NoticeThis site demonstrates one possible use of this domain. For acquisition, partnership, or investment inquiries, please use our contact form.
A concise, cross-linked encyclopedia of the core concepts that show up across every page of this site - from attention to alignment, from neuroplasticity to superintelligence.
The capacity to acquire knowledge, reason about it, and use it to achieve goals in varied environments. There is no single agreed definition - psychometric, computational, and behavioural traditions each emphasise different facets.
A hypothetical system that can learn and reason across the full range of cognitive tasks at or above human level, transferring knowledge fluidly between domains.
A hypothetical system whose general cognitive performance substantially exceeds the best human minds across essentially all domains.
An umbrella term for systems that perform tasks normally requiring human intelligence: perception, language, planning, problem solving.
Systems that learn statistical patterns from data rather than being explicitly programmed with rules.
Multi-layer neural networks that learn hierarchical representations - the engine behind modern AI capabilities.
A neural network architecture introduced in 2017 that uses self-attention to model relationships across long sequences. The backbone of contemporary large language models.
A transformer-based model trained on large corpora of text to predict tokens, producing fluent language and showing emergent reasoning behaviour at scale.
A computational model loosely inspired by biological neurons, composed of layers of weighted connections trained to minimise prediction error.
The process of drawing conclusions from premises, evidence, or prior knowledge - including deductive, inductive, abductive, and analogical forms.
The act of deriving new information that is not explicitly stated, from data that is. Distinct from memorisation or pattern matching.
Applying knowledge learned in one context to new, unseen situations. A central criterion in evaluating both biological and artificial intelligence.
The selective allocation of limited cognitive or computational resources to a subset of available information. Implemented biologically in the brain and algorithmically in transformer models.
The encoding, storage, and retrieval of information over time. In humans, divided into sensory, working, short-term, and long-term systems.
Higher-order cognitive control: planning, inhibition, working memory, cognitive flexibility, and goal management.
Generating outputs that are both novel and useful. Involves divergent search, recombination of existing concepts, and selective filtering.
Subjective, first-person experience. Whether and how artificial systems could possess it remains an open scientific and philosophical question.
The brain's ability to reorganise its structure and function in response to experience, learning, or injury.
A system that establishes a direct communication channel between neural activity and an external device, for control, restoration, or augmentation.
The technical and conceptual problem of ensuring that advanced AI systems pursue goals consistent with human values and intentions.
The discipline of building AI systems that behave reliably and avoid causing harm, especially as capabilities scale.
A framing that emphasises AI as a tool that extends human cognitive capacity rather than replacing it.
Intelligence that emerges from the coordinated activity of many agents - humans, machines, or both.
An alternative term for non-biological intelligence, emphasising that it is engineered rather than evolved.
The hypothetical instantiation of conscious experience in a computational substrate. Currently speculative and contested.