Source report:
“DeepSeek Debates: Chinese Leadership On Cost, True Training Cost, Closed Model Margin Impacts” by SemiAnalysis
DeepSeek took the world by storm. For the last week, DeepSeek has been the only topic that anyone in the world wants to talk about. As it currently stands, DeepSeek daily traffic is now much higher than Claude, Perplexity, and even Gemini.
But to close watchers of the space, this is not exactly “new” news. We have been talking about DeepSeek for months (each link is an example). The company is not new, but the obsessive hype is. SemiAnalysis has long maintained that DeepSeek is extremely talented and the broader public in the United States has not cared. When the world finally paid attention, it did so in an obsessive hype that doesn’t reflect reality.
We want to highlight that the narrative has flipped from last month, when scaling laws were broken, we dispelled this myth, now algorithmic improvement is too fast and this too is somehow bad for Nvidia and GPUs.
While it does seem like DeepSeek has made some important achievements in terms of compute efficiency (although I still remain sceptical), it’s funny how so many people (including some in the media) uncritically believed their initial claims.
A similar situation to all the astro surfing around users switching to [competing Chinese social network] with respect to the possible US TikTok ban. I don’t deny that this may have happened on some level, but it’s clear that the internet is an easy avenue for botting and creating fake sentiment.
You’re using an article with 0 sources from people who are long in NVDA and have a vested interest in deflating the excitement for deepseek. You just happen to have a dislike for deepseek and chose to buy what the idiots at semianalysis are selling because they use some technical jargon and write with the confidence and certainty that only a fool can muster.
Is there any public source regarding SemiAnalysis being long on Nvidia?
There is no need to get personal. I haven’t tested DeepSeek yet and I am happy to be proven wrong or be corrected. What I do have is experience with propaganda in authoritarian revanchist states (I have lived in russia for a decade) and this is not something I will ignore.
Why not just share your opinion and critiques instead of making assumptions about me? It was genuinely not my intent to make a strong statement regarding DeepSeek; it was more of an observation on the nature of social media sentiment.
Dylan had his portfolio (NVDA heavy) on commonstock early in 2022.
And here it is, you wanted to believe “deepseek bad” so you latched on to the idiots at semianalysis. This is called post hoc rationalisation and is prevalent in online discourse. It’s also why we’re in a post truth society, people don’t look for info, they look for confirmation.
Deepseek shared an invaluable tool with the community as did mistral and meta before them. This is because without wide adoption there’s no monetisation. There’s no big conspiracy to be had.
I did not know one of the authors was long on Nvidia. This is an important point.
I do think you are misinterpreting my position. This was not a post hoc rationalisation and I explained why.
Here’s another clue they’re talking out of their arse. They throw FUD on the Huawei partnership with DeepSeek. Turns out, they were full of shit. A real analytics company, not a grifter who professionalised bribing the supply chain for tidbits, would have seen this coming months in advance due to the foundry traces all the way down the supply chain. They’re just aping Demerjian without any of the contacts or industry knowledge but enough VC money to buy puff pieces on them in business insider.