I've been pushing it all along before@ Mira_Network Also with Kaito Linked together Is it finally my turn to kiss? Mira fast TGE Recently AI I am actually focused on the 'data' aspect of the project, Agents I really played it all out, the most important thing in training a model is still to regress data labeling。 I think labeling data on the chain is the right direction, especially Mira We need to solve it AI「 The problem of hallucinations is currently causing LLM The reason for hallucinations is that the data is too mixed and false, Mira I want to put the process of data labeling and human verification on the chain, so that the model can consume trustworthy and traceable data. AI The problem with being strong is not computing power or parameters, but data quality, which is fundamental. Besides, I am also very concerned about the same DATA Related projects@ SaharaLabsAI ， But because I was too tired a few days ago po Wen did not reach the priority quota, KYC Remember to do it, I will vote at the beginning.