Proprietary Data and Network Effects as AI Moats

much discussion about proprietary data as the moat in AI Interesting to watch AI startups create proprietary data by being simply earlier to market. Then find a wedge, get momentum, then get proprietary interaction data. These are all the edge cases, corrections, and human-in-the-loop decisions your competitors never see. The theory is to build new models, UX, and capabilities around that The question is if these moats are enough versus the upgrades in the SOTA models I am still interested to see how folks get network effects going. Openclaw a recent example of a dev community spawning around a framework and people picking it bc the functionality will be richest. That’s the natural moat but others will be found too