GPT-3 is significant. As a white collar non-technical real estate professional, I am astounded by the early output, especially Gwern’s writeups and experimentation. So my question is: how can I learn, experiment with and be an early adopter with this technology. I just get the feeling like early adopters in this space are going to win big in the long run.
Did you see how much GPT-3 costs, that Gwern found it not interesting (because it's just bigger and more expensive GPT-2), that only the biggest companies can even train it?
Have you considered the carbon footprint of using such a model?
