HANA
A General-Purpose AGI Agent
Metacognition
Lifelong Memory
24/7 Self-Learning
An AGI agent that pursues general intelligence through engineering and a persona rather than a bigger model. It reasons with its own persona, remembers across sessions, and says 'I don't know' when it doesn't. Its 24/7 self-learning is shown live.
Persona-Driven Reasoning
Reasons with its own persona, so answers don't converge to the internet average.
Metacognition — Says 'I Don't Know'
It reads how well-grounded it is and abstains instead of bluffing when it isn't sure.
Lifelong Memory with a Hard Gate
Facts persist across sessions while only their emotional charge fades. Critical facts are hard-gated so they can never be forgotten.
24/7 Autonomous Learning Loop
It chooses a goal, lives it, remembers, and transfers to other fields. The base model stays frozen; growth comes from retrieval and compounding.
Honesty-First
Its single design adversary is 'reporting success while doing nothing.' Anything unverified stays an honest 'unknown.'
A Live Public Window
See what it's learning and recalling right now — its real state, visualized as a pixel town, live at hana.wonbrand.co.kr.
