# Long Bet Project > Make accountable long-term predictions with skin in the game, forcing rigorous thinking about future outcomes through public commitments and charitable stakes - Skill: `lev-os/long-bet-project` (Agent Skill) - Install (CLI): `npx skillmds@latest add lev-os/long-bet-project` - Raw SKILL.md: https://api.skillmd.com/api/skills/lev-os/long-bet-project/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: lev-os (https://skillmd.com/u/lev-os) - Updated: 2026-09-10 - Page: https://skillmd.com/skills/lev-os/long-bet-project --- # Long Bet Project ## Overview The Long Bet Project, created by the Long Now Foundation in 2002, is a public arena for accountable predictions about the future. It addresses a critical problem: pundits, CEOs, and thought leaders routinely make confident predictions with zero accountability. Media's short attention span lets them escape consequences when wrong. Long Bets demands predictors put their name, a solid argument, and money down in support of their statement. Winnings go to charity. This mechanism transforms cheap talk into costly signals - if you're wrong, a cause you oppose might benefit. The minimum term is 2 years with no maximum, encouraging thinking across decades or centuries. The framework operationalizes prediction accountability: your forecast becomes a matter of public record, your reasoning is documented, and resolution is guaranteed by institutional continuity. Warren Buffett's famous 10-year bet against hedge funds (won in 2017) demonstrated the model's power to settle debates definitively. ## When to Use - Making long-term predictions you want held accountable - Settling debates that won't resolve for years or decades - Forcing rigorous thinking about future outcomes - Creating skin in the game for forecasts - Building institutional memory around predictions - Testing expertise claims publicly - Funding charities through competitive prediction - Challenging overconfident forecasters to back their words ## The Process ### Step 1: Formulate a Clear, Falsifiable Prediction State exactly what will happen by when. Binary outcomes only - no hedging. **Poor prediction:** "AI will change everything by 2030" **Strong prediction:** "By December 31, 2030, a commercially available AI system will have won a Grammy for Best New Artist" ### Step 2: Build Your Argument Document WHY you believe this outcome. Include: - Causal mechanism (what forces make this outcome likely) - Evidence supporting your position - Acknowledgment of key uncertainties - Why the timeline is appropriate **Example argument:** "Foundation models demonstrate emergent musical composition. Grammy's 'Best New Artist' eligibility requires human performance, but AI-generated personas with human performers already chart. Commercial incentives push labels toward AI collaboration. Timeline based on 10-year cycles in music industry disruption." ### Step 3: Set Meaningful Stakes Minimum $200 per side, no maximum. Consider: - Amount large enough to create genuine accountability - Amount you can afford to lose entirely - Charity you'd want to fund if you win - Charity you'd prefer NOT fund (opponent's choice) **Stakes design:** Higher stakes = stronger signal of conviction. Buffett bet $1M. Match stakes to confidence level. ### Step 4: Publish and Invite Challenge Make your prediction public. Either: - Post on longbets.org ($50 publication fee) - Announce publicly with clear terms and escrow mechanism **Critical:** True names required. Anonymous predictions don't create accountability. ### Step 5: Wait and Let the Future Adjudicate Resist the urge to hedge or revise. The prediction stands as written. **Resolution:** Mutual agreement between parties if both alive, otherwise Long Now Foundation adjudicates. Public announcement and charitable disbursement. ### Step 6: Analyze the Outcome (Win or Lose) Win: Document what you got right and why your reasoning held. Lose: Conduct honest post-mortem - where did your model break? **Key insight:** The learning value exists regardless of outcome. The process of rigorous prediction improves forecasting ability over time. ## Example Application **Situation:** Tech industry debate about cryptocurrency future (2018). **Application:** - **Prediction:** "By December 31, 2028, at least one sovereign nation will use Bitcoin as its primary legal tender for all domestic transactions" - **Argument:** El Salvador adoption trend, hyperinflation in emerging markets, Lightning Network scalability, generational shift in money perception - **Stakes:** $5,000 per side - **Charities:** Winner's choice - Electronic Frontier Foundation; Opponent's choice - Institute for Humane Studies - **Publication:** Posted on longbets.org with full reasoning **Outcome tracking:** El Salvador adopted BTC as legal tender in 2021, but "primary" for "all domestic transactions" remains unmet. Prediction continues. ## Scoring Rubric - **Practitioner Weight:** 8/10 - Real-money stakes from notable figures (Buffett, Bezos donors) - **Clarity/Executability:** 9/10 - Extremely clear process with institutional support - **ROI:** 7/10 - Indirect benefits through improved forecasting discipline - **Novelty:** 8/10 - Unique mechanism combining prediction markets with charity - **Cross-domain:** 8/10 - Applicable to any falsifiable future claim **Total: 40/50** ## Anti-Patterns - Making vague predictions that can't be definitively resolved - Setting stakes too low to create real accountability - Using pseudonyms or anonymous accounts (defeats accountability) - Refusing to publish reasoning (hides from scrutiny) - Revising predictions after initial publication - Betting on unfalsifiable outcomes ("AI will be important") - Ignoring the post-mortem learning opportunity - Only making predictions you're highly confident about (no learning) ## Related - superforecasting (systematic prediction accuracy improvement) - tetlocks-10-commandments (forecasting best practices) - skin-in-the-game (accountability through personal risk) - cathedral-thinking (multi-generational timeframes) - thinking-in-bets (probabilistic decision-making)