We know from the kinic team and services hosting Llama2 that ICP can host LLMs and store additional knowledge by embedding and provide specialized AI services, but is it possible to attempt to generate LLMs from training? Also, is it possible to host an alpha version of openCog, which SingularityNet will release in Q1 2024?
It seems very important to provide AI services on-chain for transparency and verifiability, but if we can commit to AGI that may come out earlier than expected, it becomes even more important and we can contribute.
It is our understanding that Dfinity does not currently have a specific roadmap with respect to AI. It does not have to be specific, but we would like to know how you intend to get involved in the AI field.
1 Like