been thinking about how mira and cysic ai approach verifiability differently mira uses a multi-ai model to fact-check each other before sending out output (more like a peer review) cysic ai uses a multi-proof model via the combination of TEE and ZK proofs to verify ai actions not saying one is better than the other, just think they're designed for different use cases while mira can be used for research and analysis, cysic ai will excel in the agentic economy just thinking out loud btw, tell me wyt?