II. Background: Leadership in Debate Interfaces
- Open research sharing: The author shared camp backfiles for six seasons, a norm debate coaches later formalized in the camp file-sharing initiative.
- The paperless template: The 2010 Word & Excel template, featured by Wired, pioneered the timer, Tilde-key speech docs, the Virtual Tub, the cite creator and Excel flowing, features later maintained by Verbatim and CardMirror.
- Virtual word-limit debates: A 2008 forum format capped constructives at 2,000 words and rebuttals at 1,200, sparking over 3,000 debates through 2022.
- A 2009 thought experiment: Web-like argument maps, keyword tagging, argument gardens, in-round quote suggestions, literature monitoring, integrated research and a public knowledge service, a vision that waited until language models arrived in 2023.
The author designed and developed the platform described here; this paper presents it as a case study. Innovation in how we debate and organize current-event issues has driven modern debate. The paper's author shared extensive camp backfiles for six seasons, providing research materials that built a vital community norm, then later the debate coaches followed with the camp file-sharing initiative (Gulakov, 2007). Then, the author innovated the paperless debate template (Gulakov, 2010) for Word & Excel, featured in a Wired article (Williams, 2012) and in the all-time most viewed video about debate (Wired, 2013), which pioneered core paperless debate features then maintained by the Verbatim template and later CardMirror web app. This includes the timer used for a decade, managing speech docs with the Tilde key, the Virtual Tub system built on the filesystem, designing a custom Word toolbar, the cite creator extension, many formatting fixes like small-sizing unread text, and Excel flowing shortcuts and casebook. The code was released under GPL license granting right to “convey verbatim copies of the Program's source code.” These features became standardized over decades, demonstrating a capacity to innovate ideas that endure, and led to the next step of speech sharing. Next, the author launched decades ago the first virtual debate platform on a popular debate forum, pioneering structured online debates with word limits (Gulakov, 2008). Constructives were capped at 2000 words, followed by cross-examination periods. Rebuttals were limited to 1200 words, with research allowed during rounds. Speeches were expected within 2-3 days, or delays announced to avoid forfeiture. This system sparked over 3,000 debates from 2008 to 2022 when the site went defunct as people moved to Reddit, providing an early model for later virtual debate formats. Finally, the paper's author imagined future interfaces in a popular debate blog essay, which would take decades of innovation to unlock (Gulakov, 2009). The essay was a childhood dream imagining the idealized vision of debate one could hope for:
- Web-like argument mapping: Central argument with branching responses in rings, mimicking mind's relative connections over top-down files.
- Keyword tagging: Highlight web text, tag with keywords to build dynamic "argument roots" storing full sources and analytics.
- Argument gardens: Shared, mergeable research hubs; private team or public/open versions for collaborative knowledge.
- In-round quote suggestions: Enter opponent keywords to auto-surface relevant snippets, analytics, and responses during rounds.
- Automated literature base monitoring: Full sources stored locally; updates propagate, enable contextual indicts tagged to sources.
- Integrated research: Tag articles continuously for all topic areas; no pre-split file creation vs. debating phases.
- Public knowledge service: Open gardens organize web info by logical arguments, like Wikipedia but multi-viewpoint.
After the author's policy debate experiences, which included debating alongside and defeating future NDT & TOC champions (Barkley Forum for High Schools, 2008), the author moved on to apply those research skills to AI development and design in Silicon Valley, from Google to a lead developer in AI startups. Because most observers underestimated the AI timeline, it was not until 2023 that language models made automated in-round research even thinkable. Beginning in 2023, the author applied prior software experience toward developing the platform described here: debate outlines as a model for AI governance and for news media streaming, providing global citizens relevant arguments that follow after every news article they read or report they watch. Additionally, the news reader can use AI Broker agents that use these outlines to auto-invest in stock and prediction markets (autoinvestment.broker).
