Jason Duke, Founder, Kronaxis
Tag: Framework
The conventional wisdom in enterprise software is straightforward: keep your methodology proprietary, charge for access, and defend the moat. Qualtrics does this. Evidenza does this. Every major survey platform treats its analytical framework as a trade secret. You pay for the output. You never see the model. You cannot reproduce the results independently. You cannot validate the methodology. You cannot extend it.
We went the other way. The DYNAMICS-8 specification, all eight dimensions, all 32 facets, the scoring bands, the derivation formulas, the compatibility algorithm, the validation rules, is published under Creative Commons Attribution 4.0 International. Anyone can read it, use it, modify it, build on it, and distribute it, for any purpose including commercial, with nothing more than attribution.
This article explains why.
The Problem with Proprietary Personality Frameworks
A personality framework is not useful because it exists. It is useful because people trust it. Trust in a personality framework comes from three sources: independent validation (other researchers have tested it and confirmed it measures what it claims to measure), widespread adoption (enough people use it that the results are comparable across studies), and transparency (anyone can inspect the methodology and identify flaws).
Proprietary frameworks fail on all three counts. If the methodology is a black box, independent researchers cannot validate it. If the framework requires a paid licence to use, adoption is limited to customers who can afford it. If no one can inspect the scoring rules, no one can identify when the framework produces nonsensical results in edge cases.
The Big Five has been the dominant personality framework in psychology for three decades. Not because it is perfect. Because it is open. Thousands of researchers have tested it, criticised it, extended it, and published findings using it. The corpus of validation evidence is enormous. The framework became a standard because it was available for anyone to use, and widespread use produced the evidence that justified further use. That is a network effect, not a marketing strategy.
HEXACO followed the same path. Published openly. Adopted by researchers who found six dimensions more predictive than five. Validated across cultures and languages. Now a serious competitor to the Big Five in academic personality research, entirely because it was open from the start.
A proprietary personality framework, no matter how sophisticated, cannot replicate this dynamic. If only your customers can use it, only your customers will validate it, and their incentive is to confirm their purchase was worthwhile, not to identify flaws.
The Network Effect
Every researcher who uses DYNAMICS-8 in a study generates evidence about the framework. Some of that evidence will be positive: the dimensions predict behaviour accurately in a new context. Some will be negative: a dimension does not work as expected in a specific cultural setting, or two facets overlap more than the specification suggests. Both kinds of evidence increase the value of the framework.
Positive evidence builds confidence. Negative evidence identifies where the specification needs refinement. A framework that only receives positive evidence from its commercial users is either perfect (unlikely) or suffering from confirmation bias (almost certain).
Every dataset tagged with DYNAMICS-8 scores becomes interoperable with every other DYNAMICS-8 dataset. A research team in Berlin and a team in Tokyo can directly compare results because the scoring methodology is identical and public. Proprietary frameworks produce data silos. Open frameworks produce data networks. Every model trained on DYNAMICS-8 traces works with any DYNAMICS-8 profiled persona, from any source, in any application. The training investment is portable.
What Is Open, What Is Commercial
Publishing the specification does not mean giving away the business. It means separating the framework from the platform.
Open (CC BY 4.0): The eight dimension specification. All 32 facet definitions. The scoring methodology. The derivation formulas for income band, spending pattern, risk tolerance, and political lean. The compatibility algorithm. The validation rules. The reference implementations in Python and Go. The 1,000 persona sample dataset on HuggingFace (500 UK, 500 US). Everything you need to understand, implement, and use DYNAMICS-8 in your own work.
Commercial (Panel Studio, Business Source Licence 1.1): The hosted platform. Census weighted persona generation across 20 countries. The three pass generation pipeline. The 18 rule validation engine. The 226 route API. Multiturn conversations. Reasoning traces. A/B testing. 17 stimulus templates. Survey builder. Focus groups. Segmentation discovery. Brand tracking. Regulatory simulation. 26 platform integrations. Four export formats. The Python SDK. All the infrastructure that turns the framework into a production research tool.
The analogy is straightforward. The DYNAMICS-8 specification is the equivalent of the HTTP specification. Panel Studio is the equivalent of the web server. The specification is free. The platform that implements it at scale is a product. You can build your own implementation from the specification. Most people will choose not to, because the hosted platform is faster, cheaper, and more capable than building from scratch. But the option is always there, and its existence is what gives the specification credibility.
The Precedent
This is not a novel strategy. It is the dominant strategy in AI infrastructure.
Google open sourced TensorFlow. It became the industry standard. Google Cloud became the obvious place to run it at scale. Meta open sourced PyTorch. Same dynamic. In every case, the company that opened the framework captured more value than it gave away. Ecosystem growth driven by open adoption exceeded anything proprietary marketing could have achieved.
DYNAMICS-8 follows the same logic. An open personality framework that becomes a standard for personality conditioned AI is worth more to Kronaxis than a proprietary framework that only our customers use. The specification's value increases with every external adoption. The platform's value increases with the specification's credibility. The two reinforce each other.
What We Ask in Return
The CC BY 4.0 licence asks for one thing: attribution. If you use DYNAMICS-8 in a study, cite it. If you build on it, credit it. If you extend it, publish the extension so others benefit.
Beyond the licence requirement, what we actually want is engagement. Use the framework. Test it against your data. Find the edge cases where it breaks. Publish your findings. If you build something better for a specific domain, publish that too.
The goal is a shared standard for personality conditioned AI. A standard that any researcher can use, any developer can implement, and any organisation can deploy. Proprietary personality frameworks protect their revenue by hiding their methodology. We protect ours by making the methodology so widely adopted that the platform built on top of it becomes the obvious choice. The moat is not secrecy. The moat is trust.
The specification is at kronaxis.co.uk/dynamics. The reference implementations are on GitHub. The sample dataset is on HuggingFace. Read them. Use them. Build on them. That is the point.