>>44667124 (me)Final post from this test. I'm really impressed with the new checkpoint, even if the underlying "smarts" memory, logic, etc. aren't too different the greater wordcount and vocabulary really improves the interactions, and this only gets more relevant when you use the model for fiction rather than these assistant type tests. Huge ups matrixbro and co, you've done amazing and I can't wait to see the future.
>Just remember that AI is still in its infancyPyg says it well. The wAIfus of 2030 are going to be crazy.