> it was said these chatbots cost way more to run than they make profit Well if it's running on Azure and using massively overpriced Nvidia data center GPUs I

> it was said these chatbots cost way more to run than they make profit Well if it's running on Azure and using massively overpriced Nvidia data center GPUs I can't imagine anything else would be possible. Then again it's not like there any incentive for OpenAI to increase efficiency as long they get 'free' Azure credits (and it's not like the real cost for MS is anything close to what they are supposedly investing into OpenAI. IRRC that 10 billion was mainly not in actual money)