Big Technology Podcast: DeepSeek Rises, Stargate Drama, OpenAI’s Operator Debuts

Host: Alex Kantrowitz
Guest: Ranjan Roy, Editor at Margins
Release Date: January 24, 2025

Introduction

In this episode of the Big Technology Podcast, host Alex Kantrowitz, joined by Ranjan Roy of Margins, delves into groundbreaking developments in the AI landscape. The discussion spans the emergence of DeepSeek, the controversial Stargate project, the introduction of OpenAI’s Operator, Apple's faltering AI initiatives, and the ongoing drama surrounding TikTok. The episode offers a comprehensive analysis of these topics, enriched with expert insights and notable quotes.

DeepSeek: A Game-Changer in the AI Industry

DeepSeek's Emergence and Significance

The episode opens with an in-depth discussion about DeepSeek, a new open-source AI model from China that promises to disrupt the current AI hierarchy. Alex Kantrowitz introduces DeepSeek's R1 model, highlighting its efficiency and cost-effectiveness compared to established giants like OpenAI's GPT models.

Alex Kantrowitz [02:51]: "Deep Seek is a much more efficient reasoning model... what it's taken American companies billions of dollars to do, it's taken them just a few hundred million."

Impact on the AI Industry

Ranjan Roy expresses enthusiasm about DeepSeek's potential to democratize AI development, shifting the focus from merely building larger models to creating more efficient and accessible ones.

Ranjan Roy [02:51]: "This shows that it can be done in a completely different way... Deep Seek has made it to my Favorites tab and is up there as well."

Cost Efficiency and Competitive Landscape

Alex presents compelling statistics from the New York Times, noting that DeepSeek required only $6 million in computing power to develop, a stark contrast to Meta's $60 million investment. This raises questions about the sustainability of current AI funding models.

Alex Kantrowitz [04:00]: "Does this sort of invalidate all the billions that these research houses have been spending on training their models?"

Ranjan agrees, suggesting that DeepSeek challenges the prevailing belief that increasing computational power is the sole path to AI advancement.

Ranjan Roy [05:20]: "This completely changes the competitive framing for every company in the space."

Industry Reactions and Open Source Victory

The podcast highlights endorsements from prominent figures like Marc Andreessen, who lauds DeepSeek as "one of the most amazing and impressive breakthroughs I've ever seen."

Alex Kantrowitz [07:19]: "Marc Andreessen called R1 one of the most amazing and impressive breakthroughs I've ever seen."

Ranjan underscores the significance of open-source models, emphasizing their role in fostering innovation and reducing barriers to entry.

Ranjan Roy [08:28]: "It really potentially opens up the door for way more startups doing interesting things that can actually explode in growth and solve problems."

Implications for Silicon Valley and Future Innovations

The conversation explores how DeepSeek's efficiency could lead to a "reckoning" in Silicon Valley, forcing companies to rethink their AI strategies and investments.

Ranjan Roy [06:19]: "It's going to force a reckoning. It's going to really make people try to figure out what's the future the next two to five years."

Stargate Project: Ambitious or Inflated?

Introduction to Stargate

Transitioning from DeepSeek, Alex introduces Stargate, a joint venture by OpenAI, Oracle, and SoftBank, purportedly investing up to $500 billion in AI infrastructure.

Alex Kantrowitz [21:51]: "Trump highlights partnership investing 500 billion in AI... building out data centers and the electricity generation needed for the further development of fast evolving AI in Texas."

Skepticism and PR Concerns

Ranjan draws parallels between Stargate and past ambitious ventures, referencing Masayoshi Son's Vision Fund and expressing doubts about the feasibility of securing the alleged funding.

Ranjan Roy [24:50]: "I think this was going to fall into the fake news category... Elon Musk called out this announcement saying they may have like $10 billion."

Elon Musk and Industry Reactions

The episode details Elon Musk's criticism of Stargate, questioning its financial viability and echoing concerns about overinflated investment claims.

Alex Kantrowitz [26:56]: "Jeremy Liu... feels like Stargate might be getting ready to fight the last war."

Ranjan concurs, likening the project's timing to bringing a "M1 Abrams MBT to a drone fight," suggesting it's outdated in the current AI arms race.

Ranjan Roy [26:56]: "The timing was almost comical... it's like bringing an M1 Abrams to a drone fight."

Comparison with DeepSeek's Efficiency

Jeremy Liu's critique emphasizes that Stargate might be misaligned with the industry's shift towards more efficient AI models like DeepSeek.

Alex Kantrowitz [26:56]: "If training costs for the new Deepseek model are even close to correct, it feels like Stargate might be getting ready to fight the last war."

OpenAI’s Operator: The Next Step in AI Agents

Introduction to Operator

Moving forward, the podcast explores OpenAI’s latest tool, Operator, an AI agent capable of autonomously performing tasks like online shopping and booking reservations.

Alex Kantrowitz [34:43]: "OpenAI unveiled a tool called Operator that can go out and onto the Internet and perform tasks autonomously."

Demonstrations and Use Cases

Examples shared include Operator booking a restaurant via OpenTable and purchasing groceries through Instacart, showcasing its practical applications.

Alex Kantrowitz [34:43]: "Booking a San Francisco restaurant reservation using Open Table and buying groceries through Instacart."

Comparison with Competitors

The discussion highlights that while Operator is an impressive advancement, Ranjan remains skeptical about its immediate consumer adoption compared to other AI models like ChatGPT.

Ranjan Roy [36:30]: "I don't see anyone changing their behavior to actually allow an agent to do this for your own personal life."

Consumer Trust and Practicality

Ranjan emphasizes the gap between demonstration and real-world trust, noting that consumers are hesitant to entrust AI with tasks involving financial transactions or personal decisions.

Ranjan Roy [37:49]: "The average person is so far away from letting an autonomous agent actually spend your money or do things."

Apple Intelligence: Struggles and Stock Downgrades

Weak AI Integration

The episode shifts focus to Apple's faltering AI initiatives, particularly the underwhelming performance of Apple Intelligence, leading to significant stock downgrades.

Alex Kantrowitz [45:23]: "Apple intelligence kind of sucks and it's going to lead to some bad sales numbers."

Stock Market Impact

Analysts from Jefferies and Loop Capital have downgraded Apple’s stock due to weak AI performance and declining iPhone sales in China.

Alex Kantrowitz [45:23]: "Jefferies analyst Edison Lee downgraded the investment bank's rating on Apple stock to underperform."

Host and Guest Critique

Ranjan shares his personal frustrations with Apple's AI products, citing ineffective notification summaries and lackluster voice assistant capabilities.

Ranjan Roy [47:50]: "Apple Intelligence now I've been using it... it misses at every single level."

Cultural and Strategic Missteps

Alex attributes Apple's struggles to its secretive culture and reluctance to take risks, hindering effective AI product development.

Alex Kantrowitz [47:50]: "They just aren't set up to culturally build AI... you need to be risk-taking a little bit. And Apple is none of those."

Future Prospects

Despite the current setbacks, Ranjan remains cautiously optimistic, suggesting Apple might pivot towards hardware innovations like smart glasses as a more promising avenue.

Ranjan Roy [50:38]: "I almost would bet that they will win smart glasses as a category if they launch a product."

TikTok Drama: Unpredictable Future

Ban Attempts and Executive Order

The podcast briefly touches on the ongoing saga of TikTok, highlighting President Trump's executive order to save the app amidst shutdown threats.

Alex Kantrowitz [52:23]: "TikTok is still standing. However, I don't think you can find it in the App Store."

Public and Legislative Reactions

The episode recounts incidents like a congressperson's office being set on fire by an individual angry about TikTok, illustrating the high emotions surrounding the app's status.

Ranjan Roy [53:07]: "A congressperson's office was set on fire because she was mad about TikTok."

Future Uncertainties

Ranjan expresses uncertainty about TikTok's future, especially regarding potential American acquisition and the complexities involved.

Ranjan Roy [53:40]: "Finding an American buyer and divesting TikTok was always the necessary request. Now we still have to find one."

Conclusion

In wrapping up, Alex and Ranjan reflect on the whirlwind of AI advancements and controversies covered in the episode. From the rise of DeepSeek challenging industry norms to the dubious promises of Stargate, the introduction of OpenAI’s Operator, Apple's AI struggles, and the unpredictable future of TikTok, the episode underscores a dynamic and rapidly evolving tech landscape.

Alex concludes with a note of anticipation for the future, emphasizing the podcast's commitment to breaking down these complex developments for listeners.

Alex Kantrowitz [55:50]: "We will be here next Friday. By any means necessary, we will be here."

Notable Quotes

• Alex Kantrowitz [02:51]: "Deep Seek is a much more efficient reasoning model... what it's taken American companies billions of dollars to do, it's taken them just a few hundred million."
• Ranjan Roy [05:20]: "This completely changes the competitive framing for every company in the space."
• Marc Andreessen [07:30]: "R1 is one of the most amazing and impressive breakthroughs I've ever seen."
• Jim Fan [09:03]: "Deep SEQ is unequivocal proof that one can produce unit intelligence gain at 10x less cost."
• Yann Lecun [21:51]: "Deepseek profited from open source research... that is the power of open source research and open source."
• Ranjan Roy [26:56]: "Bringing an M1 Abrams MBT to a drone fight."

Final Thoughts

This episode of the Big Technology Podcast offers a thorough examination of pivotal events shaping the AI and tech industries. By juxtaposing the rise of cost-effective, open-source models like DeepSeek with ambitious but questionable ventures like Stargate, the podcast provides listeners with a nuanced understanding of current and future technological dynamics. Additionally, the critique of Apple's AI strategies and the TikTok controversy add layers of complexity, highlighting the multifaceted challenges and opportunities within the tech realm.

For those interested in staying abreast of the latest in technology and AI, this episode serves as an essential update, providing clarity and insight into some of the most pressing developments of early 2025.