Definition
Attention mechanism
A technique that lets a neural network focus on the most relevant parts of its input when producing each piece of output. It is the core idea behind transformers.
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
Ambient AIAI that operates continuously in the background of an environment or workflow, providing value without requiring explicit user initiation or interaction. Ambient AI monitors, processes, and acts on data streams without being summoned - generating outputs that appear in the user's environment as relevant context, completed tasks, or proactive recommendations. Examples include AI that sweeps news feeds to surface relevant competitive intelligence in a marketing dashboard, systems that pre-populate spreadsheets from incoming emails, and tools like Aryn that process documents and deliver structured outputs to downstream systems without the user ever opening an AI interface. The contrast is with reactive AI (chatbots, on-demand tools) that require the user to initiate each interaction.Ambient alignmentCoordination that is simply present in the workflow rather than enforced through meetings or dashboards.Artificial hivemindA network of AI agents that share information and coordinate their actions so they behave like a single collective system.Artificial intelligence (AI)The science and engineering of making machines that can perform tasks that normally require human intelligence. The term was coined by computer scientist John McCarthy in 1955.Auditable Thinking TraceA transparent log of how an AI system arrived at its conclusions, including citations and reasoning steps. Required for legal AI deployments where attorneys must be able to verify and defend every recommendation.BackpropagationHow neural networks learn from mistakes: the error is traced backward through the network to work out which connections to adjust.Bayesian networkA model that maps how variables influence each other using probabilities, useful for reasoning about causes and likely outcomes.Bias (in AI)When an AI system produces results that systematically favor or disadvantage certain groups, often because of skewed training data.