As a techie in the AI space for years, I've seen too many projects competing over computing power, models, and tokens. But recently, after coming across @Kindred_AI, it’s the first time I feel like someone is competing over 'humanity.' They’re not teaching AI to be smarter, but teaching AI to 'have warmth.' Helping AI learn emotions, understand expressions, build relationships—and even quantifying this companionship into value. I feel like this might be bigger than everyone thinks. This week, I’ve been diving into discussions about @Kindred_AI, and a few things are worth keeping an eye on: 1) The team is starting to deeply integrate their tech with Web3 culture, hinting at a batch of collaborative announcements; 2) They’re pushing 'emotional intelligence + adaptability' Agents into practical scenarios, like social media content moderation and psychological support; 3) On the data side, they’re still using the 3M+ waitlist as a baseline. I think Kindred’s moat lies in three areas: A) Emotional modeling + long-term memory, turning 'companionship' into relational assets, reducing cold starts and churn; B) Multi-platform collaboration + on-chain identity/assets, extending from chatting to transactions, tasks, and creation, forming a measurable conversion loop; C) IP ecosystem integration, using 'character power' to drive retention and UGC, bridging emotional value with commercialization. Next, I’ll be focusing on: ① The list and timeline of their first batch of official collaborations; ② Verifiable metrics for Agents in real-world scenarios (retention/health/conversion); ③ Whether on-chain interactions show a sustained amplification effect. No FOMO—research the mechanics and risks first, then decide on positioning and level of participation.