Almost all statistical question that ask about the "most people" tend to have the same results as asking: "Which countries have the most population?" So that answer is most likely "India."
What you might really be driving at is which countries have the highest percentage of their population that are users. But you'd need the # of users broken down by country to answer that, and would likely need to contact the various LLM companies to ask for that info. And if you are truly interested, you may as well do so - that sounds like data that would be good marketing, so they just might share it.
Thank you. Indeed, if we had user counts broken down by country, we could test many different hypotheses — and it seems India would lead.
Through this exchange with you, I realized that what I really wanted to know was: “Which countries have more people asking deep questions?”
However, to investigate that, we would probably need to know “how much computation AI is running in parallel to generate a response.” I’m thinking of asking OpenAI about this as a test.
Almost all statistical question that ask about the "most people" tend to have the same results as asking: "Which countries have the most population?" So that answer is most likely "India."
What you might really be driving at is which countries have the highest percentage of their population that are users. But you'd need the # of users broken down by country to answer that, and would likely need to contact the various LLM companies to ask for that info. And if you are truly interested, you may as well do so - that sounds like data that would be good marketing, so they just might share it.
Thank you. Indeed, if we had user counts broken down by country, we could test many different hypotheses — and it seems India would lead.
Through this exchange with you, I realized that what I really wanted to know was: “Which countries have more people asking deep questions?” However, to investigate that, we would probably need to know “how much computation AI is running in parallel to generate a response.” I’m thinking of asking OpenAI about this as a test.
[dead]