I'm building Noesia, an AI that learns and reproduces how you think
Hi everyone, I’m currently building a beta product called Noesia.
Most AI products are designed to help people get answers faster. With Noesia, I’m exploring a slightly different direction: using AI to make a person’s way of thinking more visible, and gradually reproduce their thought patterns and values over time.
Noesia learns from a user’s answers, then accumulates their decision-making style, values, and thinking tendencies as a persona model. The goal is not just to chat with AI, but to create an experience where AI helps you understand yourself more deeply.
Noesia has three core modes: a question mode that builds the persona, a talk mode that lets you interact with it, and a persona comment mode that explains it in text. In general, the best flow is to answer questions first, then check the persona comment, and then try the conversation mode.
I’m also Japanese, so I can’t fully judge on my own whether the AI’s English behavior feels natural in subtle ways. Because of that, I’d especially appreciate feedback on whether the wording, tone, and personality feel natural and consistent when used in English.
A few things Noesia is trying to do:
Learn from accumulated responses, not just one-off prompts
Build a persistent persona model over time
Let users interact with the persona as it becomes more defined
Use AI not only as a productivity tool, but also as a mirror for self-understanding
It’s still in beta, so I’d especially love feedback on:
Does this concept feel genuinely useful?
Does it feel like your way of thinking and your values are actually being reproduced?
What kind of users do you think this would resonate with most?
Does the AI’s behavior in English feel natural, especially in phrasing and tone?
What would make the first experience easier to understand?
If this sounds interesting, I’d really appreciate any honest feedback.
https://noesia.jp
The concept is interesting because it is not just “AI that answers,” it is closer to a personal reasoning model that becomes more accurate as the user keeps giving it signal. That distinction matters. If the first experience only feels like a chatbot, people may miss the deeper idea. I’d make the onboarding prove one thing quickly: “this AI noticed a thinking pattern about me that I did not explicitly say.”
One naming point: Noesia has a thoughtful/self-knowledge feel, but the product may eventually move beyond reflection into persistent AI persona infrastructure. If that is the long-term direction, a name like Viryxa.com could carry the AI/persona-model layer more sharply.
Thanks, this is really helpful.
I think your point about onboarding is especially important. If the first experience feels like “just another chatbot,” people will miss the real value. I like your framing that the product should quickly show: “it noticed a thinking pattern about me that I didn’t explicitly state.”
The naming point is also interesting. Right now I’m still thinking of Noesia more from the self-understanding side, but I agree that if it grows into persona infrastructure, the naming layer may need to become sharper.
Really appreciate the thoughtful feedback
Exactly. Noesia works if the product stays mostly in the self-understanding/reflection lane.
But if the product becomes something closer to a persistent persona model that learns how someone thinks, decides, reacts, and reasons over time, then the name probably has to carry more technical weight.
That is the distinction I’d keep watching.
Reflection product: Noesia can work.
Persona intelligence layer: the name may need to feel sharper, more ownable, and less soft.
Viryxa.com fits that second direction better because it sounds more like AI infrastructure around identity/persona than a journaling or self-reflection app.
I wouldn’t force the change now, but I’d keep it in mind before the product gets too far into the market under one frame.
That distinction makes a lot of sense. I’m increasingly thinking in the persona model / infrastructure direction, so this is something I’ll keep in mind carefully. Really appreciate you taking the time to articulate it so clearly.
That makes sense.
If you are increasingly leaning toward persona model / infrastructure, then I’d definitely keep the naming layer in mind before Noesia gets too deeply associated with reflection only.
The product category people remember early can be hard to unwind later.
Also, just to be transparent, I own Viryxa.com. I mentioned it because it genuinely fits the sharper persona infrastructure direction better than the softer self-understanding frame.
Not pushing a change now, but if you ever want to seriously compare the naming direction, happy to discuss it directly. If there is a real fit, I can keep the pricing very founder-friendly and make the transfer simple.
A quick note on the 3 modes in Noesia:
Question Mode (Learn Mode)
The AI asks you questions, and you answer them in your own words. This is the mode that builds your persona model. As a rough guide, I recommend continuing until around Lv50 so the model has enough data to reflect your thinking more reliably.
Talk Mode
This is the conversation mode. You talk with the AI after it has learned from your answers, so the responses are based on the persona that has been formed.
Persona Comment Mode
This mode explains your formed persona in text. It gives you a readable summary of your tendencies, values, and style of thinking.
Recommended flow:
Question Mode -> Persona Comment Mode -> Talk Mode