This looks like a first generation model to bootstrap future models from, not a competitive product at all. The knowledge cutoff is pretty old as well. (2023, seriously?)
If they wanted to train it to have some character like Anthropic did with Claude 3... honestly I'm not seeing it, at least not in this iteration. Claude 3 was/is much much more engaging.
If they wanted to train it to have some character like Anthropic did with Claude 3... honestly I'm not seeing it, at least not in this iteration. Claude 3 was/is much much more engaging.