I wish I would not already be so leveraged.
Nvidia’s $589B DeepSeek rout
271–280 of 1001 posts
Re: Nvidia’s $589B DeepSeek rout
#272Wow, literally bought more Nvidia shares last week. Just goes to show the stock market is 80% gambling. Me believing in the value of a company is overshadowed by the hype of growth and "future valuation".
If this is gambling, what isn’t gambling? Sometimes things just don’t go the way you want. A Chinese company coming up with a cheaper alternative to a cutting-edge technology out of nowhere, is an outcome that is hard to predict. In hindsight, betting on Nvidia maintaining its monopoly on a resource crucial for such an important technology as AI, might not be the best of ideas, but then again, who knows.
buying an individual stock is gambling. buying an broadly-diversified index is not.
Re: Nvidia’s $589B DeepSeek rout
#273Earlier quoted context omitted.
> Where has Trump's agenda collapsed? I might have missed the press release. Well, it's too early to say if it has happened in this case. But we've seen it happen again and again, so it will not be a surprise if it happens. > And why would a curveball on AI throw off an agenda around trade, immigration and military engagement? I don't follow. Well, trade restrictions against China have just backfired spectacularly. S…
> Well, it's too early to say if it has happened in this case Ah ok, so you’re just guessing. You wrote it as if it had already happened. And I think you’re confusing the stock market with the broader economy. The economy as a whole is unaffected by AI at this point. And the trade war with China involved hundreds of industries beyond AI. Most of it is manufacturing. I’m not sure how failure of AI sanctions (questiona…
> Ah ok, so you’re just guessing. You wrote it as if it had already happened.
That's not a generous reading. I'm saying this is what has happened historically.
Edit: the use of the words "might" and "could" made this pretty clear.
> The economy as a whole is unaffected by AI at this point.
True in terms of no marked effect on GDP. But the stock market and the broader economy are very much linked. See 2008. And the stock market is high on AI.
> I’ve heard so many claims that the Chinese government doesn’t control research or businesses like TikTok.
Even western countries tell tech companies what to do. Do you think an authoritarian government is going to do that more or less? There's a reason they didn't want to sell TikTok.
Re: Nvidia’s $589B DeepSeek rout
#274ASML plunge indicates a hysterical/irrational component to the response, right? They aren’t going anywhere. If it turns out training is easier than expected, they make the devices that make the devices that do inference too… If the field is going to produce anything useful, cheap training gets us there faster.
Not making the news in western media yet is PRC claims to have started mass producing their indigenous 28nm litho (70% of global wafer use) this month... the estimated cost is 1/30th of ASML machines. Extrapolate and they're on trend of produce 14nm machines at comparable fraction cost in next few years.
Re: Nvidia’s $589B DeepSeek rout
#275Earlier quoted context omitted.
Right and LLMs will not be able to generate their own high quality training data. There are no perpetual motion machines.
> LLMs will not be able to generate their own high quality training data. Humans certainly did. We did not inherit our physics and poetry books from some aliens.
Humans and LLMs are different things.
LLMs can not reason - many people seen to believe that they can.
Re: Nvidia’s $589B DeepSeek rout
#276ASML plunge indicates a hysterical/irrational component to the response, right? They aren’t going anywhere. If it turns out training is easier than expected, they make the devices that make the devices that do inference too… If the field is going to produce anything useful, cheap training gets us there faster.
I'm guessing the sentiment is not about training, but about the R&D capability of China. If Chinese can figure out how to build a good enough model faster and cheaper, they may be able to come up with an ASML competitor as well.
Re: Nvidia’s $589B DeepSeek rout
#277I don't quite understand the sentiment. Lowered cost of training and inference means more companies can join the game, and therefore more demand for the hardware. Basically Jevon's paradox in play. What will more likely fall instead in the long-term should be OAI's valuation, if they don't have other killer products.
IMO reaction to short term semi cycle - a lot of reshored/redundant fab expansions last few years outside of TW. If compute costs goes down 10-50x there may not be enough use case to fill Jevon in next infra hardware cycle of ~5 years. People are doing more with same, but maybe use case not sufficiently more to go fill/justify acquiring more hardware in near term time frame.
Re: Nvidia’s $589B DeepSeek rout
#278ASML plunge indicates a hysterical/irrational component to the response, right? They aren’t going anywhere. If it turns out training is easier than expected, they make the devices that make the devices that do inference too… If the field is going to produce anything useful, cheap training gets us there faster.
Not really, if demand for high end high margin chips goes down considerably than margins and demand will go down across the board.
Re: Nvidia’s $589B DeepSeek rout
#279ASML plunge indicates a hysterical/irrational component to the response, right? They aren’t going anywhere. If it turns out training is easier than expected, they make the devices that make the devices that do inference too… If the field is going to produce anything useful, cheap training gets us there faster.
I think everyone was just a bit overpriced because of the AI hype and it's just letting out some air.
Re: Nvidia’s $589B DeepSeek rout
#280Earlier quoted context omitted.
$50 billion can be afforded by WAY WAY more sites and companies so Nvidia will simply delivered to >10x more data centers. Or instead of shipping 100k GPUs to Meta, they will ship 100k GPUs to 10 different customers. For Nvidia, it is great news because now finally, the concentration of GPUs at Hyperscalers will end and every Fortune company can finally get their local data center to train their AI Models. Because if…
> $50 billion can be afforded by WAY WAY more sites and companies Spending $50 billion to do $6 million worth of Ai training seems like a good way to trigger a golden parachute and "spend more time with your family" as a CEO.
How about spending $50 billion in capex to do $100 billion worth of inference?
Everyone is hung up on the cost of training R1 but it’s 685 billion parameters. We still need all of those GPUs to actually use the model.