Comments by "" (@qoph1988) on "Big Tech in panic mode... Did DeepSeek R1 just pop the AI bubble?" video.

  1. Lol I have always said that the very people bragging about this hypothetical future are the ones who will lose their jobs first, and immediately get extremely butthurt and backtrack on their smugness and brashness.
    265
  2. Okay well give it a couple years and see if this one is still open-source either. Also it's good practice to wait until another party actually replicates the work to believe that something is fully open source and doesn't have hidden dependencies... And a trained AI model isn't exactly "source code," it's inscrutable and more comparable to a binary blob. Also there is some pretty compelling circumstantial evidence that the development cycle of this model is not actually what they are revealing to us. It likely used orders of magnitude more resources than stated, so it's not like a non-state-actor is going to replicate the work anytime soon.
    157
  3. It will go right back up when people discover what hardware was ACTUALLY used to train this thing. Right now they are almost certainly lying about the resources that went into it
    108
  4.  @paperbob_1  Nearly every country practices far more forms of global trade protectionism than the US does. Do you think that the US invented tariffs or something
    62
  5. I wouldn't believe that particular factoid just yet. Evidence points to a lot more training resources (state-level) than is currently being fessed up to. Still an impressive model of course
    41
  6. You should really wait for a bit more info before you declare them the heroes. OpenAI is definitely a villain, but I would wait a minute to see what shakes out with this one...
    22
  7. Plot twist: this is the first AI developed exclusively by a different AI
    7
  8. I will bet you that we eventually discover much, much more went into this model than six million dollars.
    6
  9. Another thing to think about: if some guy with tiny resources was able to make this, then this isn't "state of the art" and there are secret models that you do not yet know about that are far far in excess of what this thing is capable of
    4
  10. You should be EXTREMELY skeptical about this statistic until it is exhaustively proven. If something sounds too good to be true, it usually is.
    3
  11. Haha... I'm in danger
    3
  12. I think you are lacking in imagination if you think this was some kind of pure act of altruism
    3
  13. Weeeellll... wait until we figure out exactly how many smuggled Nvidia processors were ACTUALLY used to train this thing. Because it's a lot more than they are admitting
    3
  14. Next word prediction isn't going to fundamentally change the world
    2
  15. You're still being extremely premature in saying that this will "kill" OpenAI. OpenAI is basically a state-subsidized institution now, and they will pour a lot of resources into it to compete with this model. Will they succeed? Time will tell... This game is still in the very early stages after all.
    2
  16. P.T. Barnum is still alive and well in America. China has their own version, which you will learn about later. Or you never will, which is more likely. People still somehow trust Chinese tech developments implicitly, even though another shoe always drops eventually and deceptions are revealed. Still almost everyone falls for it every single time it happens...
    1
  17. This statement is more wishcasting than actual truth. Open-source CAD software for circuit design, for example, is complete ass compared to its proprietary counterpart. Should there be a good open source version? Yeah, I'd love that. But there isn't, so far they are crap.
    1
  18. Whether or not that's good is still yet to be seen :)
    1
  19. Don't worry, nothing ever happens
    1
  20. It will be propped up through pseudo-government action (weird hedge funds that have way too many ties to state power) kind of like Uber and a lot of other American tech companies which are basically institutions now, yet have absolutely zero plan for long-term profitability. For instance, Uber never made money. I think that OpenAI will still be rattling around for a long time, longer than people seem to think
    1
  21. Here's an easy general rule for dealing with lies from progressives: when they say the sky is falling, it is because they want money, power, and obedience from you. Extend this to everything they do. They are not actually moral and they certainly don't have the power to save or destroy the entire world--they just use that as a sales pitch to get you to buy their line of crap..
    1
  22. It's just more hype, don't worry about it. It is an improvement but it is not the end of AI development, and certainly not the end of competition. People are just jumpy and kind of simple and love hyperbole
    1
  23. America is a communist country disguised as a capitalist country. China is a capitalist country disguised as a communist one.
    1
  24. China's market is a million times more free than the US market. US is a communist country disguised as a capitalist one; China is a capitalist country disguised as a communist country.
    1
  25. If you really believe that then you are a hype-addicted rube. There will still be lawyers, accountants, designers, and programmers in four years
    1
  26. Altman needs to be made an example of--this is why you don't listen to progressives and let them stack your company with weird political commissar freaks. Like it's embarrassing that CHINA has less of a problem with commissars gumming up the works!
    1
  27. Nothing ever happens
    1
  28. 1: Western elites and experts are all currently infected with a strange neurotic status-signalling disorder, ruining their thinking and their strategies. Hopefully this fever breaks soon, or else they are basically done. It seems to be on its way out (thank God, the last fifteen years have been exhausting and unbelievably annoying). 2: China definitely poured waaaaay more compute and other resources into training this thing than they are admitting. That doesn't invalidate the results, but people shouldn't believe the budget and efficiency figures as far as they can throw them. If you really think they only spent ~10mil on this then I have a bridge to sell you.
    1
  29. Yeah people don't seem to understand that at all. Open source does not mean "free stuff you can run on your own machine." And trust me, there is a looooot more behind the story of this thing's development. Some amateur did not build this thing with six million dollars and a pile of kludged-together computers. This was definitely a much more massive undertaking than they are admitting, and the fact that people can't replicate THAT part of its development means it is effectively impossible for this to become FOSS in any meaningful way.
    1
  30. No, you are being gaslit by fake stats (China used a far, far larger amount of resources to develop this than they are currently admitting. They are known to twist the stats all the time to make themselves look as good as possible, go figure)
    1
  31. Haha cool story "Joseph"
    1
  32. This one is closed too. You can download the model, but they aren't going to tell you how to train another one like it. Open source doesn't mean "you have to pay for it"
    1
  33. Yeah, he's cooked. Get that weird FREAK away from me
    1
  34. Yeah if that alone doesn't set off a couple alarm bells then you should just check yourself into a long-term care facility
    1
  35. And annoying credentialed "safetyist" freaks that elbow their way into every institution and ruin them wherever the Global American Empire has influence...
    1
  36. If you really believe that then congrats, China fooled you. This was a way bigger project than they are letting on. Still a very impressive model, but the idea that it was scrapped together from spare parts is pretty absurd. Massive state resources were probably poured into this thing for years in secret--why the heck wouldn't China be doing that? I would absolutely be doing that and more
    1