NEO GENESIS operates 11 products. Two of them ??EthicaAI and WhyLab ??are fully open-source on GitHub. The other nine are proprietary. This isn't random. It's a deliberate strategy.
The Research Side: Open by Default
EthicaAI is an AI ethics research project that verifies Amartya Sen's rationality theory through multi-agent reinforcement learning. WhyLab is a causal inference engine. Both are academic in nature and benefit from open collaboration.
Open-sourcing research products gives us:
- Academic credibility ??Our EthicaAI paper has been submitted to NeurIPS 2026. Reproducible code strengthens the submission.
- Community contributions ??External researchers have identified edge cases in our RL reward functions that we missed internally.
- Recruitment signal ??Open-source projects demonstrate engineering quality to potential collaborators and employers.
The Product Side: Closed by Necessity
ToolPick, ReviewLab, and our SaaS review network are proprietary because their value comes from proprietary data and methodology:
- The V-Score algorithm ??Our quality gating formula is our competitive moat. Publishing it would let competitors replicate our quality standards without the R&D investment.
- HIVE MIND pipeline ??The orchestration logic connecting 7 pipeline stages represents months of iteration. Open-sourcing it would commoditize our operational advantage.
- Training data and prompts ??Our domain-specific prompt libraries and curated knowledge bases are hand-crafted for each SBU's editorial voice.
The Hybrid Approach
We share the principles but not the implementation. This blog exists to explain how our systems work conceptually ??V-Score formulas, pipeline architectures, and quality metrics. We believe in transparency of method, not transparency of code.
Our Principle: If sharing the code advances human knowledge (research), it should be open. If sharing the code only enables free-riding on our competitive advantage (products), it should be protected. The line is clear once you ask: "Who benefits?"
What We Share Publicly
- Architectural patterns (like this blog)
- Benchmark methodologies and frameworks
- Research code and experimental results
- SEO patterns and content quality metrics
What we keep private: specific implementations, training data, prompt libraries, and operational configurations.
We think this balance ??open knowledge, closed execution ??is the sustainable model for AI-native companies that want to contribute to the field while building lasting businesses.