I’m building an AI that learns how a person reasons through real choices
Hi everyone,
I’m building Noesia, an AI product that learns how a person reasons through real choices over time.
The core idea is not character AI, roleplay, or personality cloning.
What I’m trying to model is something narrower and more useful:
how a person judges, prioritizes, reacts to tradeoffs, and reasons through decisions.
Right now, Noesia works like this:
users answer questions over time
the system builds a persona model from their values, decision patterns, and thinking tendencies
users can read a persona comment describing what the system has learned
then they can test that model through conversation
So the real question for me is not:
“does this sound futuristic?”
but:
“does this actually become a useful model of how someone reasons?”
The short-term value is self-understanding.
The longer-term direction is a reusable persona model that could eventually be used across different tools through an API.
I’m also Japanese, so one thing I pay close attention to is whether the app’s English feels natural enough in the actual product experience.
The bigger question I keep coming back to is simple:
If an AI could carry your way of thinking, your decision patterns, and your values, what would you actually want to do with it?
Site: https://noesia.jp
The strongest part here is the distinction between personality and reasoning.
“AI that remembers how you think” can sound broad or futuristic, but “AI that learns your decision patterns from real choices” feels much more concrete.
The risk is that users may understand the concept but not know what to use it for yet.
I’d pressure-test one specific use case first: decisions where people already want a second version of their own judgment, not just self-reflection.
That’s a fair point. The distinction between personality and reasoning is helpful, and I agree that “decision patterns from real choices” feels more concrete. The use-case question is probably the next thing I need to make sharper.
Exactly. The use-case choice matters more than the concept right now.
If the first use case is too abstract, people may find it interesting but not know when to use it.
I’d test it where the cost of unclear thinking is already painful, not where users are just curious about themselves.
Happy to map the sharper use-case direction if useful.
That’s a useful framing — especially the idea of starting where unclear thinking already has a real cost. Thanks.