What is Cognicism? A Glossary
Cognicism: An adaptive framework and meta-ideology built on the foundation of generative language models and decentralized data storage to foster intelligent collaboration and informed decision-making within communities. Central to the framework are Iris — a generative belief encoding model — and Ŧrust which represents credibility within communities. By learning to allocate attention effectively through communal feedback, cognicism fosters collaboration and ethical decision-making rooted in shared wisdom instead of tokenized value or short-term gains.
Ŧrust: Within the cognicism framework Ŧrust represents a form of power and influence that transcends traditional notions of money. While money can accumulate and concentrate power in the hands of a few, Ŧrust operates as a decentralized mechanism that distributes attention and credibility within a community. Ŧrust takes into account factors such as the reliability of sources, their alignment with communal values, and their past performance. At the heart of Ŧrust lies the Iris model, which serves as a communal interface for pooling knowledge and forming direction together. Through the allocation of Ŧrust, the Iris model enables informed decision-making and fosters collaboration rooted in shared wisdom and collective intelligence within the community. It is a democratic value signal regulated by communal oversight, fostering collaboration and ethical decision-making rooted in shared wisdom.
Iris: A generative belief encoding model within the cognicism framework that utilizes source embeddings and attention distribution parameters to track community contributions over time and enhance understanding of community narratives and actions. Iris is an essential component that drives Ŧrust by processing and evaluating information from various sources to facilitate informed decision-making and align with communal values in a dynamic, transparent manner.
FourThought: A dialectic process and belief-tracking system within the cognicism framework that emphasizes accurate predictions, insightful questions, helpful reflections on the past, and informative statements for decision-making and community growth. The FourThought protocol is designed to interface with democratic large language models (LLMs), providing a structured schema for input and output and facilitating decentralized scientific inquiry and governance. Built upon the idea of staking claims and asking questions in a coherent and dynamic manner, the protocol includes two voting mechanisms for each claim: valence and uncertainty. Valence reflects the alignment of a thought with one’s morality, while uncertainty represents the alignment with one’s sense of truth. FourThought enables the logging and staking of beliefs about the past, present, and future for future assessment, fostering a collaborative and rigorous approach to knowledge creation and sharing within communities.
Belief Staking: A mechanism within cognicism where individuals “stake” their beliefs, expressing their confidence in a particular outlook, impacting their Ŧrust and influence within the community, as part of the FourThought protocol.
Social Proof of Impact: A mechanism within the cognicism framework that leverages societal validation and consensus to amplify the influence of ideas and actions deemed socially beneficial over time. By nurturing a community that values long-term foresight and collaboration, it helps demonstrate an individual’s or group’s positive influence or contributions within the community and can influence their Ŧrust and credibility, thereby promoting the impact of ideas and actions that align with societal goals and values.
Prophet Incentive: A motivating factor within the cognicism framework that rewards accurate predictions, fostering the pursuit of truth, reliability, and constructive community contributions, as part of a paradigm shift from short-term profit to long-term impact, emphasizing fuzzy textual predictions, and community-driven focus on sustainability and the collective good.
Semantic Ledger: An essential component of the cognicism framework that efficiently stores, retrieves, and tracks collective knowledge through the concept of latent space embeddings in a shared, immutable, distributed, and decentralized record of interactions between Irises.
Loss Function: A mathematical formula that quantifies the difference between the model’s predictions and the actual target values, used to train and optimize the model’s performance, and within Iris, it adjusts attention to sources considering past accuracy and social value.
Social Feedback: The input provided by human evaluators within the cognicism framework that shapes attention allocation, Ŧrust distribution, and the model’s learning and alignment with communal values. Serving as a core incentive structure, it focuses on the distribution of attention rather than the exchangeable tokenization of value, ensuring a community-centered approach to knowledge and decision-making.
Source Embeddings: Vector representations for unique information sources within the cognicism framework, capturing essential relational characteristics like reliability and expertise over time for evaluating claims in the context of past claims.
Temporal Embeddings: Continuous and differentiable representations of time, including sinusoidal embeddings, used by the Iris model within the cognicism framework to capture cyclical patterns, nuanced temporal patterns, and dependencies in data. These embeddings enable Iris to understand community narratives and actions over time in a time-aware manner, enhancing its ability to write informed output that align with the community’s needs and values.
Uncertainty: A measure of the level of confidence or ambiguity associated with a model’s output or prediction and a voting mechanism in FourThought for assessing the level of doubt surrounding a claim.
Moral Valence: The ethical import or significance of particular beliefs, ideas, or actions within the cognicism framework and a voting mechanism in FourThought for evaluating the ethical implications of a claim.
Regenerative Wisdom Tensors: Mathematical representations of the compressed, useful information derived from community knowledge and interactions in the form of tensors that contribute to rebuilding and renewal within a community in the cognicism framework. Learned and compressed through the iterative communal learning process called social proof of impact.
Prediction: An assertion or statement made within the cognicism framework about a future event, scenario, or outcome, as a type of thought in the FourThought protocol.
Reflections: Speakers’ evaluation and understanding of past experiences, events, or decisions within the cognicism framework, as a type of thought in the FourThought protocol.
Statements: Assertions made by speakers within the cognicism framework that convey information or opinions related to a topic or discussion, as a type of thought in the FourThought protocol.
Questions: Inquiries posed by speakers within the cognicism framework designed to stimulate thought, discussion, and learning, as a type of thought in the FourThought protocol.
