Not quite, I believe this sell off was caused by DeepSeek showing with their new model that the hardware demands of AI are not necessarily as high as everyone has assumed (as required by competing models). I've tried their 7b model, running locally on a 6gb laptop GPU. Its not fast, but the results I've had have rivaled GPT4. Its impressive.
None of the models other than the 600b one are R1. They’re just prev gen models like llama or qwen trained on r1 output making them slightly better
"Slightly" is an understatement, though. Distillations of R1 are significantly better than the underlying models.
Curious thought: could those large price movements have something to do with the fact that DeepSeek is financed by a hedge fund (rather than the more typical VC)? It is unclear how DS will make money from its current strategy of sharing much of the secret sauce that went into training as well as releasing the results under permissive licenses. But if the play was "short major tech stocks and then release surprising results in a way that maximally undermines their current growth story", then it could make a lot more sense.
IMO this is less about DeepSeek and more that Nvidia is essentially a bubble/meme stock that is divorced from the reality of finance and business. People/institutions who bought on nothing but hype are now panic selling. DeepSeek provided the spark, but that's all that was needed, just like how a vague rumor is enough to cause bank runs.
Not quite, I believe this sell off was caused by DeepSeek showing with their new model that the hardware demands of AI are not necessarily as high as everyone has assumed (as required by competing models). I've tried their 7b model, running locally on a 6gb laptop GPU. Its not fast, but the results I've had have rivaled GPT4. Its impressive.
To me this rings a lot like “640KB ought to be enough for anybody”
Similarly, as fast as processors have gotten, people still complain their applications are slow. Because they do so much more.
Generally applicable ML is still in its infancy, and usage is exploding. All those newfound spare cycles will get soaked up fairly quickly.
Nvidia valuation = $3000B Nvidia profits during AI madness last year = $65B (last 4 quarters) Nvidia profits during normal year = $5B This stock could drop 90% from here, and still be expensive. The numbers are absolutely crazy and make no sense at all. Stargate project is aiming to invest $500B over 4 years. Those $500B are a pipe dream, but let's suppose for a second that all of that $500B will be Nvidia profits an…
Are you suggesting "AI madness" is a one time event that will not continue? AI is dramatically improving productivity and not just for developers. The "madness" will continue until every job is fully automated. That requires a lot of chips.
"AI", once we get there, is indeed the future and will continue to play an increasing role everywhere.
But what happened to Nvidia profits last year is a one time event and will get back to normal sooner or later.
Not quite, I believe this sell off was caused by DeepSeek showing with their new model that the hardware demands of AI are not necessarily as high as everyone has assumed (as required by competing models). I've tried their 7b model, running locally on a 6gb laptop GPU. Its not fast, but the results I've had have rivaled GPT4. Its impressive.
I believe you that it had to do with the selloff, but I believe that efficiency improvements are good news for NVIDIA: each card just got 20x more useful
They only got more useful if the AI goldrush participants actually strike, well, gold. Otherwise it's not useful at all. Afaict it remains to be seen whether any of this AI stuff has actual commercial value. It's all just speculation predicated on thoughts and prayers.
Google and Microsoft both bundled their AI offering into their office suites because they weren't getting traction as an add-on. Startups don't have that option.
They did that because 1) Their initial AI offerings weren't real products customers would use or pay for 2) They weren't seeing sufficient adoption to justify the expense 3) They have insane levels of distribution in their existing product lines and can incrementally add AI features This is entirely orthogonal to whether or not other startups can build AI-first products or whether they can position themselves to comp…
Yes as you point out in #1, they failed to develop an AI product and #2 they failed to get traction in the market.
What moat does an AI startup have that will prevent them from being crushed by big tech?
Not quite, I believe this sell off was caused by DeepSeek showing with their new model that the hardware demands of AI are not necessarily as high as everyone has assumed (as required by competing models). I've tried their 7b model, running locally on a 6gb laptop GPU. Its not fast, but the results I've had have rivaled GPT4. Its impressive.
I believe you that it had to do with the selloff, but I believe that efficiency improvements are good news for NVIDIA: each card just got 20x more useful
When your business is selling a large number of cards to giant companies you don't want them to be 20x more useful because then people will buy fewer of them to do the same amount of work
NVIDIA sells shovels to the gold rush. One miner (Liang Wenfeng), who has previously purchased at least 10,000 A100 shovels... has a "side project" where they figured out how to dig really well with a shovel and shared their secrets. The gold rush, wether real or a bubble is still there! NVIDA will still sell every shovel they can manufacture, as soon as it is available in inventory. Fortune 100 companies will still…
Jevon's paradox would imply that there's good reason to think that demand for shovels will increase. AI doesn't seem to be one of those things where society as a whole will say, "we have enough of that; we don't need any more". (Many individual people are already saying that, but they aren't the people buying the GPUs for this in the first place. Steam engines weren't universally popular either when they were introdu…
The other thing is that if this pushes the envelope further on what AI models can do given a certain hardware budget, this might actually change minds. The pushback against generative AI today is that much of it is deployed in ways that are ultimately useless and annoying at best, and that in turn is because the capabilities of those models are vastly oversold (including internally in companies that ship products with them). But if a smarter model can actually e.g. reliably organize my email, that's a very different story.