The New Internet

🌈 Abstract

The article discusses the vision and big picture behind Tailscale, a networking company founded by Avery Pennarun and the Davids. It explores the evolution of the internet, the problems with modern software development, and Tailscale's approach to building a "New Internet" that addresses these issues.

🙋 Q&A

[01] The Vision for Tailscale

1. What is the big vision for Tailscale, according to the article?

• Tailscale is building a "New Internet" that avoids the problems of the "Old Internet" by:
  • Providing every device with a certificate, IP address, DNS name, end-to-end encryption, and identity, allowing them to be peers rather than just clients
  • Enabling connectivity and bypassing firewalls without adding latency or overhead
  • Removing layers of complexity and "gunk" that have built up in modern software development and cloud computing

2. Why does the author believe Tailscale is uniquely positioned to pursue this vision?

• The founders, being "old" and having experienced the evolution of computing and networking over the decades, have the context to see the problems with the current state of software development and the internet
• Their experience allows them to take a different approach and build solutions that remove unnecessary complexity rather than just adding more layers on top

3. What is the "chain of dominoes" that the author believes Tailscale can start to fall?

• Fixing the underlying connectivity issues will address the centralization and rent-seeking behavior of cloud providers, allowing for a new wave of distributed computing and innovation

4. What is the "catch" or challenge the author identifies in realizing this vision?

• Tailscale needs to achieve widespread adoption to create a "New Internet" that developers can build upon, otherwise it will suffer from a "chicken-and-egg" problem where not enough people have Tailscale, so no one builds apps for the "post-Tailscale world"

[02] The Problems with Modern Software Development

1. What are the key problems the author identifies with modern software development?

• Developers are too focused on scaling and building for hypothetical "billion user" scenarios, even for simple applications
• This leads to unnecessary complexity, overhead, and suboptimal solutions that are slower and harder to build and maintain
• The industry has spent more time making the "hard things possible" rather than making the "easy things easy"

2. How does the author characterize the current state of the tech industry?

• The "tower of complexity" has become so tall that developers are considering using large language models (LLMs) to write the "incomprehensible code in the incomprehensible frameworks"
• Modern software development is mostly "junky overhead" rather than actually solving the original problem

3. What does the author see as the underlying cause of these problems?

• The connection of local networks (LANs) to the internet, leading to increased firewalls, attackers, and degradation of simplicity
• The internet has made networking fundamentally complex, and attempts to paper over these problems with additional layers of complexity have only made the situation worse

[03] Tailscale's Approach and the "New Internet"

1. How does the author describe Tailscale's approach to addressing these problems?

• Tailscale is building a "New Internet" at the OSI layer 3 level, based on new assumptions that avoid the problems of the "Old Internet" in the first place
• This allows Tailscale to remove unnecessary layers of complexity and enable simpler, more efficient solutions like the Taildrop file transfer example

2. What does the author see as the key to unlocking the potential of the "New Internet"?

• Widespread adoption of Tailscale, so that developers can build applications that take advantage of the improved connectivity and reduced complexity it provides
• The author compares this to the adoption of the internet itself, which was slow at first but then became ubiquitous

3. What is the author's vision for the future impact of Tailscale's "New Internet"?

• Just as the distributed computing enabled by Microsoft's Windows led to new innovations, the "New Internet" enabled by Tailscale will allow developers to invent things that were "unthinkable in the old centralized world"
• This will disrupt the current centralized cloud computing model and the rent-seeking behavior of major providers like AWS