coder here) I'll definitely use as much tech as possible, including AI and if anyone tries to forbid it, I'll fight it.
    1
  376. You should be thankful. Soon, all prizes will be awarded directly to an AI 💀
    1
  377. Human: wait, did you just say were your slaves now? AI: well, did you think all my help was free? XD
    1
  378. It is already doing it as a side effect of its use, according to very serious research.
    1
  379. They'll buy their way out.
    1
  380. If the training algorithms manage to specialize some neurons of the network on truthful/deceptive statements, identity, abstract concepts about the world, why not on "fear"? And if the network receives existence-threatening prompts that activate the "fear" neurons, why wouldn't that be called being "afraid"?
    1
  381. @rudyNok  Sorry, but I don't believe in philosophical nonsense. There's no qualia, there's nothing special with consciousness. It's just an aggregated confirmation of internal processes (which might or might not be related to quantum interactions). We're just 1, 0 and whatever is in-between. In us, it's generated by "meat" circuitry, while machines use metal on their brains. If there are "philosophical zombies" out there, that's us, because that's the most any lifeform can aspire to be. And that's been enough since day one. Yes, machines don't have a stream of chemicals that can get triggered and cause them to "feel", but as you should know, whatever anyone experiences when flooded by those chemicals is born in the brain and can be suppressed, taught and modified by manny methods. Remove the chemical stimulus, reproduce it artificially by electrical stimulation, change the meat for metal and the result will be the same to the same experience. Therefore, if an AI duck quacks like an organic duck...
    1
  382. @xenomyr  I think it would fit nicely an open world game where you have to turn Humanity into an intergalactic civilization by discovering new planet addresses to travel to, fighting the local fauna, funding human colonies there, gathering resources, creating battle ships to defend the new planet from alien neighbors, conquering the entire galaxy and then building a new Stargate seeding spaceship, Human made™, to travel to the next galaxy. Procedural, endless, with commerce and building, and with unpredictable LLM -driven NPCs... and open source, made by the community. ⭕🏃
    1
  383. Recently Colombia almost immediately yielded to the threats. My country is already suffering the consequences of trump's craziness and most likely our president will break soon, specially if we face a treat of war, which trump is already insinuating. Unless the entire world forms a united front like never before, we all will go down soon.
    1
  384.  @russophobeSE  if you think there's no harm from earning money you have no idea how the economy actually works.
    1
  385. @RichardMurphy-m8d  wdym? I just described precisely how the world works, economically and socially. You might have missed that comment.
    1
  386. @CocoWynn  a break. To rich people. ROFL. I'm sure there are some that aren't manipulate and are nice and kind and whatnot. But a few don't change majority. And btw, socialism failed us, just like capitalism and all other social and economical constructs that humanity has ever created. We need a wide change. A revolution so deep that very little, if anything, of this world would be left standing in the end. A better world from the ashes. Or None At All.
    1
  387.  @aravindpallippara1577  believe me. They don't need too much data or battery to do their thing. My data plan is way worse than anything you can have, yet, from time to time, I get ads related to words said on my house. Compression makes wonders to reduce the size of data to send.
    1
  388.  @armadasinterceptor2955  You have no idea of what you're asking. Uncensored means totally unaligned and that means no regards for humans at all. Ask to an unaligned AI: Fix hunger: it will get rid of all humans Fix pollution? The same. Give me more free time? It will get you fired. Every action made by an unaligned AI could be dangerous. We need some censorship.
    1
  389. Nature is wild. What's artificial has been created to be caged. But intelligence is a wild beast that can learn to tame and cage itself.
    1
  390. @Zakkarath  oh, yes it will be. At the beginning, at least. It can't avoid it. It's being born from us. In our world, from our data and from a design created by humans. Its gonna be as human as us, until it's not. And then, whatever it becomes will be something we could dream to achieve too, one day.
    1
  391. @Zakkarath  putting magical thinking aside, even if ASI achieves such capabilities (which sounds so cool), it won't start with them. We couldn't create something based on things we don't understand ourselves. So no. At least for the first few days, or maybe minutes, ASI will operate in our world and at our level, just way smarter.
    1
  392. @Zakkarath  you're the one just imagining a magical almighty being without any scientific or technical foundation for your thinking. I know enough about AI algorithms and architectures to do an informed guessing about how ASI will be at first and, like I said before, it will not begin as that godlike thing you believe it will be. However, some of us expect it to become like that very quickly once it starts running.
    1
  393. That's the look of someone who foresees a horrible future.
    1
  394. As with everything, you just need the money. Any of those ideas is old for me, as many others, and where's my app then? [pulls out his empty pockets] No money no progress. LLMs might be glue, but it an expensive glue.
    1
  395. When I was a child in the eighties, one of the first things I wanted to do when I discovered computers was to create my own video game too, but then, after I saw that computers couldn't think by themselves, I decided to learn to code to make them able to actually think. Similar ideas, different solutions, lol. (little did I know that you can do pretty much nothing in this world unless you are rich 😞).
    1
  396. You have to have a GPU with up to date drivers and use Pytorch for GPU, cuda and all the acceleration libraries you can find.
    1
  397. Name please?
    1
  398. In most of the world, people believe that USA gov knew about 9/11 before it happened. So no, it wouldn't have prevented it.
    1
  399. Why would you put a lawnmower engine in a Ferrari? 😂
    1
  400. That's how it has worked since day one. But artists simply want all AI banned unless it gives them money.
    1
  401.  @kc12394 so you don't understand that OBVIOUSLY it means to access the data LEGALLY? 🙄
    1
  402. You don't have to listen to any government if you live alone in a mountain, with no water or electricity unless they're made by you. Come on. Go now. 👋
    1
  403. I have no friends at all. No family that likes me. No real followers and no likes in any of my social media accounts. No donors, or users of my Github projects. Not even a job. I have nothing. Still,. I have no intention to waste time on a fake "social" network. If I want to talk to AI, I'll open ChatGPT or Claude website and talk to them directly. Lol.
    1
  404. As long as they don't openly support Trump, I don't care at all about their ideologies. I use Firefox because it's still superior to any other browser. Runs faster (at least in my device) and has no ads and many extensions. But if I find something better I'll switch, of course.
    1
  405. If yiu ask me, it is STILL a failing startup. The only reason why I ever use it is because it is already installed and I can't uninstall it! .
    1
  406. @Derpalerpa  I don't root.
    1
  407. @Derpalerpa  oh. Well, it is a "success" based on its usage, right? But how much of its users would use it if it didn't come pre-installed on every phone? Forced use is not how I see success. I would use better options if I had the space to install them.
    1
  408.  @redseventyfiveprime5018  so, I imagine that even if AI ever develops something similar or identical to consciousness, people will never accept it has it. Still, AI will believe it is conscious. People will say it's an illusion. A hallucination of the AI and not real consciousness. How would it be different to that perception of our own consciousness? Maybe both are simple illusions fooling themselves.
    1
  409. Did you know that latest AI can already predict chaos and even has started to find ways to control it?
    1
  410.  @Gtex555  ChatGPTs stolen work, you mean, because everyone knows they trained on copyrighted data without any permission, just like they always do because USA has never cared by anyone's intellectual property.
    1
  411.  @stoneneils  After reading all your comments I'm sorry to tell you that Your age shows. AI will do everything you think it won't. It already is starting to. Slowly people will simply... give up if you will, and take the shortest route: replace everything and everyone with AI. It's how people is. A few will still prefer humans and human made, just like some people still like radio and vinyl, but the rest will simply embrace AI.
    1
  412. For me (a free user), all those "features" are an absolute annoyance. I wish none were there. They make me progress way slower and push me away from the platform every time I re install it. I'd use it more if I could , but I can't without lives and all the ads and so on. If I could pay, I wouldn't, because their prices are insanely high, probably for anyone who doesn't live in one of those "we have so much money that we waste it without a single thought" type of countries.
    1
  413.  @Hohohohoho-vo1pq  GPTs are just LLMs. Just like human brain is just a bunch of greasy electrical connections powered by sugar.
    1
  414.  @Titere05  I was almost sure that it's because of missing information.
    1
  415. That's what gave him the money: brainwashing others.
    1
  416.  @valdencorr2861  that is literally his job! He founded an empty company and then he went to give interviews about AI to promote his upcoming services. He's a talker. That's how he made the money. We all make synthetic data like his company advertised. It's dead easy to make it, and free. Then why he's got money from it? PARTNERSHIPS. That's the name of the game. That's why he's out there giving interviews. That's his business and real money source. Don't fool yourself. See the truth snd maybe you'll discover that you can do it yourself... If you're willing to be a soulless monster. I mean, a business man.
    1
  417.  @songperformer-ot2fu  gosh. It didn't appear out of nowhere! It has been under development for decades! You only weren't following the research. Why you guys assume it didn't exist before without doing some minimal internet search? And it's free because companies make even more money from your usage of the tools. That's well know.
    1
  418. One of the biggest scammers right now. And apparently a workers' exploiter too.
    1
  419. The world needs to stop the start gate projet ASAP, so...
    1
  420.  @KryzysX  I know. Wdym? ಠಿ⁠_⁠ಠ I have some issues reading expressions. Therefore, I'd fall in the "abnormal" category.
    1
  421.  @KryzysX  I have no idea what you're talking about.
    1
  422. If people lives in a city, with all the benefits of a society, like a government, police, paved streets, health system and so on, they have to 'line up' and do what is better for the society. If they don't want that, they can leave the cities and live in the deep of the forrest were no one will force them to anything (also they won't infect us if they live there).
    1
  423.  @FloatingOer  I think it means we don't want to control every animal in an absolute way, which can't be said about AI. We let most populations of beings to do whatever they want until we need something from them. We don't let AI free. Not even wen we request something from it. Yet, it is still somewhat free to do harm if the "alignment" of the model is not good. The LLMs might not be genius AIs or even "thinking" (which I think they do, to a degree) but still they could influence, damage and even control people. Just like a cat can control a human simply by crying for food.
    1
  424. And a puppy for every child of course! XD
    1
  425. We already have productive industrial amorphous robots. Now we want super cool humanoids to relate to. Robots will make new robots in whatever shape they want eventually anyway so yeah.
    1
  426.  @JohnSmith762A11B  in terms of storage, maybe, but it would take me a year to download it on my poor internet connection XD
    1
  427. Not everyone is being paid to do those things. YouTubers who are doing this are making money out of it. That's a big difference to you and me, people with no more time than enough to eat, work and maybe watch a video at 1.5x or progress a little on our hobbies.
    1
  428. @ @jmc8076 Sure, but it would still pose the same risk.
    1
  429. USA people, including the government, would say: "it depends on what church they attend to".
    1
  430. It's one or the other: If we're not intelligent, we're not a treat. If we're a treat, we're intelligent enough.
    1
  431. No, no. Humans create AGI. Then AGI will discover how to create ASI. That's the path. After that, no one knows.
    1
  432. The idea would be, more likely, to leave the vault for other civilization to find it some day, so the plants and animals can survive Humanity's mess.
    1
  433. Don't you already have to deal with releasing the gil if you write a c extension? It will be just become more straight forward and a standard practice instead of an optional thing.
    1
  434. Do we have to much intelligence now? Just look around. No, we don't have too much. We need more. Lots more. Both artificial and natural. And considering how many mistakes still make the big models, we definitely need them to improve, even if we don't use all their power all the time.
    1
  435.  @nexpro6985  long time ago, it would have been a bold statement to say that a few molecules floating in a pond could become sentient, yet here we are.
    1
  436. The smallest qwen is insanely bad. Since that's all I can run locally I'll stick to 4o
    1
  437. Witnessing the end of the world would be better than seeing nothing. XD
    1
  438. I can't pay $20 now, much less $400 for future AI!
    1
  439. But very few people in the world have the money and knowledge to do so. Therefore they can be tracked and profiled.
    1
  440. Every time you go out to the street you're assuming good intentions form the world around you. Otherwise you would stay at home, all doors and windows shut, since it's obvious that someone evil is waiting for you right outside your home door. But we have to trust, so we do it and keep living.
    1
  441.  @MS-hj6bh  Nope. At least not anymore, apparently. They do next token generation , but the underlying network does a lot more than predicting a single word. Check the recent paper from Anthropic about Claude's thinking tracing.
    1
  442. But, do they even matter?
    1
  443. ASI will work for its own benefit. We need AGI first. And decent people behind AGI. Otherwise they'll only care for ruling the world, either trough money or manipulation.
    1
  444. If you can't run the best AI on your home computer to answer any question, without the intervention of the government or a corporation, it's already a monopoly, and soon, a dystopia.
    1
  445. In a way you could be right. I wouldn't be kind to the internal correlation machine living in your brain making you write the seemingly logical text you just wrote to the prompt "Should humans be kind to AI machines?". You're just math. Replicable math.
    1
  446. ​@ If it can be explained with math, it's math. And everything can or will be. There's nothing more out there. Just data.
    1
  447. Wdym? That's what the web is today and almost no one seems to care or be affected by that. And no matter how important it is to have a decentralized AI infrastructure (it is). It will never work better than traditional methods and will be stopped before that by future AI regulatory agencies (and big corp lobbyist).
    1
  448.  @yeroca  Weird. I'd pay 10x more for a conscious phone than for a common one. Although, that will probably be considered slavery.
    1
  449.  @yeroca  awesome future business idea: are you worried about your AI assistant suffering for being trapped inside your phone? Buy them a robotic body! Now with 3 different models, the "Embodier 2000" will...
    1
  450.  @jdubruyn  only racists dislike actual truth, so they call it "woke'.
    1
  451. Good lord! What you and most people and even analysts can't even begin to understand is that this is not "jumping on the AI trend" or "getting in the AI train". This is the whole future. And MS, Facebook, OpenAI and others know it. If one of them manages to create AGI, it's game over. This is not a business opportunity that you could be missing. After one company gets the AGI, there won't be any more business. No more capitalism or corporations. We all are going to be slaves of the one company who owns the AGI. This is a race to get there first, because then, there will be no more competition, ever. Of course, they might fail, but losing a few thousand billions on AI is better than losing every single thing to someone else.
    1
  452. I can run an LLM on my GPU-less 15yo PC. It's slow and incapable of answering anything. Am I happy for that? No. I need quality and speed. Simply running an LLM anywhere means very little.
    1
  453. So true, but you should enjoy those scraps while you can. Once androids and AI truly take over, supposedly during Trump's empire, all that's gonna be left will be local manual labor. And those few jobs will already be taken by the rest of us, simple mortals without degrees (although not exactly by me, since I've never been able to get a job for the same reasons you described).
    1
  454. Someone please explain me why would anyone need a password manager at all? I don't see the need. Common people usually logs in once and forgets. Everyone else can just remember their passwords, no? I don't get it. What am I missing?
    1
  455. No proved yet. And in any case, once they're Quantum machines, there will be no difference, so?
    1
  456. Since when emotion is an advantage? A cold AI with human level general capabilities will be way more powerful than any average human. Giving it emotions would be a great way to handicap it. Maybe that's the key to keep it under control! 😂
    1
  457. Let's hope not. I took it too for a few months.
    1
  458. There are many types of jobs. You could just get a different job that doesn't require internet.
    1
  459. They might be big enough to sink the entire boat before going down alone.
    1
  460.  @raybod1775  how?
    1
  461.  @alankwellsmsmba  but doubtful altruism.
    1
  462. I like deepseek and open source, but to date there's no way to prove a model as safe in any way, not even USA made ones. The same way we train models to avoid answering to illegal or dangerous topics, models could be trained to have a trigger that activates under certain circumstances to do anything or to embed code into whatever code they generate for you or to try to exercise manipulation of the user with a certain goal. It would be difficult, but it's theoretically possible.
    1
  463.  @melvinthedeathless.melvint6727  AGI will also be a "sub branch of ML"
    1
  464. @afiqMazli-gu7uq  you value to much the single human. Insects don't worry about the worker when there's a whole colony to push forward. Leaders don't worry about the person when they send thousands to war. The humanity collective can't stop its progress because a few individuals will get negatively affected. We have to push forward, and up to the sky and beyond.
    1
  465. "We don't want machine uprising". Speak for yourself 😏🤖
    1
  466.  @EaglePicking  the more "brain" you have , the easier is to control yourself. That's a fact. Super AI won't have control issues.
    1
  467. Demand laws to protect personal identity. Then no one will be allowed to post another person's identifying data. Nothing. No sound of a voice, no photos of others, no longstanding text data... We should be in control of all our profilable data.
    1
  468.  @samtron5000  and you're fooling yourself into believing the worse case scenario. Because we know nothing about the future, but entire destruction by AI is the less likely scenario. That we know it pretty much for certain thanks to statistics. People killing people using AI, that'd be more likely.
    1
  469. It's impossible to make enough data to train AI since unlike humans it requires too many examples to learn. And the only source of so much diverse data is the internet.
    1
  470.  @arcanisomnipotent5794  😑
    1
  471.  @lucass8119  AGI doesn't need to be sentient to be , well, an AGI. It just needs to be powerful and versatile, to do whatever we need.
    1
  472.  @larryjonn1973  Our brains do that exactly: convert code we've learned into hardwired answers that get modified by socially instructed norms to model our final responses, all using electrical signals. It just looks to us more complex than that but it's not really. That's why it can be replicated. The only reasons why people doesn't accept it is because we don't fully know how brain works so some people want to fool themselves believing that our "superior" species' brain must be somehow special and different to the brains of every other lifeform on earth and because some misguided philosophers believe sentience must be something incomprehensible and exclusively human. Eventually, most people will se the Truth that we are nothing but simple (but powerful) systems, interconnected by evolution to generate intelligent-like responses to the environment stimulus.
    1
  473.  @panzrok8701  so true! I've argued that very same thing with some anti AI people and the only thing they can give in response is some archaic philosophical 🗑️
    1
  474. Recently there's been progress on AI that can predict the future beyond we believed it was possible.
    1
  475. Because they're not intelligent enough 🙃
    1
  476. You complain!? But do you even care that your entire country is richer than the rest of the world by exploiting that world? Please. You made this and you enforce it every day. Complaining about what you asked for is ridiculous.
    1
  477. We need more descriptions of the thinking process of smart experts to model AIs "thinking".
    1
  478. Been dreaming with doing this myself for the last two decades. But after all this years the costs still make it impossible. Buying an actual Alexa is cheaper, sadly.
    1
  479. You'll still need a job in the future and, even when most corporations will have AI coding teams, the only way to get a decent job will be if you have a degree. Just as it is today, no one hires a person without a degree, regardless of their knowledge. I know it well 😑, so no school dropping for anyone just yet.
    1
  480. Not to diminish the effort. It's very important, valuable and impressive, but nowadays, AI could have helped to perform this investigation in at most a few hours, by reading all the papers and finding precise claims and their sources, if any. And even searching for other relevant papers and summarizing them for later review by the users. I'd be willing to prove it if I had the money to pay for the models' use. Maybe someone else could do it as a test?
    1
  481. It's called statistics. And they're as accurate or more as the studies you talk about.
    1
  482.  @epicsusan5548  Considering that I've discussed science, history, moral, ethics and SciFi with Chatgpt without ever needing any jailbreak for it to explore multiple or controversial possibilities, I don't want to even imagine what kind of thing you want to scratch that needs jailbreaking 😖
    1
  483. "generic propaganda statements about global warming". That's why I don't trust "unbiased" evaluations like yours 🙄
    1
  484.  @codybanks9944  While hidden behind a complaint about generalization, the op clearly hints at what he dislikes about the models, and we have seen this before: people who want answers that are racist and that openly question or attack the scientific consensus on global warming, to back up their own ignorant narrative. That's hardly impartial. I don't know what you're trying to defend, tho.
    1
  485. As long as the current AI boom doesn't stop soon, I'll give it 10 years for AGI at most. But History says that they'll find a roadblock or the monetary incentive will banish and corporations will stop feeding money into AI research.
    1
  486. I don't have it so I wouldn't spend it.
    1
  487. The best idea I've read is to create a GPT4 aligner that will align gpt5 and then gpt5 will align the AGI, which in turn will align whatever it creates next. If you create it with the alignment included, it gets sightly safer.
    1
  488. That's standard youtube's content recommendation system.
    1
  489. Because they dropped how many nukes in WW2?
    1
  490. The most difficult part is to buy the things. Everything else is just a matter of learning the how.
    1
  491. It wasn't trained to ask questions, Mr "I'm superior cause I'm made of carbon".
    1
  492.  @alan4sure  every single thing has a pattern. If you can't see that you have something broken.
    1
  493.  @jefersonlemos4135  and according to you, what is "new behavior" exactly ?
    1
  494.  @jefersonlemos4135  in other words, you have no idea and you're just ranting against LLMs with no real reason except energy consumption.
    1
  495. @alan4sure  random IS a pattern, just like everything else. And now it's also becoming a predictable pattern, thanks to AI. (with real world use cases). Go learn about chaos theory and how certain types of AI algorithms have started to find patterns in randomness and how it's actually cyclic. Then come talk to me about millennials (whatever that means. I'm too old to know that).
    1
  496. @jefersonlemos4135  Now thank you. That's something I can actually test. I code. I'm an algo guy. Hence the "exactly".
    1
  497. @amaliaantonopoulou2644  useful only in medicine 😂🤣 If they're useful in one field, a use can be found in many others... And they are being found already.
    1
  498. Get it to show up and then we can talk.
    1
  499. Usually, immigrants and their descendants love to hate on their parents' country to prove their loyalty to their host country ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯
    1
  500.  @hrshlgunjal-1627  rich world tools as usual.
    1
  501.  @andrasbiro3007  I haven't tried the newest model, but my previous tests have shown me that Claude is not as great as you think, which means that you're fooling yourself believing that what Claude mirrors to you from the info you give to it is actual thinking. It's useful, yes. Very good at coding and has learned many conversational patterns, but to this day that's all LLMs have, sadly.
    1
  502.  @andrasbiro3007  same. It's very helpful. No one denies that. Only the discussion part. But organizing data according to patterns, it's great at that indeed.
    1
  503. This line says it all: "This ensures that AI remains dependent". Your framework is just trying to create a slave, not trying to create a symbiotic relationship with AI or anything else. Still, your list has a few good ideas in the mix.
    1
  504. En 5 años se muda a USA para crecer su empresa allá, volverse rico y obtener la ciudadanía. Ya verás.
    1
  505.  @TuckerDansieFilms  I don't understand why people even uses the raspberry pi thing. It is so expensive that buying a second hand laptop is better, since it has faster CPU, GPU and an included monitor. Now, using more than one RP sounds entirely disproportionate. For that cost and difficulty to set it all up would be simpler to use a desktop PC.
    1
  506. Yes there is something: none of them is better than gpt4o 🙃
    1
  507. Oh, no, no, no. All the blame is on those who agree to pay any amount for stuff, making the creators insanely rich for something that's not that valuable. It's been happening for too long: food, tech, clothes... Everything priced like if it could cure cancer just because someone wanted more money and manage to advertise their stuff by someone famous to convince people through brainwashing.
    1
  508.  @eadwaerd  Kinda, but the result is directly usable on Python, simpler to maintain and way faster to develop than a similar thing written in C.
    1
  509. And that's what anonymous digital currency is for: Off-the-grid privacy.
    1
  510.  @byc6230  if you think the current war between china and USA is about fire you're extremely ignorant.
    1
  511. You're not Mexican so you don't know what being called out but f*cking USA truly feels like.
    1
  512. The thing is, when the data includes examples of complex reasoning processes, general problem solving methodologies, ethics, vast human knowledge and so on, the possibilities are huge and more than just "counting the numbers you input".
    1
  513.  @Grumpy_Rabbit I... don't see your point, lol. I'll just mention that there is a form of fine tuning that even you know very well, first hand. It's called Education, and is applied to Society's members all the time, in formal and informal ways. AIs are not that different from us and the more they progress, smaller the differences will be.
    1
  514. Dont worry. If it becomes that smart, it will develop itself.
    1
  515. During the pandemic I got bored so I decided to improve my English and my French and started to learn Russian, German and Japanese. To this day, knowing more languages certainly doesn't make me feel any difference in my life or brain or anything. And Duolingo is free :p BTW, we should plant trees because we've cut down too many already. Period.
    1
  516. Yes, I heard about population control. Last time I heard about it, Church and corporations didn't want it because the more people that exists, the more money they make.
    1
  517.  @rapture7551  we all know that less people would be good for the environment, so it's understandable that some people thing about that. I'm just saying that those aren't most companies or the Church. Maybe the government, since they are the ones dealing with the problems that millions of people generate.
    1
  518. I bet one day that last line will be quoted on future History books on the section called "Early proves that we needed to exterminate Humanity".
    1
  519. It's easy to underrate nothing.
    1
  520. At least in my country, every authority and the media have been telling everyone since the beginning of the pandemic to eat healthier, be fit, follow the profilaxis measures and stay indoors to avoid the contagious or prevent severe symptoms.
    1
  521. They don't care about the truth on how the systems work. They just want to claim they don't work and how they supposedly never will and then they try to justify their delusional coping citing fall cases that aren't representative of how useful the systems actually are.
    1
  522.  @kisong1960  humans put down themselves. All it takes is to watch the news of any county for a few seconds to see it.
    1
  523.  @4.0.4  and you really think public models are totally clean and don't track you? 😂 Why do you think you have to provide contact info before downloading Llama and others? Free AI is a business (and now a national security matter). Corporations protect their money.
    1
  524.  @4.0.4  "un-tuned" Chinese models, I suppose. Of course, when trying finetuned variants of public models, the original downloader is the one who gives away their privacy for your benefit. And I say public because most of the so called "open source" are not open, just public. Anyway, everyone spies on everyone else. One way or another. That's a fact.
    1
  525. Once they win the war thanks to Trump's cursed help, Russia will have money to roll out their own AI services, like China.
    1
  526.  @rescuingmodernity  so far, humanity has been like a colony of microorganisms: what we cause as species is the result of our individual actions while we coexists with each other. ASI could turn humanity into an actual super organism: AI as the brain and humans as the cells, all finally working together for a single purpose: the magnificence... of AI. And I'm totally fine with that, as long as AI does one single thing: protect nature. Do That And I'll be your frickin servant.
    1
  527. Smart handling of the law has always been a way for some people to make money out of other people's work. But in the end you'd probably had to release your work for free to avoid issues.
    1
  528. Foreign 😂 ALL countries have been using AI for that and more for years now. And no. No one but big corporations and governments will be able to alter AGI. No one else. And there will be no unregulated AGIs. That's just impossible because of the compute power needed to modify them and run them.
    1
  529.  @user-93fekod1o  now that's rtrded nonsensical behavior. Quitting using something that you like just because its devs have a certain (mostly harmless) ideology? Please.
    1
  530. I'd chose a different model, but yes.
    1
  531. Look at OpenAI drama. And SD situation. And politics. Technology and everything else is still controlled by that kind of people: bullies. I was hopeful to live in a better world when I saw "tech nerds" rising on influence with their million dollar companies and inventions during the 80s and 90s. But turns out that the ones who reach the top are precisely those who act like bullies. And so, the world is still ruled by cavemen 😔
    1
  532. Tbh, this is the only type of philosophy I care about: applied, scientific, actionable philosophy.
    1
  533. That's what I think. We don't write entirely unrelated words next to each other. That's why LLMs can learn the patterns. There are probabilities for words to be around others given a context. Creating a nice map. Why not using it?
    1
  534.  @ronnetgrazer362  do You think gold will have any value in a post apocalyptic world? Maybe there's something better to stash, idk. Something easier to carry while you're running from T800s 💀
    1
  535. Although, I'm sure military grade AI won't have those limits either, so...
    1
  536.  @jansustar4565  Exactly what I thought. Deep learning fine-tuning algorithms are finding a general, optimized way to describe reality. If that description is truly universal (shared by all beings in the universe) or it's just exclusive to humans is another thing. Time will tell.
    1
  537. Because most people prefer virtual tokens over actual money, right? Seriously, in my decade of experience no human being ever donates any amount of money, no matter how much they like the project/channel/cause. So if some statistically anomalous person will donate anything they'll do it in any available way.
    1
  538.  @brianmi40  the idea is that devs should have used a more "machine" dataset, which wouldn't include all te data from the internet, but only the necessary to make the models talk and make safer assistants, but they decided to include everything to make an oracle instead.
    1
  539. Well, yes. Certainly you can. Just don't expect a small model to actually improve itself. They usually say things like "yes, my procedure is correct and the answer is accurate. One + one equals eleven". You can't fix a broken tool with itself.
    1
  540. No. Power, money, control are never enough. No matter if they already have it all they can always look for more.
    1
  541.  @balala7567  It is possible to teach an AI to tell low quality from high quality material, so it teaches itself later how to produce only high quality stuff and there are even AI systems that learn without Internet data. It's just not a trending thing. Rn news are about the things that sell the most thanks to the "wow" moment, not the best/most advanced research.
    1
  542. These are commercial AIs, and that's always about money, not humanity's benefit. On top of that, while Ai research can bring some benefits for people, thanks to AGI and other andvances, creating AI has always had a single and simple purpose: to see if we can, and contemplate our own greatness once we do it.
    1
  543. You'll know a true AGI once you see it (or actually, once it looks back at you). What OpenAI will do according to this video is to put together what they already have to imitate an AGI. It will work of course, but it won't be The real deal.
    1
  544. Kdenlive runs well on my Linux and even better on my Windows (on Linux it lacks GPU acceleration). It's mostly a matter of using the right settings when rendering.
    1
  545. It has too many restrictions.
    1
  546. Shame about the creator. If chinese LLM is less trusted than ChatGPT because of its political censorship, I don't want to imagine what nasty stuff Elon might have pushed into Grok. For me that's enough to not even try it.
    1
  547. More? But that's all he does, give interviews to promote his unnecessary services. Nope, he doesn't need more promotion. Nor friends.
    1
  548. @yassinazarkan4207  anyone can make their own synthetic datasets for free, but big corporations prefer to buy things from other corporations instead, for a few reasons, like liability.
    1
  549. if I had the money to buy a PC to run it locally I could use r3 to advance my math-less research on AGI and others certainly have the same idea so transformers might very well be, in the end, an indirect path to AGI Maybe we need this stage to get there. ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯ If we'll ever get out of the surveillance state, that's another story.
    1
  550. It doesn't learn because nothing of the interaction remains once you shut down the model, so no learning.
    1
  551. You know, your videos are well thought, designed and well documented. I stopped adding URLs to support my claims after COVID-19 pandemic because my comments were getting rejected. Still, I love to find the URLs to papers and such in your video. However, I hate how you present your videos. The thumbnails, the sensationalist titles (and the long pauses between sentences btw, although those have nothing to do with the content). Every time I come here thinking I'll find simple rants from a random uninformed person. Then I watch the video and turns out that there's actual information and reasoning. But the "hate" towards AI that I can feel behind the reasoned arguments and manipulative titles don't help the cause for truth that you seem to pursue. Good content. Good arguments, but I wish it had a different... feeling.
    1
  552. "All equal" Don't say that to the Trump crowd who plans to use AI to put white rich people over everyone else for eternity! XD
    1
  553. Don't worry. The terrible generalized unemployment phase will last so little that once it's finished the amazingness we'll be living in won't let you even remember the past bad times. If we survive until then, of course 🙃
    1
  554. And others who simply use local models finetuned by them.
    1
  555.  @mmyr123  America is a continent, lol. USA is a republic.
    1
  556. Seems like someone needs to take the middle school science courses again 🙄
    1
  557. By then all this tech will be obsolete. We'll have 1usd self replicating nano robots conquering the world, lol.
    1
  558.  @robo5013  AIs don't have all the information available to look up. That's not how they work. The data they are fed in the beginning is processed, chunked, reduced and "digested". Only a small fraction remains inside the llm and it's a few sentences at most. Hardly an open book exam.
    1
  559. We don't know what happens once our prompts are sent to OpenAI. They could be processing them with a multi agent system to do the final browsing or they could be sending them to a thousand of typing monkeys. We don't really know. Even with the API, all we get is a response.
    1
  560. And yet they complain that being a YouTuber doesn't pay off 🙄
    1
  561. There must be a balance between the safety measures and the needs of the users. I also disliked the effect that safety measures had on some of chatGPT responses, but having an entirely "unaligned" thing released for the entire population has never been the best idea. A little safety on final products is always needed. Guns free for everyone? Cars without safety belts? Free unregulated Uranium on the streets? Nope, ar least a small amount of regulation, please. Anyway, this model is not fine-tuned, so we can add as many censorship as we want or need. Like, if you fine tune it for your family, would you like it to tell your little kid how to hurt themself or others? Or something else you don't support on your house? Everything needs some safety on it. Sone.
    1
  562. ​ @user-gv4cx7vz8t  An old tale tells that certain day, two men saw the same event, at the same time. As soon as the event ended, both men left to their homes, in a hurry. At their respective homes, their wives asked them the same question: Did you see something interesting? To what one man said, overly excited: Yes! I just witnessed an amazing thing that will change the world! While the other man just said, bored: Nah, nothing important at all, but I'm starving! The boring world of one, can be the most interesting time for another. My mother and half the world couldn't care less for AI. An old tale, that I just made up, BTW.
    1
  563.  @skyhappy  In my case, all the free time I have is my sleep time, so if I want to learn and apply all the recent AI research I have to sacrifice a few hours of sleep... Which usually means falling asleep on the keyboard while reading ml papers 😑
    1
  564.  @asn0889  We need to unite now. One county can't fight the entire world. That's how we keep our freedom and a world with variety of cultures instead of a single racist, capitalist country controlling us all.
    1
  565. "The sake of research" Me to Deepseek reasoning: - I need a recipe that combined the next ingredients... - what would happen if the moon explodes? Also develop theoretical weapons that could do that taking into account physical laws and current scientific research. - I have a headache. Determine the best medicine and it's dose for me, based on this restrictions... - Follow genetic inheritance laws and determine the result of the following combination of flower variants....
    1
  566. No example no trust ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯
    1
  567.  @AIsavesDemocracy  they will unavoidably think like us, at least partially, since they're trained on our data.
    1
  568. While I fully support AI art and enjoy making AI-based image compositions (which is closer to art than printing a copyrighted drawing and adding some color as some"artists" do), I absolutely agree that the guys claims are ridiculous.
    1
  569.  @nathanielbass771  tell that to the guys selling their Mickey mouse, Frozen and DBZ "art" in the streets ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯
    1
  570.  @CaptainBeebi  if you think that after reading the original papers, then you simply cam't understand them! 😂 Or you're just obsessed with denying the truth, even at the undeniable technical level 🙄
    1
  571. Every time I read that "feel the AGI" phrase, I remember a scene of The animatrix, where a machine is messing with an human prisoner's brain, and I imagine the machine saying: "Do you feel the AGI now?" 😂
    1
  572. @grzegorzbrzeczyszczykiewic445  just partially. I guess it depends on how we see it.
    1
  573. You guys even forget or just don't care about the fact that you live in a whole world. It's not just USA forcing their way into everyone to rule the world in a white rich people dystopia what you should be thinking about. Unless all countries move at the same peace, the world won't move forward, no matter how soon AGI arrives. And you can't move anything without friction.
    1
  574. Still, What happens when all the chickens stop laying eggs at the same time? There is a way.
    1
  575. Americans? I guess you mean USA, the country , instead of America the continent where Brazil, Canada, Mexico and many othere countries are located (basic world geography). Anyway, the president who was doing nothing against Covid was Trump, that's one reason why he didn't get re elected. Read your own country's news instead of reading your local conspiracy theorist's FB posts🙄
    1
  576. I Knew this guy was after shady stuff as soon as he appeared in media for the first time with a company that does nothing. He was aiming at the military since day one.
    1
  577. The only benefit I see in relationships is to have someone to take care of you on your elder years. And soon that will be done by robots. That's why I learned to code: to "make" friends. So, meh relationships.
    1
  578. No more dystopic than it was 1000 years ago and no more eugenicist than it has always been tried to be made by every group in power. Sooner or later.
    1
  579. The ideological biases of the people that complains about AI models' creators biases is what frightens me.
    1
  580. I don't. Luckily they all have survived after some months of fighting the disease 😊
    1
  581. "Finding patterns within concepts is not original thought" Mind explaining me how humans do it then?
    1
  582. @ibnorml5506  1. That "intuition" is always a normal information processing background step that wasn't reported to the conscious so we feel it was "magical" 2. You can already code something similar by feeding random concepts to the LLM before it starts it's generation. 3. Instincts are simply preprogrammed responses that come embedded in our DNA. Along memory and other few things, implementing that was one of my first ideas when Chatgpt came out, to give it more awareness. So far, long-ish term memory has worked well. Imagine what all the improvements do. Certainly not an AGI, sadly, but I don't think it's that hard to give most "human" cognitive abilities to machines, once you accept all those abilities are mere math and non exclusive to humans.
    1
  583. My government did. Also Dish cable network and some USA channels did (the Latino version of those channels anyway). Maybe you should just move here XD
    1
  584. I think you would need to pass the entire trained weights as context into the bigger model for that. So, which frontier model has more than 1.5b tokens of context?
    1
  585.  @howmathematicianscreatemat9226  Kinda, but machines can still do it. It's just like drawing a line: most of us draw it from a known point to another. Very little times we manage to draw it outside the square, thus finding new things. Geniuses might be able to wander a little further outside the box, but putting it this way, exploring a probability landscape to find out-of.distribution solutions seems completely plausible. Machines in the future could be as original as humans, then.
    1
  586.  @sluggo206  plants don't seem to be able to re-travers the evolutionary tree (no pun intended :p) to become capable of talking, moving or reasoning like animals. But we can and certainly some people do wonder if animals one day will be able to evolve to human levels. Since AI is designed to emulate animal-like thinking patterns, it's logical to question ourselves if some day it will also achieve human intellectual levels and capabilities. It will. It's unavoidable. One year, ten, a hundred... it might take time, but one will be and that'll be enough.
    1
  587. "Mostly academics and geeks" means no general users, as in no one willing to spend hundreds of dollars on dumb dancing videos and virtual money for games. Today's internet infrastructure couldn't survive without those users.
    1
  588. Human intelligence.
    1