Rebuilding the Internet for Privacy | Barrett Lyon on DoxxNet
Thursday, 1 October 2026 · 3 min read · Listen to the episode ↗
Barrett Lyon argues that privacy is eroding as more communication and personal data pass through intermediaries. His DoxxNet aims to replace that model with encrypted peer-to-peer calls, chats, and file transfers that require no phone number, email, or name. Lyon says the company runs its routing and AI on infrastructure it owns, keeping contacts and AI inference within its environment rather than sending them to major AI providers.
Barrett Lyon describes the internet as reaching an inflection point with “coin-operated intelligence at scale,” while protocol design has stagnated, censorship is increasing, and privacy is eroding. He argues that more activity and personal data now pass through intermediaries’ servers. Messaging apps such as WhatsApp, Signal, iMessage, and Telegram may appear to connect people directly, he says, but messages still route through servers.
DoxxNet is Lyon’s alternative: encrypted peer-to-peer communication using transient connections, with chats that he says leave no record. It supports calling, video, chat, and file transfers without requiring a phone number, email address, or name. To join, a user solves a puzzle intended to show they are human and receives a token; inviting a friend initiates a cryptographic exchange that establishes trust and enables direct communication.
Lyon says files travel directly between phones rather than through an intermediary server. He describes transfers of up to 20 gigabytes, with no file-size limit, and connections using three or four encryption layers. He calls the system post-quantum. DoxxNet is not simply a conventional VPN, which typically sends traffic through leased servers: Lyon describes it as a broader, adjustable mesh ecosystem. It has 196 domains that do not exist on the public internet, he says, and can change domain and address leasing policies within milliseconds.
The company built its infrastructure “from software to wire” because Lyon believes organizations cannot set policy for equipment they do not own. DoxxNet’s routing and AI components run on its own hardware. Lyon says its AI environment keeps contacts and inference inside rather than sending them to major AI providers. He reports about 26 sites worldwide and about 12 people involved. The company has a US base and a Swiss company for European operations; Lyon says it chose to be open about how it operates and makes decisions. He contrasts that approach with VPN providers’ “no-logs” claims, arguing that such promises do not answer who owns the leased servers or controls their data policies.
Lyon says DoxxNet aims to reduce tracking across apps and websites. App telemetry can often be blocked without breaking the apps or most websites, he says, while warning that individually modest signals become revealing when combined. He also says intelligence agencies increasingly use data brokers and techniques associated with advertising agencies. In his view, networks should not make identity decisions: an IP address is not a driver’s license, and identity restrictions belong elsewhere in the technology stack. The speakers cite website restrictions tied to state Know Your Customer rules and a reported Cisco Umbrella incident in which a major carrier was classified as being in France, where rules required blocking more than 100 piracy sites.
Lyon expects DoxxNet may face denial-of-service attacks, including attacks enabled by users within its own network. He says privacy does not exempt users from following the law. His experience with DDoS dates to building Prolexic after attacks on his high-school servers; the business later became Akamai Prolexic. He also recalls early Prolexic work tracking Russian cybercriminals who extorted clients and caused outages, saying they went to prison for “I think eight years,” though he did not know what happened to them afterward.
This summary was generated from the episode transcript and can contain mistakes.