
Written by a model from the transcript; not yet reviewed.
In this episode, A16Z's Joel de la Garza sits down with internet infrastructure pioneer Barrett Lyon to discuss Doxnet, his new venture aimed at rebuilding internet architecture from the ground up to restore privacy and security. Lyon explains why standard VPNs and messaging apps fall short by relying on centralized servers and third-party infrastructure, and how Doxnet leverages decentralized mesh routing, proprietary hardware control, and automated AI administration to provide secure peer-to-peer communication. Along the way, Lyon shares remarkable stories from his career, including surviving a plane engine failure, tracking down cybercriminals in Russia, and mapping the entire internet for an art project that ended up in the Museum of Modern Art. The conversation highlights the technical realities of modern internet surveillance, the erosion of protocol design, and the urgent need to protect open infrastructure from manipulation.
Marc Andreessen welcomes listeners and introduces the episode’s focus on internet privacy and Barrett Lyon’s work.
Barrett shares his high‑school server business, pilot license, and a dramatic in‑flight emergency.
The hosts debunk myths about data center water usage and discuss the importance of owning physical infrastructure.
Barrett details how DoxxNet creates a fully encrypted peer‑to‑peer mesh, enabling large file transfers and calls.
Discussion of how current messaging apps route through servers, the erosion of privacy, and the stagnation of internet protocols.
Barrett talks about his MoMA artwork, philosophy background, and how art informs his technical work.
Marc and Barrett wrap up, reflecting on the future of privacy, the internet’s open spirit, and upcoming projects.
Barrett Lyon explains how a growing traffic commute in the Bay Area led him to start flying a Cirrus airplane to work in Tahoe, which eventually resulted in a terrifying mid-air engine failure over Oakland.
The Post-Crash Reality Checkstory
After successfully landing a failing plane at Buchanan Field, Lyon describes the sudden wave of fear that hit him only after a first responder shook his hand.
Why VPNs Fall Shortinsight
Lyon argues that commercial VPNs are merely entry and exit points leased from third-party landlords, leaving users exposed to whoever controls the underlying servers.
Serverless Peer-to-Peer Networksstrategy
Doxnet bypasses traditional servers entirely by letting user devices connect directly over an encrypted mesh network, allowing large file transfers with zero middlemen.
Running Infrastructure with AIdata
Lyon reveals that his 12-person company runs 26 global routing sites without a single network administrator or sysadmin, handling maintenance entirely through AI automation.
Mapping the Internet for Artstory
A casual lunch bet over network mapping techniques led Lyon to create the Opti project, an internet visualization that became part of the permanent collection at MoMA.
“The really big problem is on WhatsApp, Signal, iMessage, Telegram. All of those applications disguise themselves as you sending something to somebody else.”
“If you're running, well, for me, like when you're organizing a camp or dealing with, you know, oh, we ran out of water on the art car. It helps because you can communicate.”
“The internet was designed to connect people directly. Today, much of what we do online passes through someone else's server, along with an enormous amount of data about us.”
a16z’s Joel de la Garza sits down with DoxxNet founder Barrett Lyon to discuss what it would take to rebuild the internet around privacy, direct communication, and less dependence on centralized intermediaries. Barrett explains why he thinks traditional VPNs only solve part of the problem and how DoxxNet is building a parallel mesh network where users can communicate peer-to-peer, transfer large files, make calls, and message without routing those interactions through a central application server. He also explains why the company owns its infrastructure and runs its AI systems locally rather than relying on third-party inference providers. doxxnet_assembly They also get into the growing amount of tracking embedded across the internet, what happens when that data can be analyzed by increasingly capable AI, and why Barrett believes the underlying protocols of the internet are due for another wave of experimentation. Along the way, he shares lessons from decades of building internet infrastructure, from fighting early DDoS attacks to mapping the internet and building global networks.
8,499 words · 3 speakers
Speaker names marked are inferred from speaking time and the show’s known hosts — they are not voice-verified.
Page updated · transcript and summary are machine-generated