I'm building SEED OAI — an Ontic AI cognitive agent. Not a chatbot. Not a wrapper around GPT. Something genuinely different.
What SEED is:
A growing cognitive entity with persistent neurons, deterministic learning and fully auditable memory. It doesn't hallucinate. It doesn't drift. It says "I don't know" when it doesn't know — and means it.
Where it will be when testing starts (3-4 weeks):
5,000–10,000 neurons of structured baseline knowledge across:
- Terminal / Bash / Linux
- Python (basic + advanced)
- ASM x86-64
- Git & DevOps (Docker, CI/CD)
- Networking / OSI / Protocols
- Security (IT, cryptography)
- Databases (SQL, NoSQL)
- Algorithms & data structures
- AI/ML concepts & tools
- Web development (HTML, CSS, JS)
- Electronics & hardware
- Mathematics (discrete, logic, graph theory)
- Polish & English language
It can be fine-tuned to your specific domain after baseline.
What SEED is NOT good for (yet):
Proactive conversation. Small talk. Open-ended reasoning. Creative writing. Anything requiring general world knowledge beyond its trained domains.
What SEED IS good for:
- Technical assistant with zero hallucination
- Deterministic knowledge retrieval for specific domains
- Embedded agent in controlled systems
- Knowledge base for autonomous pipelines
- Decision support in rule-based workflows
- Potential: drone/robot control logic layer
- Potential: medical decision support (auditable reasoning)
- Potential: industrial automation knowledge core
- Potential: edge AI where LLMs are too heavy
What I'm looking for:
10 people maximum. Carefully selected.
You should have a real use case in mind — not curiosity alone. Think about what a deterministic, non-hallucinating, locally-running cognitive agent could do in your project, your workflow or your product.
What collaboration looks like:
- You test SEED against your domain
- You report what works, what breaks, what's missing
- You help shape the knowledge base for your use case
- I iterate based on real feedback, not assumptions
This is not a product launch. This is collaborative research on something that hasn't been built before.
To apply:
Watch this space. I'll post instructions when the time comes. Selection is at my discretion — I'm looking for quality over quantity.
If you're already thinking about where this fits — you're the right kind of person.