I'd love to know the demographic breakdown there. For instance I know students in some relatively isolated corners of the world are still, ubiquitously, using it to cheat. And a quick search gives around 1.4 billion students globally.
Obviously it's not just students, but the fact it [numerically] could be is rather amusing and makes me surprised that it's only 1 billion active users, since GPT has a lock on name recognition amongst low information users.
I'm naive, but is this possible? The power required per user to run those GPUs seems like enough to run an air conditioner. For a billion people everyday or so seems like enough to overload power plants.
> The power required per user to run those GPUs seems like enough to run an air conditioner.
Not true. Maybe you're thinking of computer games? Inference requires much less energy per user session than a session of your average AAA game title. And there are more active gamers at any time than active ChatGPT users.
You’re comparing 1 billion people doing bleeding edge gaming to LLM usage. Feels a bit like your thumb is on the scale here. Even under lower load GPU’s do in fact use a lot of power, as evidence by the massive power buildouts that we are seeing across multiple countries.
Yes Silicon Valley gives things away to build a high DAU count before the squeeze we’ve seen it for almost 2 decades now. That tells us nothing of the longterm health of their business.
