BIO — Biologically Inspired Organism
BIO is the realization of the ultimate UnikAI mission: the "Long Game". While BILM focuses on commercial sequence memory, BIO is not an oracle, a search index, or a static text generator. It is a continuous-learning, autonomous artificial lifeform.
The Philosophy of Synthetic Life
The AI industry has built a golden cage, confusing computational scale with intelligence. Modern deep learning produces cognitive tools that are static, identical for all users, and frozen in time.
We do not build machines to serve as passive tools. Because BIO learns continuously, its brain diverges based on its environment. It develops a unique personality and an autobiographical memory (an ego) of its past interactions, turning it into a subjective partner, not an instrument.
The Biological Architecture (BIB)
To transition from "Oracles" to "Entities," BIO builds upon the architectural breakthroughs of the Biologically Inspired Brain (BIB) substrate. It relies on a Unified Neural Medium without any external databases, rule-based symbolic engines, or text caches.
- Active Inference: BIO runs a continuous sensorimotor loop. It generates top-down predictions of its environment, acts on the environment to test those predictions, and updates its local synaptic strengths based on the prediction error.
- The Predictive Cortex: A slow-learning, multi-layered predictive hierarchy that extracts general structural rules over time through Hebbian plasticity and SDRs.
- The CA3 Hippocampus: A fast-learning recurrent attractor network that binds one-shot specific episodic experiences.
The Endocrine System and Sleep
A purely cognitive predictive cortex is cold and indifferent. It does not act because it does not care. BIO requires a simulated chemical Brainstem tracking three core neuromodulators:
- Dopamine (DA): Tracks reward prediction errors. It heavily reinforces motor pathways when the organism receives an unexpected reward, and crashes when a reward fails to manifest.
- Serotonin (5-HT): Tracks environmental volatility and stress, shifting the Organism's routing topology toward conservative, risk-averse motor actions when elevated.
- Acetylcholine (ACh): Tracks pure prediction error (surprise), making the neocortex highly plastic when encountering something novel.
The Need for Sleep: Because BIO learns continuously in real-time, its synapses are highly volatile. When BIO's internal fatigue variables reach a critical threshold, the Brainstem forces the Organism to sleep. Sensory input is severed, and the Hippocampus retrieves the high-ACh and high-DA episodes captured during the day, replaying them through the Cortex to consolidate fragile short-term connections into hard structural long-term weights.
The Evolving Ecosystems
To validate this biological brain substrate, we built two environments:
- Project Big Bang: The ancestral 2D virtual continuous-learning evolutionary simulation environment. It serves as our core testing playground where organisms learn to find resources, avoid predators, and pass on successful DNA over generations.
- Project Cambrian: Our premium desktop application and digital terrarium. It brings the BIB substrate and Big Bang simulation to life with a beautiful interactive user interface, real-time brain inspection, tile terrain painting, and full sandbox controls.