DeepSeek

“They made so much money on their Nvidia trade. It doesn't, it's irrelevant.”

“And the American markets To, I think, at least some of our surprise, take these claims at face value, and summarily lose one trillion dollars.”

“So, like, they are losing a ton of money, but they're just losing a little bit less money than everyone else is losing.”

“This DeepSeek thing is exactly what we were talking about, that by releasing Llama, you know, DeepSeek surely used it, and many others, sounds like, and the fruits of that win come back to Meta, like, overnight, right? And they can plug it back in and make mor…”

“The problem with DeepSeek is it indicates a direction towards total commoditization, which means that your flaming money defense does not work, right? And it will just be everywhere.”

“They do have this one part of one of the papers that talks about a unified paradigm around reinforcement learning, which is pretty cool. They've got a Very beautiful piece of math that kind of brings together all these different fine tuning and optimization te…”

“It's entertaining because the people who made it run an algorithmic hedge fund.”

“The value of AI is AGI meme-y. Like, it's not, like, it loses a show of money, but it's, like, it's okay, because, like, AGI, right? It's very memetically driven. But the problem is that, like, it's a very binary zero or infinity value, and so it's extremely s…”

“So China's been proven to be able to release the exact right press release that can, you know, damage the world, damage the American economy.”

“Our guess was 10%, and we guessed three hundred million. It was actually closer to six hundred million.”

“We've done a lot of reporting at the information about businesses that sort of say, you know, we're not increasing our spend right now because we're getting efficiency gains from deep seek and all these other models.”