@CodeWithAmann it shouldn’t work this way. no one wins here. neither we win both or just we all gonna lose. also most of the times gui borns from cli. cli winning = gui winning.
@danielrmay if someone’s laughing at you/trying to make fun of you but can’t and don’t want to explain why - it’s a pure signal of dumbness.
смех без причины - признак дурачины.
@Daavya_vaishnav@signalapp signal is good but sadly no privacy it’s all bs(if some app requires phone number for registration they can’t talk about privacy)
@KRinweb3@Daavya_vaishnav@signalapp loool u really think it wins on privacy? bro its literally REQUIRES phone number. no privacy. research on truly private solutions (matrix-based best choice imo)
>racism
>passport
??? what passport has to do with race hahahaha??? yep maybe you are racist but if you want to hate someone based on PASSPORT you are also nationalist. and yep in the first place you are just dumbass(haha its 2026 bro really doesn’t know nationallism/racism difference) and racism/nationalism are just direct consequences.
so ? its not about how their country looks. in the end its just about money - and any potential threats to us security have that money. tell me exactly - how are you gonna detect complicated udp proxy setup with fully fingerprint spoofing. also there is remote desktop solutions. what are you gonna do about remote desktop. okay LETS SAY YOU CAN detect any vpn(which is not true and never going to be that). what are you gonna do about remote desktop. someone from us just renting their pc to north korean hackers while ordinary jack from uk cant have fun with new model. your solution?
@Lordsaucyy@emanueledpt and what's the point of it ??? us citizens will just sell access to it. (even if it somehow becoming crime, it will just increase price).
and i don’t really know how to resist such things without releasing weights. want to know your opinion on that, maybe you can change my mind - but fable i think fable situation just proves my point. and we still early on ai and have time to build around openweight models. rly want to hear your opinion on it.
in my vision: if we really believe that ai is next big thing that will really change the world(i’m not talking about agentic “golden rush”, i mean really something that will empower humanity in general, currently imo it’s still a little niche which is sustainable for tech people and giving some impact(which is overhyped currently) for non tech people) it will be really dangerous to give any corp/govt some form of a “blackbox” control over it(i mean - any regulations, restrictions by any entities is completely okay as long as it has open policy on it, which shouldn’t be just laws or some ToS, it should be provable by code). just imagine if it is going to change peoples life as smartphones/pcs and you will end up to be forced to do something you don’t really wanna do(for example: current situation in some US areas where you have to provide your id for using a pc, what a fucking joke.)
i don’t want to gave up another sphere in my life to be controlled by some freaks whose goal is just to control your money, your opinion, your way of living etc.
While we are doing lobbying for AI laws, I propose this. All providers of closed source models are 100% liable for their outputs. Copyright, libel, conspiracy, fraud, etc... If you want safe harbor from this liability, release the weights. This is real AI safety legislation.
@natolambert Reminds me of when downloading mp3 files was banned. Some grandma somewhere is gonna download a Qwen and catch federal charges, but this won't change most people's behavior.
white pill for my nerds:
60fps e-ink display
a random guy outperformed entire eng teams by developing a pixel by pixel driver for e-ink displays that makes it 60fps.
he did that after work for months, launched it yesterday.
the future is bright
if we're gonna just sit and watch how it goes on - you are right, but if you are really worried about it, and have any glimpse of hope - you should do something now. it's time to take some action and im not losing any hope on fairly distrubition just because of i see progress of OSS models, so it's already really good. if i cant build models itself, i can build proper infrastructure around them and really save some hope that in 5-10 years we can see it disturbed not only by anthropic and us govt orders. maybe it sounds utopian, but for me - worst thing i can do now is ignore and not build.
@jtregunna@catalinmpit also i think “infra” prt is related to this article, consider reading it, maybe you can help with your vision of problem https://t.co/WDtNlH0Kbn