In late 2024, I was on a long walk with some friends along the coast of the San Francisco Bay when the question arose of just how much of a bubble we live in. It’s well known that the Bay Area is a bubble, and that normal people don’t spend that much time thinking about things like AGI. But there was still some disagreement on just how strong that bubble is. I made a spicy claim: even at NeurIPS, the biggest gathering of AI researchers in the world, half the people wouldn’t know what AGI is.

As good Bayesians, we agreed to settle the matter empirically: I would go to NeurIPS, walk around the conference hall, and stop random people to ask them what AGI stands for.

Surprisingly, most of the people I approached agreed to answer my question. ^[1] I ended up asking 38 people, and only 63% of them could tell me what AGI stands for. Some of the people who answered correctly were a little perplexed why I was even asking such a basic question, and if it was a trick question. The people who didn’t know were equally confused. Many simply furrowed their brows in confusion. Some made a valiant attempt—I heard a few artificial generative intelligences and even an Amazon general intelligence.^[2]

Judging from the response I got on X (the everything app), this was a very surprising outcome. I ended up running this experiment again at NeurIPS 2025 with an even bigger sample size (n=115).^[3]

After this first experience with surveying people, it became clear to me that the next step was to venture further outside the bubble, and survey the general US population. It turns out that this is already somewhat of a solved problem. A lot of people care about what the average American thinks. The market, in its infinite wisdom, has provided a solution—you can just pay pollsters to run random questions.

It’s impossible to actually sample from the distribution of all Americans. So you find some other approximate distribution, such as the distribution of all people who answer polls on the internet in exchange for amazon gift cards. You ask them a thousand demographic questions, like “how old are you” and “how much money do you make”. Then, since we know from the US census what these distributions are supposed to look like, you can correct for the distribution shift using importance sampling.^[4]

With this caveat in mind, I embarked on a journey to ask normal Americans a b...

justanothertuesday

justanothertuesday