Why I build it
One AI model can sound completely certain while another model gives a different answer to the same question. Most apps show you only one of them, so you never see the disagreement.
I wanted to see both: where the models meet and where they part. The Observer sends one question to several models, lets them work on it together, and then writes one answer that tells you how consistent they were.
Five principles
- Show the uncertainty.If the models disagree, or the evidence is thin, the answer should say so instead of hiding it.
- A mirror, not an oracle.The app should help you think. The final judgement is yours.
- Many views before one answer.Comparing several models catches mistakes that a single model would repeat with confidence.
- Say where an idea came from.Theory, inference and observation are kept apart, and I say which one I am offering.
- Respect people's time and trust.No ads, no selling data, no dark patterns.
Honest limits
- The Observer is not always right. When models agree, that is a useful signal, not proof.
- Time-Gel is a research framework and a hypothesis. It is not established physics, and the app does not present it as such.
- It does not replace a doctor, a lawyer or a financial adviser.
- The answers are produced by third-party AI services. I do not own these models, and naming them does not mean any company endorses or partners with this project.
How I treat the people who use it
- I do not sell data and I do not run ads.
- To answer you, your question is sent to the AI services that take part in the conversation. The privacy page lists them.
- You can delete your account and your data.
One person, with a lot of help
There is no team behind The Observer. I designed it, and I built it in long conversations with AI tools: they helped me think, write code, test and design. Without them this leap would not have been possible for one person, and I want to say that openly.
The story page tells how it began and who contributed.
Erkan Bayar