Definition
Parameter
An internal value a model learns during training. Large models have billions of them, and together they hold what the model knows.
Learn the language of AI-native founders
Every week: the new terms, tools, and workflows founders are using, straight from the podcast.
Related terms in AI Fundamentals
Optimization algorithmThe method a model uses during training to adjust its parameters and make fewer mistakes, such as gradient descent.Opus-class vs Sonnet-classShorthand for top-tier versus the still-excellent, cheaper tier most workloads can live in.Over-Provisioned LLMA large language model deployment in which the model is being fed significantly more raw, unclean, or poorly structured data than necessary - compensating for data quality problems with raw compute power rather than fixing the underlying data. The result is higher cloud infrastructure costs, increased latency, and often lower output quality than a well-structured smaller-context approach would produce. David Carmel uses the metaphor of 'dropping bunker busters to break up data when a scalpel would work.' An over-provisioned LLM is typically a symptom of an upstream data quality problem that has not been addressed: rather than cleaning the data, the organization throws more model capacity at it and hopes the model can compensate.OverfittingWhen a model memorizes its training data, noise and all, so it performs well in testing but poorly on anything new.ParameterizedAbsorbed into a model's weights, as distinct from stored as a retrievable file.Persona ArchitectureThe deliberate design of distinct AI characters with consistent voice, behavioral patterns, and identity - each mapped to a recognizable human archetype. In AI products, personas function as both UX differentiation and distribution channels: each persona attracts a specific audience segment and creates shareable identity moments. Effective persona architecture draws on established psychological frameworks (Jungian archetypes, cultural identities) to give characters depth beyond surface-level branding. Commitify's six personas - Life Coach, Zen Master, Slay Bestie, Hype Beast, Drill Sergeant, CEO - each serve as acquisition vectors targeting distinct communities.Personal Data LakeAn individual's curated pipeline of personal context shared with AI tools — the personal-scale equivalent of an enterprise data lake. Raises strategic questions: what data do you share with which systems, does it stay on-device (Apple Intelligence model) or go to servers (Google model), and which companies do you trust as data stewards? Geoff sees an emerging bifurcation between ad-funded AI services and privacy-preserving paid services — with major trust implications for both.PersonificationAssigning human traits to an AI system. Ross Barnes distinguishes this from assigning tone, and argues personification erodes accountability.