The problem CanIBetOn solves

Bonding curve for prediction market: Liquidity in most prediction markets is thin and inconsistent, making it hard for users to launch novel or niche markets. Bonding curves solve this as well as providing a means for LPs to identify promising markets to support.

Prediction market parlay: Users currently lack a way to combine multiple outcomes into a single high-reward bet. High payout bets can encourage user engagement and provide higher profit margin to the bookmakers.

Prediction market AI oracle with dispute feature: Most oracles rely on centralized or opaque resolution mechanisms, risking incorrect outcomes without a transparent or decentralized way to challenge them. This is why existing prediction markets outside sports are very limited and using AI will allow it to scale.

Challenges I ran into

In building a more scalable and engaging prediction market, we’ve encountered key challenges across liquidity, combinatorial betting, and resolution. For bonding curves, implementing LMSR on Base introduces gas cost and contract complexity issues, especially post-graduation when markets need to scale or resolve. Liquidity remains fragmented across long-tail markets, and it's difficult to design graduation thresholds that don’t open up arbitrage risks or require excessive governance. For parlays, generating conditional odds is non-trivial—naively multiplying independent probabilities leads to pricing errors, and while AI offers a path forward, it’s not fully reliable or interpretable for sensitive financial predictions.

The biggest challenge with an AI-powered oracle is building trust and dispute resolution. AI can scale market resolution beyond sports but is prone to hallucinations and overconfident misstatements, especially in edge cases. Designing a dispute process that balances transparency, cost-efficiency, and incentive alignment is complex. Without robust dispute and correction mechanisms, even small resolution errors could lead to user distrust or manipulation. Ensuring fair and explainable resolutions at scale remains the core bottleneck for non-sports markets.