Recently, I posted a reply to @ZviMowshowitz stating that it is likely that, for a number of reasons, the "DeepSeek R1 crash" in the stock market is in some way related to a scam. Rather than summarize, I'll just re-post the link: https://x.com/SteveSokolowsk2/status/1883918506111963572 .
This market will resolve to YES if, before the end of April, the spirit of the post is correct. For example, one of the following would be true, but the list doesn't end here:
A group of people intentionally pushed a narrative over social media during the weekend that DeepSeek R1 can be run on phones to profit from the crash on this old news
The Chinese government or a Chinese company has actually been smuggling banned chips into China, and the $5 million claimed for training is false
A group of researchers did not actually train the model in their spare time, or the Chinese government was involved in the training
The Chinese government had a hand in releasing the model on Trump's inauguration day for political reasons
If time runs out on the market and we never learn why the crash occurred, then it will be presumed that the commonly believed reasons were correct, there was no scam, and the market will resolve to NO.
Update 2025-01-30 (PST) (AI summary of creator comment): Clarification on the involvement of the Chinese government:
There needs to be a strong link, such as individuals being paid by the Party to perform this specific work.
Update 2025-01-30 (PST) (AI summary of creator comment): Model Usage by the Party:
The model was used by the Party to create a market panic, indicated by the unusual delay between its release and the market's reaction.
1,000
3.00A group of researchers did not actually train the model in their spare time,
This would be an extremely strange case for a YES resolution—the mainstream consensus belief is that top tier AI talent (fully focused on this work) delivered this result. Confirmation of this would not confirm a “scam angle”.
@Ziddletwix agree. They never claimed it themselves, this market is the first place I'm hearing this idea from. They released the model as a product made by a company that presumably employs professional researches for money, since it's a Chinese company, likely at least partially with government funding. The official paper on r1 cites the company as author/owner of the project, this wouldn't be the case if it was someone's "free time" project.
https://github.com/deepseek-ai/DeepSeek-R1/blob/main/DeepSeek_R1.pdf
@mathvc Your statement to me is equivalent to saying that half of the United States is Republican.
That doesn't mean that all of the Republicans have influence on Trump, though. There needs to be a strong link - i.e. the people were actually being paid by the Party to perform this specific work.
@mathvc I'm not sure it matters what specialty the people who worked on the model are. What matters is whether the model was used by the Party to create a market panic, because of the odd delay between its release and the market's reaction.
@SteveSokolowski i am talking about the point that @Ziddletwix and me have raised in this comment section
@jim If that is actually the case, then the resolution is NO.
However, I also think this might be false, as it would also cost less than the profit to be made from puts to buy old phones that were destined for recycling to run up these app store download figures.
Plus, after all this success, now they suddenly restricted user signups, claimed a DDoS attack without posting evidence, and killed their momentum the moment the stock market bottomed? It's all right in line with some sort of scam.