Delivering last mile UX for dApps with purpose-built models and agents

• Almost all centralized apps we interact with today use AI/ML models: Models to rank products on Amazon, flagging malicious content on social media platforms, recommendation for stocks by robo-advisors etc.

• For decentralized apps to compete with centralized apps, they will need AI/ML in a similar capacity to improve UX eg. @PondGNN and optimize core aspects of the offering.

• The data needed to train such models and agents will need to come from off-chain businesses who have done it successfully as well as on-chain data generated from user behaviour.

• Deploying a model for on-chain apps that are similar should take <5 minutes through composability primitives while benefitting the data owners and developers who have explicit ownership, and monetization rights accelerating the pace of innovation and adoption.