This video features Sequoia Capital partner Konstantin Buhler appearing on Bloomberg Technology to explain how important "memory" is in artificial intelligence (AI) and why the United States must remain at the forefront of algorithmic innovation. It covers a broad range of topics including the US-China AI competition, new technology trends, and real-world examples of the latest AI protocols. It thoughtfully explains how the key to future AI development lies in software innovations like "memory" and emphasizes the spirit of collaboration in the American innovation ecosystem.

1. Recent Infrastructure Investment and the National Significance of the AI Industry

The video opens with a discussion about infrastructure deals over the past 24 hours -- specifically US IT companies expanding into the Middle East (Gulf states) to build data centers and establish development hubs.

Konstantin explains this movement:

"These new partnerships and investments show that AI is no longer just about corporate success -- it has become a national-level imperative."

In other words, the influence of the AI industry has grown to the level of national importance. He recalls that just ten years ago, there were many concerns about whether the US could maintain its leading position in AI.

He identifies four pillars that effectively sustain AI development: compute, power, data, and algorithms. Among these, algorithms are actually the most important, and the reason the US was able to stay ahead of the world for a long time was precisely because of that fourth pillar -- algorithms.

"The US has the world's best researchers and engineers. It's important to remain at the forefront of algorithmic innovation going forward."

2. The Rise of AI Memory and Why It Matters

While countries are investing heavily in hardware and infrastructure, the topic gaining increasing attention in the industry is AI memory. When the host asks why this topic is rapidly gaining prominence, Konstantin explains:

"Just as important as compute, data, power, and algorithms is one more element: 'memory.' When you interact with an agent (like an AI chatbot), it's not just about the agent remembering information about 'you.' The AI must also remember itself."

In other words, memory goes beyond simple information storage or user history logging -- it refers to the AI's ability to maintain 'continuity,' 'identity,' and context on its own. He illustrates this with the example of a doctor:

"When a doctor sees a patient, they don't just look at data or vitals -- they can recall past interactions themselves and even restore the manner of communication."

Tools like 'Open Evidence' are mentioned as examples, and he emphasizes that as AI memory advances, AI will be able to remember "the way it communicates" based on past experience, much like a human expert. He points out that "memory" will become an increasingly important factor in AI going forward.

3. The US-China AI Competition and Research Environment

After a brief news interruption about the President's Qatar economic cooperation announcement, the conversation naturally shifts to China's AI industry growth and the global distribution of AI researchers. The host poses the hypothesis that "more than 50% of AI research talent is concentrated in China" and asks whether the US and Chinese AI industries face similar challenges.

While acknowledging this, Konstantin emphasizes that the US also has the most creative and talented researchers in the world, and cites a recent AI conference organized by Sequoia as an example.

"At this year's annual conference, industry leaders like Jensen Huang (NVIDIA) and Sam Altman (OpenAI), along with 150 young talents, gathered together. Among them, the standout topic was 'Tool Use' -- the ability to use tools."

Tool Use refers to new approaches that enable AI systems to collaborate and interact with each other.

4. AI Collaboration and Innovation: The Model Context Protocol (MCP) Case

The latter half of the video, where the real innovation content is concentrated, introduces the "Model Context Protocol (MCP)" which has seen significant industry progress recently.

"Think of every piece of software as an expert in a different field. A CRM (customer relationship management system) specializes in remembering past interactions with customers. But these experts -- these software systems -- don't communicate well with each other."

"MCP acts like a kind of 'universal translator' that enables these diverse AIs to communicate smoothly with each other. This is a critical turning point for the US to stay ahead in the AI competition."

As an example, he introduces 'Rocks,' a Sequoia portfolio company. Rocks' AI:

1. Uses MCP to connect software from various departments when analyzing potential customers
2. Automatically creates customized pitch decks for customers
3. When needed, directly integrates with 'Claude Code' or 'Cognition' to even produce actual demos

In this way, "collaboration" and "flexible connectivity" between AI solutions is exactly the secret by which the American innovation ecosystem maintains its competitive advantage.

"In this way, American researchers and engineers actively collaborate with each other to ultimately lead the AI future."

Conclusion

The video thoughtfully explains that AI innovation is determined not only by hardware infrastructure but equally by software advances such as algorithms, "memory," and communication methods between AIs. It concludes that America's strength lies in the fact that its top talents continually collaborate and build new protocols and collaborative ecosystems.

"Future AI becomes more powerful not by being smart alone, but by understanding each other well and joining forces."

Key takeaways:

• The future competitiveness of AI depends on memory and inter-AI collaboration protocols.
• The US is leading the latest trends such as MCP through close collaboration among the world's top innovators.
• At the heart of innovation always lies connection, collaboration, and creativity.