Me too, the problem might be that the costs to run queries might be really high? I made some back-of-the-envelope calculations as https://beta.sayhello.so http

Me too, the problem might be that the costs to run queries might be really high? I made some back-of-the-envelope calculations as https://beta.sayhello.so https://beta.sayhello.so (based on ChatGPT) results are often better than Google. The energy alone will be really high for the amount of queries I do in Google. No idea how to make this sustainable. Hopefully I'm wrong, or more efficient GPUs/TPUs come out in the near future.