The best way to have complex discussions?
A new open-source tool, CQ2, aims to structure complex online conversations by anchoring nested comment threads to specific snippets of text and allowing threads to be “concluded” with summaries. Commenters compare it to Usenet, email threading, Slack, Zulip, Kialo, and Google Docs, arguing over whether deep threading, DAG-style reply graphs, and AI-assisted summaries can genuinely make long-form, asynchronous collaboration easier to follow. Many note that social norms, moderation, and UX (especially mobile support and thread discoverability) may matter more than data structures alone, but still see value in experiments that better surface consensus, track unresolved points, and reduce repetitive or fragmented exchanges.
Comparisons to Existing Tools
- Many see CQ2 as similar to Google Docs/Quip comments, Notion, or email threads, with the main novelty being deeper, structured threading plus a “conclude” state.
- Others compare it to Zulip, Discourse, Kialo, Roam, Reddit/HN, old phpBB/vBulletin forums, and Usenet clients; opinions differ on whether CQ2 is genuinely new or mostly a rearrangement.
- Several argue Usenet-style threaded readers and classic forums already solved long-form, complex discussion better than modern chat systems.
- Some liken it to Google Wave or imageboards (4chan-style post IDs and back-references), seeing the same basic graph-of-replies idea.
Threading, Structure, and Graph Models
- Strong split between advocates of deep threading for complex topics and advocates of simple linear timelines with manual quoting.
- Repeated concern that most people don’t understand or won’t use threads properly; they default to top-level replies, making a mess.
- Multiple proposals to model discussions as DAGs or general graphs, allowing replies to multiple parents, re-parenting, and better cross-linking between related subthreads.
UX, Adoption, and Mobile
- Some find the demo confusing and navigation (scrolling, “back”, mobile) frustrating; lack of mobile support is heavily criticized.
- Others argue optimizing for desktop-only may actually help serious, focused discussion, but accept it hurts adoption.
- Suggestions include visible breadcrumbs, activity “hotspots”, better discoverability of concluded threads, and more obvious use of the “conclude” button.
AI, Summaries, and Consensus
- Strong interest in LLMs to summarize threads, cluster arguments, identify agreement/disagreement, maintain TL;DRs, and extract action items from calls.
- Ideas for AI to build an “idealized” debate map or debate tree, constrain redundancy, and help re-route new comments into existing relevant threads.
- Concerns raised about algorithmic curation hiding minority views and about surveillance and chilling effects if all speech is transcribed and stored.
Culture, Moderation, and Participation
- Many argue tooling can’t fix bad discussion culture; good faith, norms, and moderation matter more than structure.
- Some propose filtering out a large fraction of users or requiring strong identity and commitment to constructive behavior.
- Anonymity is seen as both enabling truth-telling and inviting abuse; trade-offs depend on context.
Perceived Niche and Limits
- Enthusiasm for CQ2 in domains like RFCs, academic debates, debate “flowing”, and collaborative paper revisions.
- Skeptics see it as a sideways move or reinvention of email/forum paradigms, not a breakthrough in “complex discussions.”