The Cost of Large Language Models

There is a lot of pain around latency again what we talked about in those key findings. Certain use cases just cannot be done you can't do fraud detection with large language models because the latency is a blocker. There are also ethical concerns around hallucinations and attaching your brand to these hallucinations. I jokingly put that my new year's resolution is to be as confident as chat GPT when it just spews out b****** on us.