How to progress toward Artificial General Intelligence (AGI)

  1. Basic pattern recognition. Extract features from raw sensory inputs; for example, the ears convert raw sounds to frequency maps; the eyes convert visual inputs to feature maps (used for facial recognition, etc.) Associate new perceptions with existing memories (via dimension reduction). Use algorithms to actively filter out irrelevant stimuli (habituation) and ignore stimuli after repeated exposure (adaptation).
  2. Mental representation. Mentally represent and simulate the state of the environment. Track physical objects across time, space and sensory inputs, and understand that physical objects are solid, can support other objects, and can otherwise affect other objects. Decompose objects into hierarchical sub-objects (a car has tires and a motor). (Mental representations include all mentally simulated concepts, including physical objects, intentional objects, sensory features, event metadata, goals, plans, memories, language syntax, words, etc.)
  3. Spatial navigation. Understand how objects navigate through space and time. Maintain a mental map of previous locations. Control one’s own movements using feedback loops.
  4. Event metadata. Tag objects & events with basic metadata: space, time, quantity, causality, attribution/source, confidence/probability, etc. Space is where events happen (near, far, landmark). Time is when they happen (time frame, recent, future). Quantity is how many objects there are (one, a few, many). Causality is why events happen, or who made them happen. Attribution is why you think they happened (who told you, confidence in the source).
  5. Attention, memory & belief. Maintain a working memory of recent or relevant events, and focus attention on the new, unusual, or unexpected. Consolidate short term memories into persistent long term storage. Maintain confidence levels, evidence, sources, and probabilities behind beliefs. Tag learning with uncertainty.
  6. Prediction & simulation. Mentally simulate alternative environmental states and outcomes using mental representations. Determine the aspects of the environment that best predict outcomes, and track their prediction error. Identify general laws and models from specific cases, and select best models to explain events.
  7. Agency. Understand the special properties of living creatures (intentional objects) having their own agency (motivated from within). Understand that others have preferences & intentions that differ from our own, and sometimes they make honest mistakes. Understand that there are also bad actors, cheaters and freeloaders. Understand what other people are thinking, what they know, and their credibility/reliability (Theory of Mind).
  8. Language. Understand the basic rules of grammar, and semantic primitives. Track word frequencies. Maintain shared assumptions, memories, interests, and values with others. Justify, rationalize, and explain own beliefs to others. Have ability to persuade, convince, attract, befriend, compliment, criticize, and manipulate others, and gain their trust and approval. Use language to transmit mental models, algorithms, and representations to others (e.g. “Don’t trust a guy who doesn’t love his mother.”)
  9. Plan & Act. Maintain a hierarchy of goals, sub-goals, plans, and agendas. For example, complete school, find a job, find a mate, have kids, seek life’s purpose, join a cause, etc. Maximize internal rewards for any progress toward final or interim goals.
  10. Emotions, feelings and instincts. Act on prior goals and instincts (hunger, fear, social anxiety, sex-drive, fight-or-flight, addiction, risk-taking). Take action based on internal motivation, ambition, desire, drives, instincts, passions, interests, greed, obsession. Modify behavior based on guilt, shame, fear of judgment, desire to lead or be led, desire for retribution and punishment, and need for approval, attention, power, and status.
  11. Mental models and algorithms. Manage 1000’s of concurrent models (algorithms that process mental representations), some of which correspond to “common sense”. Integrate new information into existing models; Learn from scarce data. Teach & share knowledge. Learn science, history, music, how to drive a car, a vocation, etc. Learn math, e.g., apply the “order of operations” rule (PEMDAS) when solving equations like (3⁴ - 4 * (5–2)) — Parentheses first, then Exponents, Multiplication/Division, and Addition/Subtraction).
  12. Imagination & creativity. Replay the day’s events and mentally test (simulate) hypotheses and scenarios. Generate potential new models, counterfactuals and what-ifs. Balance exploration of the unknown vs. exploitation of the known. Actively seek to fill in your knowledge gaps — know what you don’t know. (Consciousness emerges from the resonance between the actual environment — i.e., reality — and our mental simulation of it, and the dissonance introduced by our own motivations, fears and desires.)

A computer scientist with a passion for AI and Cognitive Science.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store