BrainCloud

Inspiration

The inspiration for BrainCloud comes from the persistent challenges faced by travelers worldwide in obtaining real-time, personalized travel advice. We sought to create a solution that leverages the power of community wisdom and advanced AI. Moreover, we were inspired by the potential of tokeneconomics within the realm of crypto to enable large-scale Reinforcement Learning with Human Feedback (RLHF) across the globe. We realized that by using token incentives, AI models can be fine-tuned much quicker and achieve greater accuracy than general models. Thus, the idea for BrainCloud was born - to combine the power of AI, RLHF, and tokenomics to revolutionize travel research.

What it does

BrainCloud is a ground-breaking AI-driven platform that provides up-to-the-minute, bespoke travel advice for cities across the globe. Users refine city-specific AI models with their input, earning Feedback Tokens ($FBT) in return. These tokens are then used to access exclusive travel insights, enhancing users' experiences. Holders of the FBT token can also lock their tokens to earn a percentage of all fees across all the models and also gain governance rights to vote on proposals to drive the protocol forward. Leveraging Chainlink Automation, veFBT token holders now will get rewards on a weekly basis from all the users around the world using the AI models.

How I built it

BrainCloud was built using a combination of OpenAI fine-tunable models, Solidity/Foundry for the Feedback Token system, and React/Nextjs for UX/UI design principles to ensure a smooth user experience.

Challenges I ran into

Building a robust, reliable AI model that could learn from diverse user inputs was a significant challenge. Also, integrating the blockchain-based Feedback Token system within the platform presented technical complexities.

Accomplishments that I'm proud of

We're proud to have created a platform that not only delivers personalized travel advice but also encourages user participation through a rewarding system. Users all over the world can now provide the very best up-to-date information about the area they live in and travelers can now get better information in real time. We're also proud of our city-specific AI models, which have shown remarkable performance in early tests.

What I learned

Building BrainCloud taught us valuable lessons about RLHF, AI and blockchain integration with these systems and how we could rewards users for providing valuable information.

What's next for BrainCloud

We aim to expand BrainCloud's capabilities to cover even more cities worldwide, refine our AI models further, and grow our user base. We're also planning to introduce more features to make FBTs even more valuable for our users through fiat payment methods and a more robust subscription model. We also plan to scale into other areas outside of travel/city info.

Built With

• foundry
• nextjs
• openai
• python
• react
• solidity
• typescript