Research Lemmas

These are sources of truths that I think will be in AGI.

1. Bitter lesson → just let the model learn
   • Focus on architectures that can scale with data. End to end is the way. Don’t get into the way of these models