The Rosenbergs exported the power to obliterate millions of innocent people in an instant. Huang exporting the power to break up America's attempted stranglehold on globally-relevant technology seems to me to be quite another matter. Especially since they're his own damn products.
Corvos
3hr ago
Imagine you never saw dice before, but God made some and told you to roll them. You roll them. Then they disappear, and nobody has any interest in making dice again. Does it make sense to talk about the probability distribution of the dice roll? I think it does.
I don't personally believe that thought experiment is applicable, or a wise way to think about one-off massively-multicausal events in the world as it actually is. You do, I get that, but I don't.
It could say he got lucky a few times.
I think this is hubristic. Strive to attribute to luck what was luck, sure, but assume merit before you assume luck. If you assume merit, the worst that can happen is you will overindex. If you assume luck, there is a very real danger you won't learn anything at all, to protect your ego.
Right, but you should counteract survivorship bias by looking at all the people who also did that and did not get super lucky.
What would I learn from that, beyond to temper my expectations and not to take super risky moves? Which I already know.
The fact of the matter is that some people succeed and some fail, and the ones who succeed are usually more interesting and more qualified along multiple axes than the ones who fail, and provide more to learn from.
Corvos
4hr ago
Thank you, I know what probability is. Talking about probability for one-off events where a true probability can never be established is intellectually and philosophically dubious, in my personal opinion.
Forecasting is one thing, you have no choice, but when we're talking about the past and other people's actions and the consequences thereof, I think it behooves one to show humility and study what successful people did rather than assigning their success too confidently to 'luck'. I think it produces a more accurate analysis.
Your analysis will be confounded if you can't separate the luck from the decision. A lot of people have this issue actually and it leads to cargo culting.
Certainly this can happen. To my mind, the point is to look at what they did in advance, and how much that put them in position to get whatever lucky break they ended up getting. I have no idea about Musk specifically but he's done well a few times now and that says something. Reading hagiographies uncritically can be fun too, of course, but that's more like a vicarious fantasy.
To quote Napoleon: "I like my generals lucky."
Corvos
5hr ago
I've no idea? I don't really believe that you can model the world like that - Rawlsian 'but what if you were somebody else' philosophy has never really held much interest for me.
Certainly every very rich man had a lucky break. More so now, when the world is more global and competition is fiercer. So hypothetically Musk / Huang / whoever would have some lower median net work spread over all the infinite worlds that don't exist. Unless they are actually genuinely special: I haven't met any of them. I've met someone comparable and he was definitely a cut above most people in risk-management/politics/drive/intelligence but how far is hard to tell.
In general though I'm more interested in analysing how people make good decisions to maximise their potential for lucky breaks than to what extent people do get lucky breaks.
Corvos
5hr ago
See my reply to @dailydogma. Other companies may have dropped the ball, but NVIDIA definitely picked it up early.
NVIDIA began working on high-performance parallel computation 2003-2007. They created and released CUDA in 2007, long before AlexNet, so that scientists could do high-performance calculations in a dedicated computing language instead of formatting their work as gimped shaders. In 2007 NVIDIA had already released a separate GPU for scientific research: the Tesla.
Between 2009 and 2011 NVIDIA's annual reports repeatedly made it clear that they considered scientific computing to be one of their three top priorities, on par with gaming. They saw this as investing in financial modelling, pharmaceutical modelling and others potentially large markets.
AlexNet didn't come out until 2012.
Yes, Huang was fortunate that AI ended up being based on linear algebra, but business development is about going hunting for Lady Luck and being ready to welcome Her when she arrives. NVIDIA definitely did that.
TLDR: The reason that NVIDIA is a giant and AMD is barely an afterthought (in GPUs) is because NVIDIA made a lot of expensive and tactically insightful investments in the 2000s, long before deep learning came along.
Corvos
5hr ago
Because a guy who is only half-American, who didn't grow up in America, had a Chinese girlfriend and may have told her his own company's private secrets although there is no evidence of it? The Manhattan Project was a government military project during wartime, the two are not remotely comparable.
Corvos
6hr ago
Its not like it happened by accident. Nvidia foresaw the use of their cards for scientific computing when none of their competitors did. Like all siccess stories, Huang was lucky that the future came in a particular shape, but he also made very savvy moves.
The thing is, you can't achieve anything by not talking to important people. And important people aren't necessarily safe. It's totally possible to get more value out of a relationship with someone who you know is passing on some of that information, vs. ceasing contact. Risk-reward. That's why the top levels aren't bound to the same restrictions the lower levels are.
Corvos
14hr ago
"Mademoiselle, we would like you to sleep with a foreign politician."
"Already on it! On y va!"
"And listen to him talk about his work."
"Je suis obligée?"
So take responsibility and clean it up. At least the midterms have sent the vague message that random wars are bad for Presidents, though I doubt the lesson will stick.
Corvos
1d ago
Countries are sending the message to America that they won't clean up the mess just because America shits the bed. Boomer American ego is a bigger problem right now than Iranian tolls, though that may change.
Corvos
1d ago
One of the things I was impressed by about the LW community was that they'd got pretty good at finding techniques to prevent the usual pitfalls of internet arguments. Taboo your terms, politics is the mindkiller, etc. etc. Sometimes it could get compulsive or silly, and it doesn't jive with everyone, but it did broadly help them have better discussions IMO.
Corvos
2d ago
I agree that the Motte has long since lost most of its high decouplers, and that this is a shame. That said, it is easier for everyone to decouple when the topic is historical or fantastical or obscure and thus has minimal emotional charge. If you want to discuss the legitimacy of states, somewhere like Cambodia would be a better bet.
Corvos
2d ago
Okay, but you can at least inhabit a perspective where one of them is more correct than the other, right? Yes, the interpretation of the winner tends to win out in practice because the winner can apply power to make it so, but that doesn't actually make him correct any more than the Chinese emperor can turn a deer into a horse.
I'll look into that, I wasn't aware. You may be right, but I do think that current licensing has remained extremely restrictive long after better arrangements could have been sorted out, primarily because publishers find the extra restrictions extremely beneficial in other ways. They have consistently resisted attempts to organise e.g. inheritance of digital property.
That can hardly be the only reason, because they impose way more restraints than that. And copyright law covers that case anyway: you own a physical book in all respects but you can't legally make 50,000 photocopies of it and sell them. There's no reason you can't give away, share, edit or inherit software and media in the same way that you can with the book, except that it's mildly more annoying to handle the DRM and the companies want to maximise copies sold.
In practice there's a huge amount of space devoted to control freakery: 1980 Nintendo's insistence on signing all games released for their console so as to protect their family-friendly brand, for example.
You seem to be talking past me. The point isn't what some looter thinks he is doing, the point is the actual case in question, which I have argued is different.
Do you agree that it is different in at least some respects, or do you believe that everything I'm writing is the delusion of a thief convincing himself theft is moral? In the former case we can discuss where the moral line actually sits, in the latter I don't think we have anything useful to say to each other.
if consumers are never willing to pursue alternatives.
I am not 'consumers'. I am a consumer. 'If everyone only...' never ever works.
I guess I could see the logic for jailbreaking a game you already own, or downloading a copy of it for more convenient use, or pirating a new server for an online game that's dead.
But the logic of "the big corporation are greedy and unfair and just want to take advantage of me, maaaaaaaaan, so I'm justified to take what I want without paying" is the ethics of a shoplifter.
Right, that's what I'm saying. There's a spectrum. The more boneheaded and aggressive the company's behaviour, the more justified I feel in circumventing it. For example, I own a copy of Bloodborne but no longer own a PS4, so that game is useless. I would therefore feel completely justified in downloading the ROM and playing it on an emulator. The same for something that isn't offered in my region. If you buy software at full price, but they deliberately content behind premium purchases or a system deliberately designed to feel like you can grind them out but you can't... I probably wouldn't pirate it but I would mind considerably less about people who did. Likewise lootboxes.
I feel like there is an awful lot of scope, in a world where games are not just a product but a sophisticated content delivery and monetisation system, and the walled-garden strategy is pursued ruthlessly, for business norms to produce practices that are genuinely inappropriate and don't earn my sympathy.
There is a big difference in my mind between say:
- Playing a multiplayer game on PC for free: dude, you couldn't be bothered to sign in to steam and pay $20 for it?
vs.
- Playing a multiplayer game on Switch 2 for free: dude, you couldn't be bothered to buy a $500 dollar games console to replace the $500 games console you bought 5 years ago that had basically the same functionality, and then spend $20 on the game, and then spend $10 a month on Nintendo's multiplayer service offering?
I personally believe that one should be honourable: there is scope for circumventing anti-consumer bullshit and the milking machine in different ways, but you should try to buy a copy somewhere and if you are playing everything for free then you are likely doing it wrong.
I don't believe that unilaterally forcing a healthier and more appropriate relationship between you, the supplier and the product is necessarily wrong, and I think flat out saying that, "there is no meaningful difference between downloading a cracked game that's not otherwise available on your device, and smashing the window of a store to take a TV" is overly reductive.
Not least because I would consider such a product/company so wildly unethical that I wouldn't want to support them, even spiritually by honoring their product.
Honestly, I don't think this works. There are too many whales in the world, the industry can stay unethical pretty much permanently as far as I can see. I'm not necessarily willing to accept a choice between jumping through every bullshit and expensive hoop placed before me, or just flat out boycotting massive chunks of the industry. I don't pirate, actually, but I feel increasingly less sympathetic to the corporations who complain about it. Steam proved decades ago that piracy is a service problem not a price problem, and if the industry refuses to learn that lesson because they make more money and have more control this way then I'm disinclined to oblige them.
Corvos
3d ago
Life involves thousands of decisions in which one can be pro or antisocial at little or no cost.
That cuts both ways, does it not? Are console exclusivity and restrictive EULAs pro- or antisocial?
There is literally no way that you can legally buy a game. You can only purchase a revocable, non-transferable, full-price license to play the game until the publisher changes their mind or the servers shut down. That license is gimped in a dozen different ways, not least of which is that they do not permit it to work on a general computing device but only on their special hand-marked and very expensive machine. It's as though you bought a Hermes bag but were only allowed to wear it if you had your entire outfit from the same store.
We are not talking about normal products and we are not talking about normal markets.
Okay, that's fair. They executed well, though, and my point is more that NVIDIA (and Huang) weren't just silly gamers who got the AI revolution magically dumped into their laps.
More options
Context Copy link