Keep Everyone In Sync: Effective Approaches Borrowed from Open Source Communities

## Overview

In this talk, security researcher Hong Kong (from Alibaba Group, researching IoT, V2X, and cyber-physical systems) addresses the crisis of information overload and communication breakdown in cybersecurity communities. Drawing a comparison to Nyquist’s theorem—which states that sampling must occur at double the signal rate to reconstruct a complete picture—he argues that modern researchers face an "Internet of bankruptcy": overwhelming floods of unread emails, instant messaging spam, endless RSS feeds, and a collapse of human focus. Rather than relying on modern, proprietary, and centralized communication tools, the speaker proposes adopting resilient, open, and asynchronous communication patterns from open-source software (OSS) communities and early Internet architects.

## The Information Crisis and Failure of Modern Communication

Modern tools and workflows have introduced severe friction into researcher productivity: * **The Bankruptcy Phenomenon:** Users experience email bankruptcy (dating back to practices observed around 2004 of ignoring backlogs), instant message saturation, unread read-later lists, and severe attentional degradation. According to referenced studies and *Time* magazine, internet-connected human attention spans have dropped to 8 seconds (shorter than a goldfish's reported 9 seconds). * **AI Categorizers & Algorithmic Filters:** The speaker cautions against trusting AI message categorizers. These filters exhibit bias toward serving positive updates while hiding critical feedback or chances for improvement. Linus Torvalds experienced a notable categorizer failure in 2015 when ~20% of Linux kernel emails were mistakenly tagged as spam. * **Proprietary & Centralized Chat Platforms:** Instant messaging relies on centralized servers, proprietary protocols, and push notifications. Serious discussions are displaced by stickers, superficial interactions, and lack of persistent, searchable archives. * **Hardware & UX Regression:** High-speed camera experiments show modern computer interfaces have higher input latency and less responsiveness than older terminals. In line with Andy and Bill's Law, software bloat negates hardware gains, leaving users waiting on machines instead of machines serving users.

## Lessons from Open Source Pioneers

The speaker reviews how notable figures manage high information throughput: * **Linus Torvalds:** Maintained kernel communication for over 20 years using text-based email clients `pine` and `alpine` without complex nested folders, storing messages in large unified repositories and relying on mirrors. * **Marissa Mayer:** Managed 700–800 emails daily by using `pine` for dedicated catch-up marathons lasting up to 14 hours on weekends. * **Richard Stallman:** Avoids cellular devices and browses the web asynchronously by emailing URLs to a server that returns page contents via email. * **Chris Lamb (Debian):** Created a GTD-style email system using scripts that defer messages across rotating folders (e.g., `delay.7days`) so items resurface when ready for review. * **Greg Kroah-Hartman & Donald Knuth:** Kroah-Hartman swore off IRC to eliminate distractions. Donald Knuth ceased direct email handling in 1990, relying on an assistant to screen and print essential correspondence.

## Pull Models, Open Protocols, and Community Architecture

To build a sustainable, productive community, the speaker highlights architectures that prioritize open access, asynchronous retrieval, and decentralized routing: * **Pull vs. Push Architecture:** Instead of push notifications that interrupt work, communities should use pull models. The speaker demonstrates connecting to the public UNIX system `SDF.org` (operational since 1987) and reading Linux Kernel, Debian, and Mozilla development discussions via NNTP, Usenet/Gmane bridges, and anonymous IETF IMAP servers without subscribing to noisy push mailing lists. * **Inclusive Tiering:** Member participation should accommodate seniors, regulars, newcomers, and random contributors without requiring continuous live presence or proprietary clients. * **Routing Ideas via BGP & Internal Directories:** Referencing Dunbar’s number (the theoretical limit of maintaining ~148 stable connections), the speaker compares community scaling to Border Gateway Protocol (BGP). Rather than top-down reporting hierarchies that breed information asymmetry, communities should advertise member skills and needs like BGP route metrics, passing questions and inspirations to the right nodes (mirroring Stephen Wolfram's "who knows what" internal directory at Wolfram Research).