1.h. Models hub        2 gCrawl:  Decentralized high performance algo-generated training data solution
                2.a. Crawling task as a transaction
                2.b. Crawled data as an asset
                2.c. Fault-tolerant distributed high performance crawling
                2.d. RAFT consensus algorithm
                2.e.  Decentralized data storage
                2.f.   DFSM decoding path
                2.g. Transfer learning for updating the model parameters
       3. gPredict: Decentralized gradient learning framework
                3.a. Training task as a transaction
                3.b. Trained model as an asset
                3.c. Floating Point mapping
                3.d. Hyper-Parameters routing
                3.e. Results reducing
       4. gCompute: Internet-scale AI dApps solution
                4.a. AI model as an asset
                4.b. Scalable
                4.c. Controllable   
                4.d. Difficult      
                4.e. Maintainable
                4.f. Responsive
                4.g. Concurrent