Okay, so public opinion polls haven't exactly been stellar lately, right? I mean, it's tough to trust those numbers completely. That's where this new company, Simile, comes in, shaking things up with a hefty $100 million in venture capital. Their big idea? Replacing traditional polls with AI. It sounds wild, but hear me out.

Simile claims it's building a "foundation model" that can predict what people will do in pretty much any situation. How? They train AI agents by feeding them data from chat-style interviews with real folks. These agents then become like "digital twins" or "digital clones" of those individuals. They even add in data about people’s actual behavior and shopping habits to make them more realistic. The idea is that market researchers can then chat with these digital clones to get insights.

Now, this might sound a bit like that game, The Sims. Remember spending hours micromanaging your Sim's life, from their job to their love life? Apparently, Simile took some inspiration from it! A research paper co-authored by Simile's CEO mentions creating an interactive environment where AI agents interact with each other, just like in The Sims.

CVS, for example, is already using Simile's simulated people to get insights on things like store layouts and new product designs. One CVS exec even said it's a "really big unlock" because they can ask as many questions as they want without tiring out the simulated people. I can see the appeal – imagine being able to test out new ideas without having to worry about real-world limitations.

Gallup is also partnering with Simile to simulate asking policy questions to large groups. It's basically a chatbot that lets you ask "the group" anything. The goal is to make these model decisions transparent, replicable, and validated against real-world sentiment. I don't know about you, but I'm a little skeptical about how well this will actually work. Can an AI really capture the nuances of human opinion? That's the big question.

While I'm intrigued by the potential of AI to revolutionize market research, I also have some reservations. It's important to remember that these are just simulations. They're based on data, but they're not the same as talking to real people. We need to be careful about relying too heavily on AI-generated insights and ensure that we're still grounding our decisions in the real world.