Answer Engine
Direct, sourced answers to the most-asked questions about artificial intelligence, AGI, neuroscience, neurotechnology, and the future of intelligence. Each answer links to the full BrainMatter explainer it is drawn from.
AGI is a machine system that can perform any intellectual task a human can - flexibly transferring learning across domains, reasoning about novel problems, and acquiring new skills without task-specific retraining. Today's large language models exhibit narrow generality but lack the autonomy, embodied grounding, and continual learning that AGI definitions require.
Read more - AGI cluster →AI refers to systems that perform specific cognitive tasks - translation, image recognition, code completion. AGI refers to a single system capable of the full breadth of human cognitive work. Modern AI is narrow even when impressive; AGI is a threshold of generality, autonomy, and transfer that no current system has crossed.
Read more - AI vs AGI comparison →An LLM is a transformer neural network trained to predict the next token in a sequence. During training it processes trillions of tokens and adjusts billions of parameters to minimize prediction error. At inference, it generates output one token at a time, conditioned on the prompt and its learned distribution over language.
Read more - LLM definition →A BCI is a system that translates neural activity into commands a computer can act on - or delivers stimulation back into the brain. Invasive BCIs (Neuralink, Synchron) implant electrodes directly into cortex; non-invasive BCIs (EEG headsets) read scalp signals. Clinical use cases include restoring movement, speech, and vision.
Read more - BCI definition →Neurodivergence describes cognitive variation that diverges from majority neurotypes - including autism, ADHD, dyslexia, dyspraxia, Tourette's, and giftedness. It frames these differences as natural human variation rather than disorders, while recognizing that some neurodivergent profiles include disabling features that benefit from clinical support.
Read more - Neurodivergence cluster →AI is in clinical use for radiology (detecting tumors, fractures, hemorrhages), pathology (slide analysis), drug discovery (protein folding via AlphaFold, lead generation), clinical documentation (ambient scribes), and triage. Most deployments augment rather than replace clinicians, and FDA-cleared devices undergo prospective evaluation.
Read more - Human + AI cluster →AI alignment is the research field focused on ensuring AI systems pursue the goals their designers and users actually intend - including under distribution shift, capability gain, and adversarial pressure. Sub-fields include interpretability, scalable oversight, reward modeling, and corrigibility.
Read more - Alignment definition →The connectome is the comprehensive wiring diagram of a brain - every neuron and every synaptic connection between them. The first complete mammalian connectome (a fragment of mouse cortex) was published in 2024. Connectomics is foundational to mechanistic neuroscience and biologically-grounded AI.
Read more - Connectome definition →Frontier reasoning benchmarks include MMLU (multitask academic knowledge), GPQA (graduate-level science Q&A), ARC-AGI (abstract pattern reasoning), BIG-Bench (broad capability suite), HumanEval (code generation), and SWE-Bench (real software engineering). Each measures a different facet - knowledge recall, novel reasoning, transfer, or agentic capability.
Read more - Benchmark Center →There is no scientific consensus on whether AI systems can be conscious because there is no agreed-on test for consciousness even in biological systems. Major theories - Global Workspace, Integrated Information Theory, Higher-Order Thought - disagree on whether current AI architectures could in principle support phenomenal experience.
Read more - Consciousness definition →The singularity is the hypothesized point at which recursively self-improving AI triggers an intelligence explosion that exceeds humans' ability to understand or control it. The timeline and even the possibility are contested; the term originates with mathematician I. J. Good and was popularized by Vernor Vinge and Ray Kurzweil.
Read more - Future of Humanity →Human memory is multi-system: working memory (seconds, prefrontal cortex), declarative long-term memory (facts and events, hippocampus-dependent), and procedural memory (skills, striatum and cerebellum). Consolidation moves new memories from hippocampus to distributed cortical storage, primarily during sleep.
Read more - Working memory →Looking for a deeper reference? Browse the Glossary, Research Library, or Benchmark Center.
