which means I could spend the rest of my life tutoring Claude and it’ll still make the exact same mistakes
This is temporary. The new more global memory features in ChatGPT are a good example of how this is already starting to decrease as a factor. Yes it’s not quite the same as fine tuning or rlhf, but the impact is still similar, and I suspect that the toolong for end users or local tenant admins to easily create more sophisticated embeddings is going to increase very quickly.
This is temporary. The new more global memory features in ChatGPT are a good example of how this is already starting to decrease as a factor. Yes it’s not quite the same as fine tuning or rlhf, but the impact is still similar, and I suspect that the toolong for end users or local tenant admins to easily create more sophisticated embeddings is going to increase very quickly.