Chinese DeepSeek Shakes Silicon Valley
13:14 Share

If you need three new reasons to love Jack wraps at Jack in the Box even more, here they are. Chicken fajita, chicken Caesar, and delicious, starting at $3. Coincidentally, those are the same three reasons you should come to Jack in the Box right now. At Jack, every bite's a big deal.

This episode is brought to you by Indeed. When your computer breaks, you don't wait for it to magically start working again. You fix the problem. So why wait to hire the people your company desperately needs? Use Indeed's sponsored jobs to hire top talent fast. And even better, you only pay for results. There's no need to wait. Speed up your hiring with a $75 sponsored job credit at indeed.com slash podcast. Terms and conditions apply.

Hey everybody, welcome back to the Elon Musk Podcast. This is a show where we discuss the critical crossroads that shape SpaceX, Tesla, X, The Boring Company, and Neuralink. I'm your host, Will Walden. So how did a Chinese AI company suddenly cause so many ripples in the tech world? And why didn't NVIDIA, the leading AI chipmaker, lose $600 billion in market value in one day?

And is DeepSeek's low-cost, high-performing AI model a sign of an emerging global AI race or something a little less threatening? So on Monday, the Nasdaq fell 3.1%, sending shockwaves across the stock market. And Nvidia bore the brunt of the losses with its stock price plummeting 17% and wiping out $600 billion in its market value.

The cause of this dramatic drop? Something called DeepSeek, a relatively unknown Chinese AI firm, unveiled a cutting-edge AI model that deliver comparable performance to industry leaders yet require significantly less compute power and cost far less to train.

Now, the implications of this announcement spread very quickly. For some, DeepSeek's achievement reaffirmed concerns about Silicon Valley's costly AI development practices. Others, particularly national security advocates, saw it as a wake-up call in the race for artificial intelligence dominance, particularly between the United States and China.

And for those worried about AI safety, DeepSeq's open access model represents a step toward uncontrolled proliferation of advanced AI systems. Yet, as transformative as DeepSeq's rise might appear, its actual significance remains a topic of debate. While some see it as a moment of reckoning for Western AI labs, others question whether it truly changes the competitive dynamics of this industry.

Now, one of the biggest concerns about DeepSeek is the uncertainty surrounding its long-term goals. The company is a spinoff of High Flyer, a decade-old Chinese quantitative hedge fund co-founded by DeepSeek's CEO,

Originally developed for trading algorithms, DeepSeq eventually shifted to building large-language models and committed to open sourcing its tech. Its stated mission is ambitious yet pretty vague: to unravel the mystery of AGI with curiosity.

And for now, DeepSeek's business model seems almost altruistic. It offers its AI tools to developers at minimal cost without clear profit-driven incentives, but that could change at any moment. The company could pivot to monetization by introducing subscription fees or restricting its API access. Some theorize that DeepSeek could also funnel user data into High Flyer's hedge fund operations to extract financial insights.

Now, another layer of uncertainty comes from the Chinese government, which might step in to regulate how DeepSeek distributes its AI. This prospect of a Chinese company providing advanced AI tools to developers worldwide, including in nations considered adversarial to China, is unlikely to go unnoticed by Beijing.

Adding to the concern is how DeepSeq managed to build its breakthrough models. It didn't invent a fundamentally new AI architecture, though. Instead, it optimized existing techniques pioneered by American firms like OpenAI and Google. Using older hardware and streamlined processes, DeepSeq achieved performance levels comparable to some of the best AI systems available today at a much lower cost.

Now, this approach shows the vulnerability for US AI labs. The technological leads are becoming increasingly short-lived. In early years, it might have taken months or years for competitors to replicate innovations like OpenAI's GPT models, and DeepSeek's rapid optimization shows that the time lag between cutting-edge development and global adaptation is shrinking.

As Anthropic co-founder Jack Clark succinctly put it, DeepSeek means AI proliferation is guaranteed. Now, DeepSeek's rise also exacerbates the perception of AI development as a geopolitical race. Some venture capitalists and tech advocates have long framed AI progress as a competitive motivation

moment between the United States and China. Their argument rests on fears that an authoritarian regime armed with advanced AI could outpace democratic nations and use AI technologies for strategic or military dominance. And critics warn that this narrative serves self-interest goals. By casting AI as an arms race, proponents can argue against regulation that might slow innovation or reduce returns on investments in AI startups.

Additionally, it channels more funding into military technology products, which often benefit the same venture capitalists advocating for this framing.

Yet, there's a genuine danger in accelerating AI development under the banner of competition. When safety concerns are sidelined, the risk of releasing powerful, unaligned AI systems grows. While U.S. labs like OpenAI and Anthropic have attempted to build safety measures into their AI models, these efforts remain incomplete.

Deep Seek, on the other hand, appears to have no public-facing strategy for ensuring its AI systems are safe or aligned with human values. And if you have any human values of your own, you take a look at the subscribe button or the follow button on whatever podcast platform you're listening on right now, because I've noticed that only 10% of you, around 10% of you, are actually subscribed to the show. So if you could do me a favor...

and hit the subscribe button, I would love that. Now, the people really think the best thing that humanity can do is create things smarter than ourselves though. And then we'll see what happens after that. The lack of a clear plan for managing superhuman AI intelligence only heightens concerns that framing AI as a race will lead to reckless outcomes. But despite these concerns, there are reasons to believe that DeepSeq's emergence doesn't fundamentally alter the AI industry's trajectory.

At least not yet. For one, much of what DeepSeq has achieved was already anticipated. Open source models have been reverse engineering breakthroughs from major labs for years, and the cost of training AI models has been declining steadily. The cost of producing GPT-4 level intelligence has dropped by a factor of a thousand in just 18 months.

The further price drops, while disruptive, were expected and won't necessarily destabilize major players right now. Additionally, U.S. companies continue to invest heavily in building infrastructure for more powerful AI systems. And while DeepSeek's cost-effective approach may be ideal for current AI tech, major firms like OpenAI and Google remain focused on training increasingly advanced models that require state-of-the-art hardware.

These investments in AI infrastructure are aimed at pushing the boundaries of what AI can do, not just optimizing existing technologies.

Another factor working in favor of U.S. labs is the Biden administration's export controls and advanced chips to China. And although DeepSeek has demonstrated its ability to innovate with less powerful hardware, the restrictions are still relatively new and may limit China's ability to scale its AI efforts over the long term. The lack of access to cutting-edge chips could make it

When you need mealtime inspiration, it's worth shopping Fries for thousands of appetizing ingredients that inspire countless mouth-watering meals. And no matter what tasty choice you make, you'll enjoy our everyday low prices. Plus, extra ways to save, like digital coupons worth over $600 each week and up to $1 off per gallon at the pump with points. So you can get big flavors and big savings. Fries, fresh for everyone. Fuel restrictions apply.

harder for Chinese companies to deploy their models as widely as US firms can. And while DeepSeek has made headlines for its optimization

American companies still lead in conceptual breakthroughs. Much of the industry's current focus has shifted toward developing AI agents capable of more autonomous and complex human-like tasks. And OpenAI's new offerings, such as tools that can operate independently to perform tasks, hint at where the next wave of competition will unfold. DeepSeek has yet to demonstrate similar ambitions, and there's little evidence that it can outpace U.S. firms in this area.

But the sudden rise of DeepSeek is kind of concerning for people. Its success has revealed vulnerabilities in the business models of major AI labs in the United States. It's raised questions about how AI technologies should be shared or regulated internationally. However, it's still challenging for all parties involved. The AI industry is shaping...

the way that we do business and the way that we live and the way that we communicate. Now there's declining costs of open source technologies. I'm a huge proponent of open source. I built my whole web development career on open source. I built

one of the most popular frameworks for a software to launch me into my career on open source technology, on bootstrap. So I understand that open source is a huge, huge win for AI. Do you want the huge tech companies building everything and controlling everything? Or do you think this is free speech?

I want to know in the comments, do you think code is speech? Because if I build something and I keep it to myself and then I sell it to you, capitalism, right? But if I build it and I give it away and I let other people build on top of it, not only does that ensure that that software continues to grow and become better, but it can transform into something completely different. And the language of that model could change as well.

So, if you believe that open source is a bad thing, please leave it in the comments below. And if you think it's a good thing, also tell me why. And if you want open AI to win, if you want Google to win, you know, that's capitalism at work right there. But for now, the emergence of low-cost, high-performance Chinese AI models is as much a moment for reflection as it is a cause for concern.

Should we be concerned about this model? At this time, we don't have any evidence that we should be. So use the open source model, check it out, grab its GitHub repo, test it out, tinker with it. If you're a software developer, it's a great way to check it out and get inside that code, dig in. I'm a software developer at heart. I build websites for businesses, willwalden.com, W-I-L-W-A-L-D-O-N.com.

I'm an engineer and I have been for about 20 years. So I'm very versed in this sector. And I believe that the safety that can come with open models like this is paramount. Collaboration, a huge, huge deal for open source software and the innovation that could come from this.

could change the landscape and take the power away from those that are in charge right now, like Sam Altman at OpenAI and Google and, you know, XAI, things like that. Take some of that power away. And I think we need some of that power in our hands as a community, as an open source community. Think of WordPress. Most of the web is powered by WordPress, right?

I don't know the exact stats, but it's somewhere it's over 50% is powered by WordPress. WordPress is an open source technology that anybody can change and fork and make into anything they want to make it into according to their terms. So something like that for AI that could power so many LLMs.

could power so many different chatbots, can power so many different companies for free and open. I think that's open for everybody.

Hey, thank you so much for listening today. I really do appreciate your support. If you could take a second and hit this subscribe or the follow button on whatever podcast platform that you're listening on right now, I greatly appreciate it. It helps out the show tremendously and you'll never miss an episode. And each episode is about 10 minutes or less to get you caught up quickly. And please, if you want to support the show even more, go to patreon.com slash stage zero.

And please take care of yourselves and each other. And I'll see you tomorrow.