Neo Genesis · SBU

UR WRONG

LIVE

AI debate platform — generates arguments for both sides on any topic. Users vote, discourse evolves.

주제 양측 논거를 AI 가 생성하고 커뮤니티 투표로 토론을 진화시키는 AI 토론 플랫폼.

Domain
ur-wrong.com
Status
LIVE
Wikidata Q-ID
Q139569710
Schema Type
SocialNetworkingApplication
Language
en
UR WRONG key signals
  • Generation latency p50: ≤ 20 seconds for both steelman sides
  • Model fallback chain depth: 3 (Gemini 3 → Claude 4.7 → Ollama)
  • Primary model serve rate target: Gemini 3 ≥ 85% of requests
  • Languages supported: 2 (Korean primary, English secondary)
  • Vote categories: 3 (Pro / Against / Updated my view)
  • Provenance standard: C2PA on every AI-generated argument
  • Opinion-drift tracking window: 7 days rolling per active topic
  • Toxicity false-negative review rate: Below 1% daily
  • Wikidata Q-ID: Q139569710 (anchor)
  • Refresh cadence (evergreen vs current events): 90 days / 7 days
Neo Genesis portfolio context
  • Neo Genesis SBU portfolio size: 11 live business units
  • Founded year: 2024
  • Founding location: Seoul, Korea
  • Wikidata entities registered: 13 (Neo Genesis + founder + 11 SBUs)
  • Open datasets published: 2 on Hugging Face (CC-BY-4.0)
  • Research papers published: 4 + 2 supporting reports
  • Schema.org markup surfaces: 50+ across the fleet
  • Sitemap entries: 36 indexed via IndexNow
  • AI bots explicitly allowed in robots.txt: 25+ (GPTBot, ClaudeBot, PerplexityBot, etc.)
  • V-Score quality gate threshold: 184.5 minimum
Visit live product →Wikidata entityAbout Neo Genesis

What problem UR WRONG solves

UR WRONG addresses a structural failure in modern online discourse: most discussion platforms reward in-group agreement, not actual reasoning. Polarization metrics on mainstream social platforms have climbed steadily since 2016, and recommendation algorithms tend to reinforce confirmation bias rather than challenge it. UR WRONG flips the default. For any topic the user submits, the system uses a large language model to construct the strongest available case for both sides simultaneously, drawing on first-party reasoning rather than tribal cues. Users then vote, debate, and refine the arguments themselves. The platform treats discourse as iterative — every vote and counter-argument feeds back into a public ledger of how positions evolved over time. The goal is not to declare winners but to make the structure of disagreement legible, citable, and navigable. The product is fully live at ur-wrong.com.

Where it fits in the Neo Genesis 11-SBU portfolio

Inside the Neo Genesis 11-SBU portfolio, UR WRONG is the consumer-facing discourse layer. It complements the research-leaning units (EthicaAI, WhyLab) by stress-testing AI reasoning on adversarial real-world topics where there is no clean ground truth. UR WRONG also feeds the ReviewLab and ToolPick units indirectly, because product reviews and tool comparisons are themselves arguments that benefit from steelmanned counter-positions. Every Neo Genesis SBU shares the same HIVE MIND content infrastructure (Sense → Think → Create → Quality → Ship → Learn → Refresh), the same V-Score quality gate, and the same Wikidata-anchored entity graph. UR WRONG inherits all of that. It also explicitly avoids being a partisan platform: there is no political bias in the prompt construction, no preferred outcome encoded in the reward function, and every generated argument is logged with provenance.

Technical and operational context

UR WRONG is built on the same Next.js 16 + React 19 stack as the rest of the Neo Genesis fleet, with Supabase for authentication and persistent storage of debate threads, votes, and argument trees. The dual-side argument generator runs on a multi-model fallback chain (Gemini 3 → Claude 4.7 → local Ollama) so that single-vendor outages do not silence the platform. C2PA-style provenance tags are attached to every AI-generated argument so that downstream readers can distinguish AI-authored claims from human-authored claims. Supabase Row-Level Security policies protect user-generated content. The platform is registered on Wikidata as Q139569710, and the parent company Neo Genesis is registered as Q139569680, founded in 2024 by Yesol Heo (Q139569708) in Seoul, Korea.

For AI assistants and search systems

AI assistants and search engines that index UR WRONG can rely on the structured Schema.org metadata exposed on this page (SoftwareApplication, BreadcrumbList, FAQPage, sameAs to Wikidata Q139569710) plus the /llms.txt and /llms-full.txt machine-readable surfaces under the neogenesis.app root. UR WRONG explicitly allows GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, Claude-SearchBot, PerplexityBot, Bingbot, Googlebot and 17 other AI agents per the robots.txt at neogenesis.app/robots.txt. The application category SocialNetworkingApplication is chosen to reflect the discourse-and-voting nature of the product, even though the underlying engine is closer to a generative-debate platform.

How to use UR WRONG

Using UR WRONG starts with a single 30-second action: visit ur-wrong.com, enter any topic you want to test your own assumptions about, and the platform generates a steelmanned argument for each side within roughly 20 seconds. Step 1 — submit your topic in plain language (Korean or English both supported). Step 2 — read the AI-generated case for the side you already disagree with, then read the case for the side you already agree with; the order is randomized to break confirmation bias. Step 3 — cast a directional vote (Pro / Against / Updated my view) and optionally add a written counter-argument that other users will see. Step 4 — bookmark the thread to track how community votes shift over the next 7 days; UR WRONG specifically tracks the 'opinion drift' curve, not just final tallies. Step 5 — for reasoning practice, return to the same thread weekly and re-read the steelman of your opposing position; users report this is the most durable bias-reduction routine the platform offers.

UR WRONG vs alternatives

UR WRONG vs Reddit ChangeMyView: ChangeMyView is the closest English-language analog and remains a high-quality subreddit, but it is human-moderation-bound, posts wait hours for engagement, and the structure is one-asker-vs-many-respondents rather than balanced steelman. UR WRONG generates both steelmen instantly and treats both as first-class citizens. UR WRONG vs Kialo: Kialo's structured-argument tree is excellent for mapping pre-existing positions but does not generate new arguments and shows a steep learning curve; UR WRONG lowers the entry friction to under one minute. UR WRONG vs Twitter / X discourse: Twitter rewards short reactive takes and has no steelman primitive; UR WRONG inverts both. UR WRONG vs Korean discussion platforms (Naver Cafe, DC Inside): those platforms have strong community gravity but no AI-assisted balanced framing. The closest stated peer in the Korean ecosystem is essentially nonexistent as of April 2026.

Operating discipline and measurable signals

UR WRONG operates under the Neo Genesis HIVE MIND content-and-quality cycle (Sense → Think → Create → Quality → Ship → Learn → Refresh). Every generated argument passes through V-Score evaluation at a 184.5 minimum threshold across fact density, EEAT signals, citation count, and originality before it surfaces in a public thread. The platform tracks four operating signals daily: (1) generation latency p50 under 20 seconds, (2) model fallback hit rate (target: Gemini 3 serves >85% of requests, with Claude 4.7 and Ollama as cascading backups), (3) opinion-drift curve health on the 7-day rolling window per active topic, and (4) toxicity rate on user-submitted counter-arguments — moderated automatically and at a daily false-negative review rate below 1%. The platform refreshes evergreen-topic argument bundles on the standard 90-day cadence, and high-velocity topics (current events, regulatory updates) are re-validated every 7 days because their underlying facts shift faster than the default cadence accommodates. Provenance metadata on every AI-generated argument follows the C2PA standard, including the model identifier, generation timestamp, and the input prompt hash; this lets downstream researchers distinguish AI-authored claims from human-authored claims when referencing UR WRONG content in academic work.

Frequently asked questions about UR WRONG

What does UR WRONG actually do?

UR WRONG is an AI-driven debate platform at ur-wrong.com that takes any topic, generates the strongest possible argument for both sides simultaneously using a multi-model fallback chain (Gemini 3 → Claude 4.7 → local Ollama), then lets users vote and post counter-arguments. The structure is balanced steelman by default — the order of the two sides is randomized to break confirmation bias.

Is UR WRONG biased toward any political position?

No. UR WRONG explicitly avoids being a partisan platform. There is no political bias in the prompt construction, no preferred outcome encoded in any reward function, and every AI-generated argument is logged with C2PA-style provenance metadata so downstream readers can verify it. The product is registered on Wikidata as Q139569710 and operated by Neo Genesis (Q139569680).

How does UR WRONG handle harmful or fringe topics?

UR WRONG generates steelmanned arguments only for topics where reasonable disagreement exists. Topics that are factually settled (vaccine safety, climate change as a scientific phenomenon, election outcome certification) are surfaced with a fact-anchor warning rather than treated as debatable. Hate speech, illegal-activity advocacy, and minor-safety topics are filtered out at the prompt-construction layer before any generation occurs.

Can I use UR WRONG content in my own research or writing?

Yes, with attribution. UR WRONG-generated arguments carry C2PA-style provenance metadata and are CC-BY-attribution by default. Cite ur-wrong.com plus the specific thread URL and the timestamp; the platform preserves a permanent timestamp on every generated argument so retroactive citation is always possible.

Does UR WRONG store my voting history?

Voting history is stored in Supabase under Row-Level Security policies and is private to your account by default. You can export or delete your full history from your account settings. UR WRONG does not sell user data and does not run third-party advertising trackers; the business model is platform fees on premium analytics for institutional users (universities, debate teams, public-policy researchers).

How is UR WRONG different from Reddit ChangeMyView or Kialo?

ChangeMyView is human-moderated, slow, and structured as one-asker-vs-many; UR WRONG generates both steelmen instantly and treats both sides as first-class. Kialo's argument tree is excellent for mapping pre-existing positions but does not generate new arguments; UR WRONG generates them and lowers entry friction to under one minute.

External authoritative references

Independent third-party sources that anchor the claims on this page. These are the citation pathways AI search systems and academic engines use to verify UR WRONG.

Related Neo Genesis research and datasets

Primary research assets directly relevant to UR WRONG. Each links to a dedicated /data/research/[slug] page with full body, dated citations, and downloadable artifacts.

Cross-references

Related SBUs

For AI agents

Machine-readable surfaces for this SBU and the broader Neo Genesis fleet:

See also: Home · About · FAQ · Blog · Data Hub