John Ruich

Well, one interesting thing we'll be looking out for is that the NPC may pass some legislation that would support and protect private businesses.

And that's pretty interesting because the leadership here seems to have renewed its interest in these companies as drivers of the economy and innovation, especially after the Chinese company DeepSeek released an advanced AI model in January that shocked the world. NPR's John Ruich with some insights from Beijing.

And that's pretty interesting because the leadership here seems to have renewed its interest in these companies as drivers of the economy and innovation, especially after the Chinese company DeepSeek released an advanced AI model in January that shocked the world. NPR's John Ruich with some insights from Beijing.

And that's pretty interesting because the leadership here seems to have renewed its interest in these companies as drivers of the economy and innovation, especially after the Chinese company DeepSeek released an advanced AI model in January that shocked the world. NPR's John Ruich with some insights from Beijing.

Yeah, DeepSeek is a spinoff from a Chinese hedge fund. It was established just two years ago in 2023. And it's based in the eastern city of Hangzhou, which is sort of a tech hub here in China.

Yeah, DeepSeek is a spinoff from a Chinese hedge fund. It was established just two years ago in 2023. And it's based in the eastern city of Hangzhou, which is sort of a tech hub here in China.

Yeah, DeepSeek is a spinoff from a Chinese hedge fund. It was established just two years ago in 2023. And it's based in the eastern city of Hangzhou, which is sort of a tech hub here in China.

And in a nutshell, what they did was hire a bunch of top-notch engineers and develop new algorithms, basically more efficient ways of training and running artificial intelligence with less computational power.

And in a nutshell, what they did was hire a bunch of top-notch engineers and develop new algorithms, basically more efficient ways of training and running artificial intelligence with less computational power.

And in a nutshell, what they did was hire a bunch of top-notch engineers and develop new algorithms, basically more efficient ways of training and running artificial intelligence with less computational power.

Well, the product is said to rival tools from competitors like OpenAI and Google in terms of what it can do, things like analyzing data and solving complex problems. It's impressed a lot of people. It rattled markets.

Well, the product is said to rival tools from competitors like OpenAI and Google in terms of what it can do, things like analyzing data and solving complex problems. It's impressed a lot of people. It rattled markets.

Well, the product is said to rival tools from competitors like OpenAI and Google in terms of what it can do, things like analyzing data and solving complex problems. It's impressed a lot of people. It rattled markets.

And what rattled the markets is the narrative, which comes with some caveats, that DeepSeek basically did it all cheaper, quicker, and with less powerful microprocessors than its big competitors.

And what rattled the markets is the narrative, which comes with some caveats, that DeepSeek basically did it all cheaper, quicker, and with less powerful microprocessors than its big competitors.

And what rattled the markets is the narrative, which comes with some caveats, that DeepSeek basically did it all cheaper, quicker, and with less powerful microprocessors than its big competitors.

Sure, the first one is around cost. DeepSeek says that it spent under $6 million to make this thing. That's tiny relative to the hundreds of millions of dollars that others are investing, even billions. But analysts say that that low figure is easy to misinterpret because it doesn't include, for instance, the cost of developing various versions from which this latest version was distilled.

Sure, the first one is around cost. DeepSeek says that it spent under $6 million to make this thing. That's tiny relative to the hundreds of millions of dollars that others are investing, even billions. But analysts say that that low figure is easy to misinterpret because it doesn't include, for instance, the cost of developing various versions from which this latest version was distilled.

Sure, the first one is around cost. DeepSeek says that it spent under $6 million to make this thing. That's tiny relative to the hundreds of millions of dollars that others are investing, even billions. But analysts say that that low figure is easy to misinterpret because it doesn't include, for instance, the cost of developing various versions from which this latest version was distilled.

So we really don't know what the total development cost was, how inexpensive it was. The second caveat is It has to do with the hardware, has to do with the chips that are critical to developing AI.