Definition
Traditional AI
Earlier AI built on explicitly programmed rules, logic, and human-written knowledge rather than learning from data.
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
Token spendThe compute cost of running large AI models, which Tal cut by building his own model.TokenizationBreaking text into small units called tokens, such as words, word pieces, or characters, that a language model can process.TokenomicsThe unit economics of AI usage, measuring cost per token against productivity produced.TraceabilityThe ability to follow an AI decision back to the data and logic behind it.Training dataThe examples, like text, images, or audio, used to teach a machine learning model how to do its task.Transfer learningReusing a model trained on one task as the starting point for a related task, which saves data and time.TransformerThe neural network architecture behind modern AI. It uses attention to look at all parts of the input at once instead of one piece at a time.Transparent AI ReasoningThe design requirement that an AI system articulate its reasoning — what it observed, what that implies, and what published evidence supports the inference — before its output reaches a human decision-maker. Axenya's implementation: every clinical inference surfaces with a reasoning chain and supporting literature citations; inferences the model cannot explain are suppressed entirely. Addresses two risks simultaneously: clinician liability (the physician can see and evaluate the reasoning before acting on it) and AI error rates (unexplained outputs may reflect spurious correlations rather than valid clinical patterns).