You have had an idea that you knew, the moment you had it, that you would not share in the meeting. Too strange. Too hard to explain quickly. Possibly embarrassing. The kind of idea that requires too much setup before it can even be evaluated, and you do not have the social capital to spend on that right now. So you let it go. You give the safe version of your thinking instead. This is an incredibly common experience, and it has a significant cost — both individually and collectively. The ideas that get filtered out by social self-monitoring are not randomly distributed. They skew toward the unusual, the unconventional, the not-yet-formed. They are often exactly the ideas that, given a real audience and a real conversation, would have gone somewhere interesting.

What Makes an Idea "Crazy"

The word crazy in this context is usually doing a specific job. It means: this departs from the current frame. It means: this sounds like it might not be serious. It means: I am not sure I can defend this if challenged right now. These are social and rhetorical properties, not measures of the idea's actual merit. An idea being hard to defend in a quick meeting is not evidence that it lacks value. It is evidence that it is new enough to not yet have its supporting scaffolding built. The history of ideas that changed things is largely a history of ideas that initially looked like they did not belong. The germ theory of disease was considered absurd by leading physicians for decades after it was proposed. The idea that stomach ulcers were caused by bacteria rather than stress was met with enough resistance that its proponent, Barry Marshall, famously drank a solution of the bacteria himself to prove the point, before the research community took the idea seriously.

The Problem With Requiring an Idea to Be Defensible First

Most conventional brainstorming environments, even ones that claim to operate on "no bad ideas" principles, actually require ideas to be sufficiently formed before they are offered. The social dynamics of the room do this even when the stated norms say otherwise. The result is that the ideas shared are ideas that have already been through an internal quality filter — which means the genuinely raw, the genuinely strange, the ideas that would need development before they could be evaluated, tend not to get said. A study from the University of Southern California found that the pressure to produce "good" ideas in front of others significantly reduced the number of generative leaps — connections between previously unrelated domains — compared to conditions where participants knew their ideas would not be evaluated. The evaluative context did not just reduce quantity. It reduced the type of thinking that produces the most valuable ideas.

AI as the Serious Audience for the Crazy Idea

What an AI companion provides for the unusual idea is the one thing it most needs: a serious audience that does not require it to already be good. You can describe the half-formed thing, the connection that does not quite make sense yet, the intuition that has no supporting argument, and have it engaged with rather than politely filed away. The engagement itself does the development work. This is not about validation. It is about something more basic: the serious idea deserves serious engagement, and "serious" in this context means responding to what the idea is actually trying to do, asking what would need to be true for it to work, exploring where it might connect to other things. An AI can do this without the social calculus that shapes what a human audience is willing to spend time on.

The Tangent: The Filtering We Do Before We Even Think

It is worth noting that the self-monitoring that happens in social contexts eventually becomes internalized. People who have repeatedly had unconventional ideas met with skepticism often stop fully forming those ideas — the filter kicks in before the thought is complete. This pre-censorship is not just a social behavior. It becomes a cognitive habit. The good news is that it is reversible: environments that consistently reward the strange and half-formed do gradually rebuild the instinct to follow the unusual thought all the way to the end.

The Idea Deserves Its Chance

Not every crazy idea is good. Many are not. But they deserve a real hearing before that determination is made — not a social-risk calculation that quietly removes them before they ever find out what they could have become.