Crypto x AI Yupp AI obtains preference data by incentivizing users to explore and compare the latest AI models for subsequent DPO and post-DPO model training. a16z led an investment round of $33 million Team: Founders have extensive experience in artificial intelligence and cryptocurrency. They jointly built consumer-level machine learning products early on Twitter. Pankaj Gupta was responsible for global consumer engineering at Google Pay and Coinbase. Gilad Mishne previously served as a machine learning director at GoogleX. The early team already includes senior engineers from Google, Coinbase, and top research laboratories. My perspective: High-quality data is extremely valuable for model training. In the web2 domain, Scale AI (acquired by Meta for $1.5 billion) emerged, along with Mercor, which initially recruited data industry experts for Scale AI and AI labs, gradually involving themselves in data annotation services. Currently, their business is more long-tail, primarily serving smaller model companies, but still holds potential. Web3 blockchain technology adds value to such businesses through transparency and economic incentives. User participation: yupp.ai After entering, users can earn points by annotating model output quality 6000 points can be used for withdrawal, 1000 points = 1 USDC Withdrawal is allowed only once per day Points may potentially serve as future airdrop credentials

In addition, it has received support from more than 45 individual angels and small investors, including Google Chief Scientist Jeff Dean, Twitter co-founder Biz Stone, Pinterest co-founder Evan Sharp, Perplexity CEO Aravind Srinivas, Cred CEO Kunal Shah, four professors from Stanford University (Professor Dan Boneh, Professor Chris Re, Professor Nick McKeown, Professor Balaji Prabhakar), Othman Laraki, Paul Grewal, Gokul Rajaram and Coinbase Ventures. This investment lineup is really strong

In addition, they will generate a model evaluation list service yupp.ai/leaderboard