Great article, Charles. So strange that the world and investment community specifically took these clowns' word without any verification on their spend, the chips they used and if they were distilling other platforms (e.g., OpenAI - which now sounds like they were, big surprise).
On the other hand, I'd venture that the $600 billion wipeout of Nvidia in the space of a couple of hours demonstrates that the market is now actively looking for any black swan to pin a drop on - the Mag7 valuations are stretched by any fundamental valuation and a reasonable correction would be good for the markets
Shakespeare to the English, is like Sun Tzu for the Chinese. It sits in the popular consciousness.
So taking a pound of flesh (and not a drop more) from this concept I believe this popular phrase from the Art of War sums up the DeepSeek situation succinctly:
“All warfare is based on deception. Hence, when we are able to attack, we must seem unable; when using our forces, we must appear inactive; when we are near, we must make the enemy believe we are far away; when far away, we must make him believe we are near.”
DeepSeek R1 14B may have been trained on datasets containing conversations involving Claude. If these datasets include examples where Claude refers to itself, the model might mimic that behavior.
I think a simplification of the learning process (if true) does not make the market for chips smaller but much bigger. It would turn out any company can do what only it tech gigants or highly ai-specialized companies where supposed to do. Same as with classical hardware, when PC's went from 5k to 500$ and the software supply became diverse and cheap, the market exploded.
Excellent perspective Charles
Great article, Charles. So strange that the world and investment community specifically took these clowns' word without any verification on their spend, the chips they used and if they were distilling other platforms (e.g., OpenAI - which now sounds like they were, big surprise).
On the other hand, I'd venture that the $600 billion wipeout of Nvidia in the space of a couple of hours demonstrates that the market is now actively looking for any black swan to pin a drop on - the Mag7 valuations are stretched by any fundamental valuation and a reasonable correction would be good for the markets
100%
Shakespeare to the English, is like Sun Tzu for the Chinese. It sits in the popular consciousness.
So taking a pound of flesh (and not a drop more) from this concept I believe this popular phrase from the Art of War sums up the DeepSeek situation succinctly:
“All warfare is based on deception. Hence, when we are able to attack, we must seem unable; when using our forces, we must appear inactive; when we are near, we must make the enemy believe we are far away; when far away, we must make him believe we are near.”
Sun Tzu
Another more recent example were the "Ruperts" (or Oscars) that were so crucial to D-Day. Sadly you have to visit French museums to see one today.
https://en.wikipedia.org/wiki/Paradummy
With Alibaba releasing their own bot, this looks a lot like it's going to be a sustained attack on US market sentiment. Stay tuned!
I documented my findings here that it seems to suspiciously be based on Claude. https://www.reddit.com/r/ClaudeAI/comments/1ibxolt/comment/m9lzmps/?context=3
DeepSeek R1 14B may have been trained on datasets containing conversations involving Claude. If these datasets include examples where Claude refers to itself, the model might mimic that behavior.
I think a simplification of the learning process (if true) does not make the market for chips smaller but much bigger. It would turn out any company can do what only it tech gigants or highly ai-specialized companies where supposed to do. Same as with classical hardware, when PC's went from 5k to 500$ and the software supply became diverse and cheap, the market exploded.
Perhaps in the longer term - but in the short term this would cause a confidence crash. Place your bets?