Understanding How Facebook Disappeared from the Internet: Who Owns the Generative AI Platform?

Understanding How Facebook Disappeared from the Internet: Who Owns the Generative AI Platform?

In a world where technology reigns supreme, it is important to understand the inner workings of the internet and the various platforms that drive it. Two recent events shed light on different aspects of this vast digital landscape: Facebook's sudden disappearance from the internet and the discussion surrounding the ownership of generative AI platforms. While these topics may seem unrelated, they both provide valuable insights into the complex nature of our interconnected online world.

To comprehend how Facebook vanished from the internet, we must first familiarize ourselves with the concept of Border Gateway Protocol (BGP). BGP is a mechanism that allows the exchange of routing information between autonomous systems (AS) on the internet. It ensures that internet routers know how to handle data and maintain connectivity. Without BGP, the internet would cease to function. It is the glue that holds together the network of networks that make up the internet.

Every network, including Facebook, has an Autonomous System Number (ASN) and an internal routing policy. These ASNs need to announce their prefix routes to the internet using BGP to establish connections and facilitate communication. However, on a particular day, Facebook stopped announcing the routes to their DNS (Domain Name System) prefixes. This led to the unavailability of their DNS servers, effectively disconnecting Facebook and its sites from the internet. It was as if someone had abruptly pulled the cables from Facebook's data centers, severing their connection to the outside world.

This incident highlights the critical role BGP plays in maintaining internet connectivity. It serves as a reminder that even giants like Facebook are not immune to technical glitches or human errors that can disrupt their online presence. It also emphasizes the interdependence of different networks and the importance of effective routing protocols to ensure seamless communication across the internet.

Moving on to the topic of generative AI platforms, we delve into the fascinating world of artificial intelligence and its applications. Generative AI refers to AI models that can create new content, such as images, text, or code. The potential of generative AI has captured the imagination of many, leading to the rapid growth of applications in various domains. Image generation, copywriting, and code writing are just a few areas that have already exceeded $100 million in annualized revenue.

However, despite the exponential growth of generative AI applications, the question of ownership and commercialization remains a challenge. Infrastructure vendors seem to be the major beneficiaries of this market, capturing a significant portion of the revenue. Application companies, on the other hand, struggle with issues like retention, product differentiation, and gross margins. While model providers are responsible for fueling the growth of generative AI, they have yet to achieve large-scale commercial success.

One key observation is that selling end-user apps may not be the only or best path to building a sustainable generative AI business. Technical differentiation alone may not guarantee long-term customer value. Building network effects, holding onto data, or creating complex workflows can enhance customer retention and drive differentiation. Additionally, there is a growing demand for proprietary APIs and hosting services for open-source models. These platforms facilitate the sharing and integration of models and create indirect network effects between model producers and consumers.

The generative AI market also raises ethical considerations, given its potential for both great achievements and harmful consequences. Many model providers have incorporated the public good explicitly into their mission, organizing as public benefit corporations or issuing capped profit shares. Surprisingly, this has not hindered their fundraising efforts. However, it sparks a discussion about whether capturing value should be the primary goal for model providers in this space.

When examining the financial flow in generative AI, it becomes evident that a significant portion of revenue goes to infrastructure companies. On average, app companies allocate 20-40% of their revenue to inference and fine-tuning, paid either to cloud providers or third-party model providers. Consequently, approximately 10-20% of total revenue in generative AI finds its way to cloud providers. Nvidia, a major player in the data center GPU market, has emerged as a significant winner due to the extensive use of GPUs in generative AI workloads.

The infrastructure layer in generative AI is characterized by standard moats such as scale, supply-chain, ecosystem, algorithmic, distribution, and data pipeline moats. However, the durability of these moats over the long term remains uncertain. It is yet to be seen if strong network effects will take hold in any layer of the stack. The success of generative AI is likely to be influenced by both horizontal and vertical companies, with the best approach dictated by end-markets and end-users. Verticalization, where the user-facing app is tightly coupled with a home-grown model, may be the winning strategy when the AI itself is the primary differentiator. On the other hand, horizontalization may occur when AI is part of a larger feature set.

In conclusion, the disappearance of Facebook from the internet and the ownership dilemma in generative AI provide valuable insights into the intricate workings of our digital world. The reliance on BGP for internet connectivity serves as a reminder of the importance of robust routing protocols. Meanwhile, the discussion surrounding generative AI platforms sheds light on the challenges and opportunities in this rapidly evolving field. As we navigate the ever-changing landscape of technology, it is essential to understand the underlying mechanisms and explore innovative approaches to drive progress.

Actionable Advice:

1. Prioritize the stability and security of routing protocols like BGP to ensure uninterrupted internet connectivity. Regularly review and update routing configurations to mitigate potential risks.
2. For companies venturing into generative AI, consider a multi-faceted approach that goes beyond selling end-user apps. Explore opportunities for creating network effects, retaining data, and building complex workflows to drive long-term customer value.
3. When considering the ownership and commercialization of generative AI platforms, carefully assess the balance between capturing value and incorporating the public good into your mission. Explore innovative business models that align with ethical considerations and foster sustainable growth.