What’s Wrong With Nvidia?
Again it just so happened that I got so angry about the AI fluff that I just couldn't shut up. This one's much shorter, I promise.
I just watched a little interview, where Jensen Huang, the CEO of NVIDIA, said something along the lines that America should be first in everything; that it's a fundamental logic error to think the task is equal to the job; and that you should basically buy NVIDIA chips for everything while keeping China begging for cheaper, worse ones. Sounds like he's partially correct, but not completely. And it's nothing to do with "evil China".
Look here: if you think the task is not the job, and you'll keep hiring just the same as before, then what's the benefit of AI, fool?
If AI does not allow you to get rid of that worker and therefore save money, then what's the point in buying NVIDIA chips (I mean no offense to workers)? Or, vice versa, if it does allow it, then the job might as well be reduced to the task, contrary to what Jensen says. That's the whole point.
So, regardless of which way it's going to go, Jensen's just not arriving at the correct conclusions. If you want to sell me something, then how about arguing for what I get in return for buying it? Can you even give me any guarantees? What do I get in return, besides better graphics for my games?
Stop being stupid. Either AI solutions are worth it or they're just another stupid investment that's not bringing in the savings to cover the cost.
Who's gonna pay more for computing when big companies cause shortages? Who's gonna pay more for electricity when the big companies build data centers? Who's gonna have a harder time competing with the big companies, when the only ones capable of producing the "baseline" are those same big companies?
When you have limitless money, then what you do is you cause a shortage, so that, regardless of whether you yourself actually produce anything of value, the cost of meaningful competition goes up, making it unviable.
The only solutions to that impasse are either breaking the monopoly-like entity or lowering the bar so much that the advantage no longer matters. But WOAH! The U.S. isn't doing that. Are you surprised?
What I am however surprised about is how the Finnish tech industry isn't doing anything about this either, no matter what.
And the only thing Europe's got in their pocket is "let's just use Linux." Well, better than nothing, I guess. So they're not completely leashed. But Europe is not going to have a software industry like that, is it?
Licky licky, sucky sucky.
I've noticed from the Jensen interview, as well as others, that the American big tech leadership is screwing with your psychology in the following way: they say AI running rampant is scary and insane, so therefore we need to have dialogue – who even knows what you're supposed to read between the lines there. But if you presume that simple dialogue is going to prevent China or the U.S. or anyone else from developing AI techniques in whatever way they see fit, then your eyes are not on the ball.
Turns out, all the states around the world are either developing whatever they can all the time or they're lacking resources. Isn't that the whole stupid narrative of why the U.S. went to Iran, as well? Because the Iranians were building nukes and their evil had to be stopped? I swear I heard someone say that. By that logic, France having nukes is a problem too, is it?
And what happened? A global energy crisis happened because of stupid U.S. policy.
Jensen also said 50% of Earth's AI developers are in China. That's an interesting claim, because then you immediately wonder why. You should have a coherent answer for what those people are actually doing, and why there's so many you're including in the "AI developer" category. That's actually really relevant.
You see, as a private person who's not a U.S. citizen, I don't really care about all that. Maybe if I wanted to be an AI developer myself, then I'd care just a little bit – about who my potential colleagues would be. But again, you should really sharpen your point regarding what AI techniques you mean exactly, and what those evil Chinese developers are actually working on. Are they such a threat to the American hegemony? And would a meaningful threat be found in numbers or rather in what was being developed precisely?
You see, if you believe Jensen, then you should be seriously interested in whether the bottleneck for developing AI truly is hardware after all; and if it's not, then how is it relevant what NVIDIA is allowed to sell and where?
Turns out his position is entirely political: he's positioned himself as the guardian of both U.S. security and tech leadership. Good for him then. Very convenient.
When you understand that AI itself is just a buzzword, then you start having all these awkward questions. But maybe if we'll lay this out, it will make sense.
Roughly speaking, most of them "dangerous" deep learning algorithms depend on extreme internet speeds, huge memory, huge disk space, and GPUs; Big Data, as they say, more or less. So you see, if not for the demand in deep learning, you wouldn't really need a very fast internet connection at all, thus eliminating a huge chunk of the need for technologies like 5G or 6G. If not for this demand, you wouldn't need much memory for your computers and servers either. And the only place for a GPU would be in gaming.
By the way, I just noticed yesterday my internet cost is going up. I guess you gotta cover those investments with some money after all, yes?
Self-driving cars of course best drive themselves when connected to the American Cloud; and your company's stupid recommender system only really works with tons and tons of data fed into it. Not that it's really doing anything except pattern matching over the items you have on offer, which might not be a lot and might not be what the customer wants next time, not to mention that computing text embeddings increases your energy expenditure. And of all that data that moves through the web, video is a huge chunk of it; all that video is what the smartphone is for.
That is, I've come to realize popular tech is all essentially like video games, with none of the charm. Popular AI is an entertainment product, leaking into all of adult working life and business. Because there is no critical thinking.
So all these guys are really just circle jerking each other, trying to get you to buy in on this new thing, because the need in one arena is going to fuel the need in another, despite no clear benefit to you. However, a couple American businesses dominating does not mean business is thriving. And to top it all off, it's not like you have any meaningful choice in the matter; the costs are going up regardless.
And it looks to me the only hope for poor, stupid economies like Finland is to just ride the wave with the big boys, because there is no more realism, and no more perspective.
That's because we don't really know what's coming in the future, because we're not architecting a goddamn thing. Otherwise we'd know. We're busy licking and sucking. One direction or the other.
You have no idea what's coming in 2027.
People obviously know GPUs can be used for nasty things, and I guess the Chinese know that too. I'm constantly surprised that people behave as though they didn't have an arsenal consisting of millions of drones, troll farms and humanoid robots somewhere deep underground. But that's why I'm actually more scared of the U.S. than any other country, because they're ahead, and because they're still willing to sacrifice humans for policy, despite maintaining the strongest and most well-funded military in the world. Let that sink in.
How can you say that Huawei is more of a state crony than NVIDIA is with respect to the U.S.? It seems states are all intertwined with the private sector now, and technology is mandatory in just barely keeping the economy standing on its spindly legs at all. But it's obvious that big people have already recognized just how precarious their position is, which is exactly why states all around the world are getting mixed up in all this in the first place.
Lobbying.
Put plainly: it's actually cheap to provide the same level of service the American tech giants did ~15 years ago. And so, you are made to believe that that's no longer sufficient, even though it might as well be, because there's no real consensus on value anymore.
That's why it's in the big boys' interests to continuously raise the bar to prevent further entry into the market. That's why junior coders have been having a hard time as well, and might still have. And the only barrier in front of them is the ego and self-righteousness of those already finding themselves in a good position.
You see, when you speak of infinite value in the future, then I dare to question whether you actually understand what value means. Without any reasonably democratic, majority agreement behind you, you can't really dictate that just by yourself.
Coming from me, all this criticism is actually really ironic, because I used to love NVIDIA, back when they were just an honest gaming hardware company. But I just can't respect a person whose whole agenda is to undermine every other country in the world, as though that were a good thing. Plus, it doesn't seem to bode well for anyone except the American rich anyway, together with their globalist plants and minions.
That's because if AI does succeed, then it's not going to help you pay your rent. And if AI doesn't succeed, then it's not going to help you pay your rent. Let that stew for a moment. Hell, make it boil.
Indeed, one thing I've wondered about for a while is "how long should you wait before any claims and promises are actually proven wrong?" How long should you keep trusting; why doesn't anyone ask that? Not in my country, not in the English-speaking web, nowhere. Nobody asks that, and yet you're just kinda supposed to use AI assistants now.
Because there is no deadline for anything, the projections do not matter. And of course, as anyone worth their salt knows, long-term predictions have a bigger error margin anyway.
Just stop being stupid and start asking some questions.
And stop making irrelevant decisions that'll only look nice on your report.
But, since it all ties back to politics and states anyway, let me tell you something I've learned.
Through the miracle of the internet I've come to know exactly what other countries think of us. When our dear president speaks, he delivers that same message too. First, Finland has a big military. Second, Finland has a big border with Russia. Now, what are you supposed to think about that, huh? What's one plus one?
I've also noticed that, almost immediately after joining NATO, internet propaganda about the amazing strength of the Finnish military has skyrocketed. I guess it's better than propaganda about the weakness of it, but only barely.
So, ergo, the English-speaking world mostly thinks this is the front line. Secretly it thinks that my country is expendable, but publicly it kisses ass and says we're strong. The only thing that's missing really, is a vague promise of a reward that's made before going to war, since there's a chance you won't actually have to fulfill on that anyway. – Wait a minute...wasn't there something about the AI or...something? Nah, it'll come back to me, I'm sure.
Funny, because I've known what the Swedes think for years and I never realized what the core Anglosphere thinks, with the possible exception of Canada; we'll see exactly what happens on that front.
The Swedes, you ask? The Swedes think us retarded, ugly drunkards who speak an ugly language. Sometimes they've a little bit of tact and they say it's an "interesting" language; sometimes they casually compliment our intelligence. Sometimes they make friends with the few Finns who move to Sweden. So it's actually the exact same thing as with the English-speaking world too: you say one thing while flirting with the opposite. And if you look at the world map, you'll see just how convenient that is.
It doesn't matter, as long as everything looks pretty.
I guess it's easier to justify sacrifice, when it's just some ugly, stupid thing. Thank you, internet, for letting me know.